Commit c884c12b authored by Artem Baranovskyi's avatar Artem Baranovskyi
Browse files

Final commit.

parent 81a9e9f3
import typing as t
from . import nodes
from .visitor import NodeVisitor
VAR_LOAD_PARAMETER = "param"
VAR_LOAD_RESOLVE = "resolve"
VAR_LOAD_ALIAS = "alias"
VAR_LOAD_UNDEFINED = "undefined"
def find_symbols(
nodes: t.Iterable[nodes.Node], parent_symbols: t.Optional["Symbols"] = None
) -> "Symbols":
sym = Symbols(parent=parent_symbols)
visitor = FrameSymbolVisitor(sym)
for node in nodes:
visitor.visit(node)
return sym
def symbols_for_node(
node: nodes.Node, parent_symbols: t.Optional["Symbols"] = None
) -> "Symbols":
sym = Symbols(parent=parent_symbols)
sym.analyze_node(node)
return sym
class Symbols:
def __init__(
self, parent: t.Optional["Symbols"] = None, level: t.Optional[int] = None
) -> None:
if level is None:
if parent is None:
level = 0
else:
level = parent.level + 1
self.level: int = level
self.parent = parent
self.refs: t.Dict[str, str] = {}
self.loads: t.Dict[str, t.Any] = {}
self.stores: t.Set[str] = set()
def analyze_node(self, node: nodes.Node, **kwargs: t.Any) -> None:
visitor = RootVisitor(self)
visitor.visit(node, **kwargs)
def _define_ref(
self, name: str, load: t.Optional[t.Tuple[str, t.Optional[str]]] = None
) -> str:
ident = f"l_{self.level}_{name}"
self.refs[name] = ident
if load is not None:
self.loads[ident] = load
return ident
def find_load(self, target: str) -> t.Optional[t.Any]:
if target in self.loads:
return self.loads[target]
if self.parent is not None:
return self.parent.find_load(target)
return None
def find_ref(self, name: str) -> t.Optional[str]:
if name in self.refs:
return self.refs[name]
if self.parent is not None:
return self.parent.find_ref(name)
return None
def ref(self, name: str) -> str:
rv = self.find_ref(name)
if rv is None:
raise AssertionError(
"Tried to resolve a name to a reference that was"
f" unknown to the frame ({name!r})"
)
return rv
def copy(self) -> "Symbols":
rv = object.__new__(self.__class__)
rv.__dict__.update(self.__dict__)
rv.refs = self.refs.copy()
rv.loads = self.loads.copy()
rv.stores = self.stores.copy()
return rv
def store(self, name: str) -> None:
self.stores.add(name)
# If we have not see the name referenced yet, we need to figure
# out what to set it to.
if name not in self.refs:
# If there is a parent scope we check if the name has a
# reference there. If it does it means we might have to alias
# to a variable there.
if self.parent is not None:
outer_ref = self.parent.find_ref(name)
if outer_ref is not None:
self._define_ref(name, load=(VAR_LOAD_ALIAS, outer_ref))
return
# Otherwise we can just set it to undefined.
self._define_ref(name, load=(VAR_LOAD_UNDEFINED, None))
def declare_parameter(self, name: str) -> str:
self.stores.add(name)
return self._define_ref(name, load=(VAR_LOAD_PARAMETER, None))
def load(self, name: str) -> None:
if self.find_ref(name) is None:
self._define_ref(name, load=(VAR_LOAD_RESOLVE, name))
def branch_update(self, branch_symbols: t.Sequence["Symbols"]) -> None:
stores: t.Dict[str, int] = {}
for branch in branch_symbols:
for target in branch.stores:
if target in self.stores:
continue
stores[target] = stores.get(target, 0) + 1
for sym in branch_symbols:
self.refs.update(sym.refs)
self.loads.update(sym.loads)
self.stores.update(sym.stores)
for name, branch_count in stores.items():
if branch_count == len(branch_symbols):
continue
target = self.find_ref(name) # type: ignore
assert target is not None, "should not happen"
if self.parent is not None:
outer_target = self.parent.find_ref(name)
if outer_target is not None:
self.loads[target] = (VAR_LOAD_ALIAS, outer_target)
continue
self.loads[target] = (VAR_LOAD_RESOLVE, name)
def dump_stores(self) -> t.Dict[str, str]:
rv: t.Dict[str, str] = {}
node: t.Optional["Symbols"] = self
while node is not None:
for name in sorted(node.stores):
if name not in rv:
rv[name] = self.find_ref(name) # type: ignore
node = node.parent
return rv
def dump_param_targets(self) -> t.Set[str]:
rv = set()
node: t.Optional["Symbols"] = self
while node is not None:
for target, (instr, _) in self.loads.items():
if instr == VAR_LOAD_PARAMETER:
rv.add(target)
node = node.parent
return rv
class RootVisitor(NodeVisitor):
def __init__(self, symbols: "Symbols") -> None:
self.sym_visitor = FrameSymbolVisitor(symbols)
def _simple_visit(self, node: nodes.Node, **kwargs: t.Any) -> None:
for child in node.iter_child_nodes():
self.sym_visitor.visit(child)
visit_Template = _simple_visit
visit_Block = _simple_visit
visit_Macro = _simple_visit
visit_FilterBlock = _simple_visit
visit_Scope = _simple_visit
visit_If = _simple_visit
visit_ScopedEvalContextModifier = _simple_visit
def visit_AssignBlock(self, node: nodes.AssignBlock, **kwargs: t.Any) -> None:
for child in node.body:
self.sym_visitor.visit(child)
def visit_CallBlock(self, node: nodes.CallBlock, **kwargs: t.Any) -> None:
for child in node.iter_child_nodes(exclude=("call",)):
self.sym_visitor.visit(child)
def visit_OverlayScope(self, node: nodes.OverlayScope, **kwargs: t.Any) -> None:
for child in node.body:
self.sym_visitor.visit(child)
def visit_For(
self, node: nodes.For, for_branch: str = "body", **kwargs: t.Any
) -> None:
if for_branch == "body":
self.sym_visitor.visit(node.target, store_as_param=True)
branch = node.body
elif for_branch == "else":
branch = node.else_
elif for_branch == "test":
self.sym_visitor.visit(node.target, store_as_param=True)
if node.test is not None:
self.sym_visitor.visit(node.test)
return
else:
raise RuntimeError("Unknown for branch")
if branch:
for item in branch:
self.sym_visitor.visit(item)
def visit_With(self, node: nodes.With, **kwargs: t.Any) -> None:
for target in node.targets:
self.sym_visitor.visit(target)
for child in node.body:
self.sym_visitor.visit(child)
def generic_visit(self, node: nodes.Node, *args: t.Any, **kwargs: t.Any) -> None:
raise NotImplementedError(f"Cannot find symbols for {type(node).__name__!r}")
class FrameSymbolVisitor(NodeVisitor):
"""A visitor for `Frame.inspect`."""
def __init__(self, symbols: "Symbols") -> None:
self.symbols = symbols
def visit_Name(
self, node: nodes.Name, store_as_param: bool = False, **kwargs: t.Any
) -> None:
"""All assignments to names go through this function."""
if store_as_param or node.ctx == "param":
self.symbols.declare_parameter(node.name)
elif node.ctx == "store":
self.symbols.store(node.name)
elif node.ctx == "load":
self.symbols.load(node.name)
def visit_NSRef(self, node: nodes.NSRef, **kwargs: t.Any) -> None:
self.symbols.load(node.name)
def visit_If(self, node: nodes.If, **kwargs: t.Any) -> None:
self.visit(node.test, **kwargs)
original_symbols = self.symbols
def inner_visit(nodes: t.Iterable[nodes.Node]) -> "Symbols":
self.symbols = rv = original_symbols.copy()
for subnode in nodes:
self.visit(subnode, **kwargs)
self.symbols = original_symbols
return rv
body_symbols = inner_visit(node.body)
elif_symbols = inner_visit(node.elif_)
else_symbols = inner_visit(node.else_ or ())
self.symbols.branch_update([body_symbols, elif_symbols, else_symbols])
def visit_Macro(self, node: nodes.Macro, **kwargs: t.Any) -> None:
self.symbols.store(node.name)
def visit_Import(self, node: nodes.Import, **kwargs: t.Any) -> None:
self.generic_visit(node, **kwargs)
self.symbols.store(node.target)
def visit_FromImport(self, node: nodes.FromImport, **kwargs: t.Any) -> None:
self.generic_visit(node, **kwargs)
for name in node.names:
if isinstance(name, tuple):
self.symbols.store(name[1])
else:
self.symbols.store(name)
def visit_Assign(self, node: nodes.Assign, **kwargs: t.Any) -> None:
"""Visit assignments in the correct order."""
self.visit(node.node, **kwargs)
self.visit(node.target, **kwargs)
def visit_For(self, node: nodes.For, **kwargs: t.Any) -> None:
"""Visiting stops at for blocks. However the block sequence
is visited as part of the outer scope.
"""
self.visit(node.iter, **kwargs)
def visit_CallBlock(self, node: nodes.CallBlock, **kwargs: t.Any) -> None:
self.visit(node.call, **kwargs)
def visit_FilterBlock(self, node: nodes.FilterBlock, **kwargs: t.Any) -> None:
self.visit(node.filter, **kwargs)
def visit_With(self, node: nodes.With, **kwargs: t.Any) -> None:
for target in node.values:
self.visit(target)
def visit_AssignBlock(self, node: nodes.AssignBlock, **kwargs: t.Any) -> None:
"""Stop visiting at block assigns."""
self.visit(node.target, **kwargs)
def visit_Scope(self, node: nodes.Scope, **kwargs: t.Any) -> None:
"""Stop visiting at scopes."""
def visit_Block(self, node: nodes.Block, **kwargs: t.Any) -> None:
"""Stop visiting at blocks."""
def visit_OverlayScope(self, node: nodes.OverlayScope, **kwargs: t.Any) -> None:
"""Do not visit into overlay scopes."""
"""Implements a Jinja / Python combination lexer. The ``Lexer`` class
is used to do some preprocessing. It filters out invalid operators like
the bitshift operators we don't allow in templates. It separates
template code and python code in expressions.
"""
import re
import typing as t
from ast import literal_eval
from collections import deque
from sys import intern
from ._identifier import pattern as name_re
from .exceptions import TemplateSyntaxError
from .utils import LRUCache
if t.TYPE_CHECKING:
import typing_extensions as te
from .environment import Environment
# cache for the lexers. Exists in order to be able to have multiple
# environments with the same lexer
_lexer_cache: t.MutableMapping[t.Tuple, "Lexer"] = LRUCache(50) # type: ignore
# static regular expressions
whitespace_re = re.compile(r"\s+")
newline_re = re.compile(r"(\r\n|\r|\n)")
string_re = re.compile(
r"('([^'\\]*(?:\\.[^'\\]*)*)'" r'|"([^"\\]*(?:\\.[^"\\]*)*)")', re.S
)
integer_re = re.compile(
r"""
(
0b(_?[0-1])+ # binary
|
0o(_?[0-7])+ # octal
|
0x(_?[\da-f])+ # hex
|
[1-9](_?\d)* # decimal
|
0(_?0)* # decimal zero
)
""",
re.IGNORECASE | re.VERBOSE,
)
float_re = re.compile(
r"""
(?<!\.) # doesn't start with a .
(\d+_)*\d+ # digits, possibly _ separated
(
(\.(\d+_)*\d+)? # optional fractional part
e[+\-]?(\d+_)*\d+ # exponent part
|
\.(\d+_)*\d+ # required fractional part
)
""",
re.IGNORECASE | re.VERBOSE,
)
# internal the tokens and keep references to them
TOKEN_ADD = intern("add")
TOKEN_ASSIGN = intern("assign")
TOKEN_COLON = intern("colon")
TOKEN_COMMA = intern("comma")
TOKEN_DIV = intern("div")
TOKEN_DOT = intern("dot")
TOKEN_EQ = intern("eq")
TOKEN_FLOORDIV = intern("floordiv")
TOKEN_GT = intern("gt")
TOKEN_GTEQ = intern("gteq")
TOKEN_LBRACE = intern("lbrace")
TOKEN_LBRACKET = intern("lbracket")
TOKEN_LPAREN = intern("lparen")
TOKEN_LT = intern("lt")
TOKEN_LTEQ = intern("lteq")
TOKEN_MOD = intern("mod")
TOKEN_MUL = intern("mul")
TOKEN_NE = intern("ne")
TOKEN_PIPE = intern("pipe")
TOKEN_POW = intern("pow")
TOKEN_RBRACE = intern("rbrace")
TOKEN_RBRACKET = intern("rbracket")
TOKEN_RPAREN = intern("rparen")
TOKEN_SEMICOLON = intern("semicolon")
TOKEN_SUB = intern("sub")
TOKEN_TILDE = intern("tilde")
TOKEN_WHITESPACE = intern("whitespace")
TOKEN_FLOAT = intern("float")
TOKEN_INTEGER = intern("integer")
TOKEN_NAME = intern("name")
TOKEN_STRING = intern("string")
TOKEN_OPERATOR = intern("operator")
TOKEN_BLOCK_BEGIN = intern("block_begin")
TOKEN_BLOCK_END = intern("block_end")
TOKEN_VARIABLE_BEGIN = intern("variable_begin")
TOKEN_VARIABLE_END = intern("variable_end")
TOKEN_RAW_BEGIN = intern("raw_begin")
TOKEN_RAW_END = intern("raw_end")
TOKEN_COMMENT_BEGIN = intern("comment_begin")
TOKEN_COMMENT_END = intern("comment_end")
TOKEN_COMMENT = intern("comment")
TOKEN_LINESTATEMENT_BEGIN = intern("linestatement_begin")
TOKEN_LINESTATEMENT_END = intern("linestatement_end")
TOKEN_LINECOMMENT_BEGIN = intern("linecomment_begin")
TOKEN_LINECOMMENT_END = intern("linecomment_end")
TOKEN_LINECOMMENT = intern("linecomment")
TOKEN_DATA = intern("data")
TOKEN_INITIAL = intern("initial")
TOKEN_EOF = intern("eof")
# bind operators to token types
operators = {
"+": TOKEN_ADD,
"-": TOKEN_SUB,
"/": TOKEN_DIV,
"//": TOKEN_FLOORDIV,
"*": TOKEN_MUL,
"%": TOKEN_MOD,
"**": TOKEN_POW,
"~": TOKEN_TILDE,
"[": TOKEN_LBRACKET,
"]": TOKEN_RBRACKET,
"(": TOKEN_LPAREN,
")": TOKEN_RPAREN,
"{": TOKEN_LBRACE,
"}": TOKEN_RBRACE,
"==": TOKEN_EQ,
"!=": TOKEN_NE,
">": TOKEN_GT,
">=": TOKEN_GTEQ,
"<": TOKEN_LT,
"<=": TOKEN_LTEQ,
"=": TOKEN_ASSIGN,
".": TOKEN_DOT,
":": TOKEN_COLON,
"|": TOKEN_PIPE,
",": TOKEN_COMMA,
";": TOKEN_SEMICOLON,
}
reverse_operators = {v: k for k, v in operators.items()}
assert len(operators) == len(reverse_operators), "operators dropped"
operator_re = re.compile(
f"({'|'.join(re.escape(x) for x in sorted(operators, key=lambda x: -len(x)))})"
)
ignored_tokens = frozenset(
[
TOKEN_COMMENT_BEGIN,
TOKEN_COMMENT,
TOKEN_COMMENT_END,
TOKEN_WHITESPACE,
TOKEN_LINECOMMENT_BEGIN,
TOKEN_LINECOMMENT_END,
TOKEN_LINECOMMENT,
]
)
ignore_if_empty = frozenset(
[TOKEN_WHITESPACE, TOKEN_DATA, TOKEN_COMMENT, TOKEN_LINECOMMENT]
)
def _describe_token_type(token_type: str) -> str:
if token_type in reverse_operators:
return reverse_operators[token_type]
return {
TOKEN_COMMENT_BEGIN: "begin of comment",
TOKEN_COMMENT_END: "end of comment",
TOKEN_COMMENT: "comment",
TOKEN_LINECOMMENT: "comment",
TOKEN_BLOCK_BEGIN: "begin of statement block",
TOKEN_BLOCK_END: "end of statement block",
TOKEN_VARIABLE_BEGIN: "begin of print statement",
TOKEN_VARIABLE_END: "end of print statement",
TOKEN_LINESTATEMENT_BEGIN: "begin of line statement",
TOKEN_LINESTATEMENT_END: "end of line statement",
TOKEN_DATA: "template data / text",
TOKEN_EOF: "end of template",
}.get(token_type, token_type)
def describe_token(token: "Token") -> str:
"""Returns a description of the token."""
if token.type == TOKEN_NAME:
return token.value
return _describe_token_type(token.type)
def describe_token_expr(expr: str) -> str:
"""Like `describe_token` but for token expressions."""
if ":" in expr:
type, value = expr.split(":", 1)
if type == TOKEN_NAME:
return value
else:
type = expr
return _describe_token_type(type)
def count_newlines(value: str) -> int:
"""Count the number of newline characters in the string. This is
useful for extensions that filter a stream.
"""
return len(newline_re.findall(value))
def compile_rules(environment: "Environment") -> t.List[t.Tuple[str, str]]:
"""Compiles all the rules from the environment into a list of rules."""
e = re.escape
rules = [
(
len(environment.comment_start_string),
TOKEN_COMMENT_BEGIN,
e(environment.comment_start_string),
),
(
len(environment.block_start_string),
TOKEN_BLOCK_BEGIN,
e(environment.block_start_string),
),
(
len(environment.variable_start_string),
TOKEN_VARIABLE_BEGIN,
e(environment.variable_start_string),
),
]
if environment.line_statement_prefix is not None:
rules.append(
(
len(environment.line_statement_prefix),
TOKEN_LINESTATEMENT_BEGIN,
r"^[ \t\v]*" + e(environment.line_statement_prefix),
)
)
if environment.line_comment_prefix is not None:
rules.append(
(
len(environment.line_comment_prefix),
TOKEN_LINECOMMENT_BEGIN,
r"(?:^|(?<=\S))[^\S\r\n]*" + e(environment.line_comment_prefix),
)
)
return [x[1:] for x in sorted(rules, reverse=True)]
class Failure:
"""Class that raises a `TemplateSyntaxError` if called.
Used by the `Lexer` to specify known errors.
"""
def __init__(
self, message: str, cls: t.Type[TemplateSyntaxError] = TemplateSyntaxError
) -> None:
self.message = message
self.error_class = cls
def __call__(self, lineno: int, filename: str) -> "te.NoReturn":
raise self.error_class(self.message, lineno, filename)
class Token(t.NamedTuple):
lineno: int
type: str
value: str
def __str__(self) -> str:
return describe_token(self)
def test(self, expr: str) -> bool:
"""Test a token against a token expression. This can either be a
token type or ``'token_type:token_value'``. This can only test
against string values and types.
"""
# here we do a regular string equality check as test_any is usually
# passed an iterable of not interned strings.
if self.type == expr:
return True
if ":" in expr:
return expr.split(":", 1) == [self.type, self.value]
return False
def test_any(self, *iterable: str) -> bool:
"""Test against multiple token expressions."""
return any(self.test(expr) for expr in iterable)
class TokenStreamIterator:
"""The iterator for tokenstreams. Iterate over the stream
until the eof token is reached.
"""
def __init__(self, stream: "TokenStream") -> None:
self.stream = stream
def __iter__(self) -> "TokenStreamIterator":
return self
def __next__(self) -> Token:
token = self.stream.current
if token.type is TOKEN_EOF:
self.stream.close()
raise StopIteration
next(self.stream)
return token
class TokenStream:
"""A token stream is an iterable that yields :class:`Token`\\s. The
parser however does not iterate over it but calls :meth:`next` to go
one token ahead. The current active token is stored as :attr:`current`.
"""
def __init__(
self,
generator: t.Iterable[Token],
name: t.Optional[str],
filename: t.Optional[str],
):
self._iter = iter(generator)
self._pushed: "te.Deque[Token]" = deque()
self.name = name
self.filename = filename
self.closed = False
self.current = Token(1, TOKEN_INITIAL, "")
next(self)
def __iter__(self) -> TokenStreamIterator:
return TokenStreamIterator(self)
def __bool__(self) -> bool:
return bool(self._pushed) or self.current.type is not TOKEN_EOF
@property
def eos(self) -> bool:
"""Are we at the end of the stream?"""
return not self
def push(self, token: Token) -> None:
"""Push a token back to the stream."""
self._pushed.append(token)
def look(self) -> Token:
"""Look at the next token."""
old_token = next(self)
result = self.current
self.push(result)
self.current = old_token
return result
def skip(self, n: int = 1) -> None:
"""Got n tokens ahead."""
for _ in range(n):
next(self)
def next_if(self, expr: str) -> t.Optional[Token]:
"""Perform the token test and return the token if it matched.
Otherwise the return value is `None`.
"""
if self.current.test(expr):
return next(self)
return None
def skip_if(self, expr: str) -> bool:
"""Like :meth:`next_if` but only returns `True` or `False`."""
return self.next_if(expr) is not None
def __next__(self) -> Token:
"""Go one token ahead and return the old one.
Use the built-in :func:`next` instead of calling this directly.
"""
rv = self.current
if self._pushed:
self.current = self._pushed.popleft()
elif self.current.type is not TOKEN_EOF:
try:
self.current = next(self._iter)
except StopIteration:
self.close()
return rv
def close(self) -> None:
"""Close the stream."""
self.current = Token(self.current.lineno, TOKEN_EOF, "")
self._iter = iter(())
self.closed = True
def expect(self, expr: str) -> Token:
"""Expect a given token type and return it. This accepts the same
argument as :meth:`jinja2.lexer.Token.test`.
"""
if not self.current.test(expr):
expr = describe_token_expr(expr)
if self.current.type is TOKEN_EOF:
raise TemplateSyntaxError(
f"unexpected end of template, expected {expr!r}.",
self.current.lineno,
self.name,
self.filename,
)
raise TemplateSyntaxError(
f"expected token {expr!r}, got {describe_token(self.current)!r}",
self.current.lineno,
self.name,
self.filename,
)
return next(self)
def get_lexer(environment: "Environment") -> "Lexer":
"""Return a lexer which is probably cached."""
key = (
environment.block_start_string,
environment.block_end_string,
environment.variable_start_string,
environment.variable_end_string,
environment.comment_start_string,
environment.comment_end_string,
environment.line_statement_prefix,
environment.line_comment_prefix,
environment.trim_blocks,
environment.lstrip_blocks,
environment.newline_sequence,
environment.keep_trailing_newline,
)
lexer = _lexer_cache.get(key)
if lexer is None:
_lexer_cache[key] = lexer = Lexer(environment)
return lexer
class OptionalLStrip(tuple): # type: ignore[type-arg]
"""A special tuple for marking a point in the state that can have
lstrip applied.
"""
__slots__ = ()
# Even though it looks like a no-op, creating instances fails
# without this.
def __new__(cls, *members, **kwargs): # type: ignore
return super().__new__(cls, members)
class _Rule(t.NamedTuple):
pattern: t.Pattern[str]
tokens: t.Union[str, t.Tuple[str, ...], t.Tuple[Failure]]
command: t.Optional[str]
class Lexer:
"""Class that implements a lexer for a given environment. Automatically
created by the environment class, usually you don't have to do that.
Note that the lexer is not automatically bound to an environment.
Multiple environments can share the same lexer.
"""
def __init__(self, environment: "Environment") -> None:
# shortcuts
e = re.escape
def c(x: str) -> t.Pattern[str]:
return re.compile(x, re.M | re.S)
# lexing rules for tags
tag_rules: t.List[_Rule] = [
_Rule(whitespace_re, TOKEN_WHITESPACE, None),
_Rule(float_re, TOKEN_FLOAT, None),
_Rule(integer_re, TOKEN_INTEGER, None),
_Rule(name_re, TOKEN_NAME, None),
_Rule(string_re, TOKEN_STRING, None),
_Rule(operator_re, TOKEN_OPERATOR, None),
]
# assemble the root lexing rule. because "|" is ungreedy
# we have to sort by length so that the lexer continues working
# as expected when we have parsing rules like <% for block and
# <%= for variables. (if someone wants asp like syntax)
# variables are just part of the rules if variable processing
# is required.
root_tag_rules = compile_rules(environment)
block_start_re = e(environment.block_start_string)
block_end_re = e(environment.block_end_string)
comment_end_re = e(environment.comment_end_string)
variable_end_re = e(environment.variable_end_string)
# block suffix if trimming is enabled
block_suffix_re = "\\n?" if environment.trim_blocks else ""
self.lstrip_blocks = environment.lstrip_blocks
self.newline_sequence = environment.newline_sequence
self.keep_trailing_newline = environment.keep_trailing_newline
root_raw_re = (
rf"(?P<raw_begin>{block_start_re}(\-|\+|)\s*raw\s*"
rf"(?:\-{block_end_re}\s*|{block_end_re}))"
)
root_parts_re = "|".join(
[root_raw_re] + [rf"(?P<{n}>{r}(\-|\+|))" for n, r in root_tag_rules]
)
# global lexing rules
self.rules: t.Dict[str, t.List[_Rule]] = {
"root": [
# directives
_Rule(
c(rf"(.*?)(?:{root_parts_re})"),
OptionalLStrip(TOKEN_DATA, "#bygroup"), # type: ignore
"#bygroup",
),
# data
_Rule(c(".+"), TOKEN_DATA, None),
],
# comments
TOKEN_COMMENT_BEGIN: [
_Rule(
c(
rf"(.*?)((?:\+{comment_end_re}|\-{comment_end_re}\s*"
rf"|{comment_end_re}{block_suffix_re}))"
),
(TOKEN_COMMENT, TOKEN_COMMENT_END),
"#pop",
),
_Rule(c(r"(.)"), (Failure("Missing end of comment tag"),), None),
],
# blocks
TOKEN_BLOCK_BEGIN: [
_Rule(
c(
rf"(?:\+{block_end_re}|\-{block_end_re}\s*"
rf"|{block_end_re}{block_suffix_re})"
),
TOKEN_BLOCK_END,
"#pop",
),
]
+ tag_rules,
# variables
TOKEN_VARIABLE_BEGIN: [
_Rule(
c(rf"\-{variable_end_re}\s*|{variable_end_re}"),
TOKEN_VARIABLE_END,
"#pop",
)
]
+ tag_rules,
# raw block
TOKEN_RAW_BEGIN: [
_Rule(
c(
rf"(.*?)((?:{block_start_re}(\-|\+|))\s*endraw\s*"
rf"(?:\+{block_end_re}|\-{block_end_re}\s*"
rf"|{block_end_re}{block_suffix_re}))"
),
OptionalLStrip(TOKEN_DATA, TOKEN_RAW_END), # type: ignore
"#pop",
),
_Rule(c(r"(.)"), (Failure("Missing end of raw directive"),), None),
],
# line statements
TOKEN_LINESTATEMENT_BEGIN: [
_Rule(c(r"\s*(\n|$)"), TOKEN_LINESTATEMENT_END, "#pop")
]
+ tag_rules,
# line comments
TOKEN_LINECOMMENT_BEGIN: [
_Rule(
c(r"(.*?)()(?=\n|$)"),
(TOKEN_LINECOMMENT, TOKEN_LINECOMMENT_END),
"#pop",
)
],
}
def _normalize_newlines(self, value: str) -> str:
"""Replace all newlines with the configured sequence in strings
and template data.
"""
return newline_re.sub(self.newline_sequence, value)
def tokenize(
self,
source: str,
name: t.Optional[str] = None,
filename: t.Optional[str] = None,
state: t.Optional[str] = None,
) -> TokenStream:
"""Calls tokeniter + tokenize and wraps it in a token stream."""
stream = self.tokeniter(source, name, filename, state)
return TokenStream(self.wrap(stream, name, filename), name, filename)
def wrap(
self,
stream: t.Iterable[t.Tuple[int, str, str]],
name: t.Optional[str] = None,
filename: t.Optional[str] = None,
) -> t.Iterator[Token]:
"""This is called with the stream as returned by `tokenize` and wraps
every token in a :class:`Token` and converts the value.
"""
for lineno, token, value_str in stream:
if token in ignored_tokens:
continue
value: t.Any = value_str
if token == TOKEN_LINESTATEMENT_BEGIN:
token = TOKEN_BLOCK_BEGIN
elif token == TOKEN_LINESTATEMENT_END:
token = TOKEN_BLOCK_END
# we are not interested in those tokens in the parser
elif token in (TOKEN_RAW_BEGIN, TOKEN_RAW_END):
continue
elif token == TOKEN_DATA:
value = self._normalize_newlines(value_str)
elif token == "keyword":
token = value_str
elif token == TOKEN_NAME:
value = value_str
if not value.isidentifier():
raise TemplateSyntaxError(
"Invalid character in identifier", lineno, name, filename
)
elif token == TOKEN_STRING:
# try to unescape string
try:
value = (
self._normalize_newlines(value_str[1:-1])
.encode("ascii", "backslashreplace")
.decode("unicode-escape")
)
except Exception as e:
msg = str(e).split(":")[-1].strip()
raise TemplateSyntaxError(msg, lineno, name, filename) from e
elif token == TOKEN_INTEGER:
value = int(value_str.replace("_", ""), 0)
elif token == TOKEN_FLOAT:
# remove all "_" first to support more Python versions
value = literal_eval(value_str.replace("_", ""))
elif token == TOKEN_OPERATOR:
token = operators[value_str]
yield Token(lineno, token, value)
def tokeniter(
self,
source: str,
name: t.Optional[str],
filename: t.Optional[str] = None,
state: t.Optional[str] = None,
) -> t.Iterator[t.Tuple[int, str, str]]:
"""This method tokenizes the text and returns the tokens in a
generator. Use this method if you just want to tokenize a template.
.. versionchanged:: 3.0
Only ``\\n``, ``\\r\\n`` and ``\\r`` are treated as line
breaks.
"""
lines = newline_re.split(source)[::2]
if not self.keep_trailing_newline and lines[-1] == "":
del lines[-1]
source = "\n".join(lines)
pos = 0
lineno = 1
stack = ["root"]
if state is not None and state != "root":
assert state in ("variable", "block"), "invalid state"
stack.append(state + "_begin")
statetokens = self.rules[stack[-1]]
source_length = len(source)
balancing_stack: t.List[str] = []
newlines_stripped = 0
line_starting = True
while True:
# tokenizer loop
for regex, tokens, new_state in statetokens:
m = regex.match(source, pos)
# if no match we try again with the next rule
if m is None:
continue
# we only match blocks and variables if braces / parentheses
# are balanced. continue parsing with the lower rule which
# is the operator rule. do this only if the end tags look
# like operators
if balancing_stack and tokens in (
TOKEN_VARIABLE_END,
TOKEN_BLOCK_END,
TOKEN_LINESTATEMENT_END,
):
continue
# tuples support more options
if isinstance(tokens, tuple):
groups: t.Sequence[str] = m.groups()
if isinstance(tokens, OptionalLStrip):
# Rule supports lstrip. Match will look like
# text, block type, whitespace control, type, control, ...
text = groups[0]
# Skipping the text and first type, every other group is the
# whitespace control for each type. One of the groups will be
# -, +, or empty string instead of None.
strip_sign = next(g for g in groups[2::2] if g is not None)
if strip_sign == "-":
# Strip all whitespace between the text and the tag.
stripped = text.rstrip()
newlines_stripped = text[len(stripped) :].count("\n")
groups = [stripped, *groups[1:]]
elif (
# Not marked for preserving whitespace.
strip_sign != "+"
# lstrip is enabled.
and self.lstrip_blocks
# Not a variable expression.
and not m.groupdict().get(TOKEN_VARIABLE_BEGIN)
):
# The start of text between the last newline and the tag.
l_pos = text.rfind("\n") + 1
if l_pos > 0 or line_starting:
# If there's only whitespace between the newline and the
# tag, strip it.
if whitespace_re.fullmatch(text, l_pos):
groups = [text[:l_pos], *groups[1:]]
for idx, token in enumerate(tokens):
# failure group
if token.__class__ is Failure:
raise token(lineno, filename)
# bygroup is a bit more complex, in that case we
# yield for the current token the first named
# group that matched
elif token == "#bygroup":
for key, value in m.groupdict().items():
if value is not None:
yield lineno, key, value
lineno += value.count("\n")
break
else:
raise RuntimeError(
f"{regex!r} wanted to resolve the token dynamically"
" but no group matched"
)
# normal group
else:
data = groups[idx]
if data or token not in ignore_if_empty:
yield lineno, token, data
lineno += data.count("\n") + newlines_stripped
newlines_stripped = 0
# strings as token just are yielded as it.
else:
data = m.group()
# update brace/parentheses balance
if tokens == TOKEN_OPERATOR:
if data == "{":
balancing_stack.append("}")
elif data == "(":
balancing_stack.append(")")
elif data == "[":
balancing_stack.append("]")
elif data in ("}", ")", "]"):
if not balancing_stack:
raise TemplateSyntaxError(
f"unexpected '{data}'", lineno, name, filename
)
expected_op = balancing_stack.pop()
if expected_op != data:
raise TemplateSyntaxError(
f"unexpected '{data}', expected '{expected_op}'",
lineno,
name,
filename,
)
# yield items
if data or tokens not in ignore_if_empty:
yield lineno, tokens, data
lineno += data.count("\n")
line_starting = m.group()[-1:] == "\n"
# fetch new position into new variable so that we can check
# if there is a internal parsing error which would result
# in an infinite loop
pos2 = m.end()
# handle state changes
if new_state is not None:
# remove the uppermost state
if new_state == "#pop":
stack.pop()
# resolve the new state by group checking
elif new_state == "#bygroup":
for key, value in m.groupdict().items():
if value is not None:
stack.append(key)
break
else:
raise RuntimeError(
f"{regex!r} wanted to resolve the new state dynamically"
f" but no group matched"
)
# direct state name given
else:
stack.append(new_state)
statetokens = self.rules[stack[-1]]
# we are still at the same position and no stack change.
# this means a loop without break condition, avoid that and
# raise error
elif pos2 == pos:
raise RuntimeError(
f"{regex!r} yielded empty string without stack change"
)
# publish new function and start again
pos = pos2
break
# if loop terminated without break we haven't found a single match
# either we are at the end of the file or we have a problem
else:
# end of text
if pos >= source_length:
return
# something went wrong
raise TemplateSyntaxError(
f"unexpected char {source[pos]!r} at {pos}", lineno, name, filename
)
"""API and implementations for loading templates from different data
sources.
"""
import importlib.util
import os
import posixpath
import sys
import typing as t
import weakref
import zipimport
from collections import abc
from hashlib import sha1
from importlib import import_module
from types import ModuleType
from .exceptions import TemplateNotFound
from .utils import internalcode
if t.TYPE_CHECKING:
from .environment import Environment
from .environment import Template
def split_template_path(template: str) -> t.List[str]:
"""Split a path into segments and perform a sanity check. If it detects
'..' in the path it will raise a `TemplateNotFound` error.
"""
pieces = []
for piece in template.split("/"):
if (
os.path.sep in piece
or (os.path.altsep and os.path.altsep in piece)
or piece == os.path.pardir
):
raise TemplateNotFound(template)
elif piece and piece != ".":
pieces.append(piece)
return pieces
class BaseLoader:
"""Baseclass for all loaders. Subclass this and override `get_source` to
implement a custom loading mechanism. The environment provides a
`get_template` method that calls the loader's `load` method to get the
:class:`Template` object.
A very basic example for a loader that looks up templates on the file
system could look like this::
from jinja2 import BaseLoader, TemplateNotFound
from os.path import join, exists, getmtime
class MyLoader(BaseLoader):
def __init__(self, path):
self.path = path
def get_source(self, environment, template):
path = join(self.path, template)
if not exists(path):
raise TemplateNotFound(template)
mtime = getmtime(path)
with open(path) as f:
source = f.read()
return source, path, lambda: mtime == getmtime(path)
"""
#: if set to `False` it indicates that the loader cannot provide access
#: to the source of templates.
#:
#: .. versionadded:: 2.4
has_source_access = True
def get_source(
self, environment: "Environment", template: str
) -> t.Tuple[str, t.Optional[str], t.Optional[t.Callable[[], bool]]]:
"""Get the template source, filename and reload helper for a template.
It's passed the environment and template name and has to return a
tuple in the form ``(source, filename, uptodate)`` or raise a
`TemplateNotFound` error if it can't locate the template.
The source part of the returned tuple must be the source of the
template as a string. The filename should be the name of the
file on the filesystem if it was loaded from there, otherwise
``None``. The filename is used by Python for the tracebacks
if no loader extension is used.
The last item in the tuple is the `uptodate` function. If auto
reloading is enabled it's always called to check if the template
changed. No arguments are passed so the function must store the
old state somewhere (for example in a closure). If it returns `False`
the template will be reloaded.
"""
if not self.has_source_access:
raise RuntimeError(
f"{type(self).__name__} cannot provide access to the source"
)
raise TemplateNotFound(template)
def list_templates(self) -> t.List[str]:
"""Iterates over all templates. If the loader does not support that
it should raise a :exc:`TypeError` which is the default behavior.
"""
raise TypeError("this loader cannot iterate over all templates")
@internalcode
def load(
self,
environment: "Environment",
name: str,
globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
) -> "Template":
"""Loads a template. This method looks up the template in the cache
or loads one by calling :meth:`get_source`. Subclasses should not
override this method as loaders working on collections of other
loaders (such as :class:`PrefixLoader` or :class:`ChoiceLoader`)
will not call this method but `get_source` directly.
"""
code = None
if globals is None:
globals = {}
# first we try to get the source for this template together
# with the filename and the uptodate function.
source, filename, uptodate = self.get_source(environment, name)
# try to load the code from the bytecode cache if there is a
# bytecode cache configured.
bcc = environment.bytecode_cache
if bcc is not None:
bucket = bcc.get_bucket(environment, name, filename, source)
code = bucket.code
# if we don't have code so far (not cached, no longer up to
# date) etc. we compile the template
if code is None:
code = environment.compile(source, name, filename)
# if the bytecode cache is available and the bucket doesn't
# have a code so far, we give the bucket the new code and put
# it back to the bytecode cache.
if bcc is not None and bucket.code is None:
bucket.code = code
bcc.set_bucket(bucket)
return environment.template_class.from_code(
environment, code, globals, uptodate
)
class FileSystemLoader(BaseLoader):
"""Load templates from a directory in the file system.
The path can be relative or absolute. Relative paths are relative to
the current working directory.
.. code-block:: python
loader = FileSystemLoader("templates")
A list of paths can be given. The directories will be searched in
order, stopping at the first matching template.
.. code-block:: python
loader = FileSystemLoader(["/override/templates", "/default/templates"])
:param searchpath: A path, or list of paths, to the directory that
contains the templates.
:param encoding: Use this encoding to read the text from template
files.
:param followlinks: Follow symbolic links in the path.
.. versionchanged:: 2.8
Added the ``followlinks`` parameter.
"""
def __init__(
self,
searchpath: t.Union[
str, "os.PathLike[str]", t.Sequence[t.Union[str, "os.PathLike[str]"]]
],
encoding: str = "utf-8",
followlinks: bool = False,
) -> None:
if not isinstance(searchpath, abc.Iterable) or isinstance(searchpath, str):
searchpath = [searchpath]
self.searchpath = [os.fspath(p) for p in searchpath]
self.encoding = encoding
self.followlinks = followlinks
def get_source(
self, environment: "Environment", template: str
) -> t.Tuple[str, str, t.Callable[[], bool]]:
pieces = split_template_path(template)
for searchpath in self.searchpath:
# Use posixpath even on Windows to avoid "drive:" or UNC
# segments breaking out of the search directory.
filename = posixpath.join(searchpath, *pieces)
if os.path.isfile(filename):
break
else:
raise TemplateNotFound(template)
with open(filename, encoding=self.encoding) as f:
contents = f.read()
mtime = os.path.getmtime(filename)
def uptodate() -> bool:
try:
return os.path.getmtime(filename) == mtime
except OSError:
return False
# Use normpath to convert Windows altsep to sep.
return contents, os.path.normpath(filename), uptodate
def list_templates(self) -> t.List[str]:
found = set()
for searchpath in self.searchpath:
walk_dir = os.walk(searchpath, followlinks=self.followlinks)
for dirpath, _, filenames in walk_dir:
for filename in filenames:
template = (
os.path.join(dirpath, filename)[len(searchpath) :]
.strip(os.path.sep)
.replace(os.path.sep, "/")
)
if template[:2] == "./":
template = template[2:]
if template not in found:
found.add(template)
return sorted(found)
class PackageLoader(BaseLoader):
"""Load templates from a directory in a Python package.
:param package_name: Import name of the package that contains the
template directory.
:param package_path: Directory within the imported package that
contains the templates.
:param encoding: Encoding of template files.
The following example looks up templates in the ``pages`` directory
within the ``project.ui`` package.
.. code-block:: python
loader = PackageLoader("project.ui", "pages")
Only packages installed as directories (standard pip behavior) or
zip/egg files (less common) are supported. The Python API for
introspecting data in packages is too limited to support other
installation methods the way this loader requires.
There is limited support for :pep:`420` namespace packages. The
template directory is assumed to only be in one namespace
contributor. Zip files contributing to a namespace are not
supported.
.. versionchanged:: 3.0
No longer uses ``setuptools`` as a dependency.
.. versionchanged:: 3.0
Limited PEP 420 namespace package support.
"""
def __init__(
self,
package_name: str,
package_path: "str" = "templates",
encoding: str = "utf-8",
) -> None:
package_path = os.path.normpath(package_path).rstrip(os.path.sep)
# normpath preserves ".", which isn't valid in zip paths.
if package_path == os.path.curdir:
package_path = ""
elif package_path[:2] == os.path.curdir + os.path.sep:
package_path = package_path[2:]
self.package_path = package_path
self.package_name = package_name
self.encoding = encoding
# Make sure the package exists. This also makes namespace
# packages work, otherwise get_loader returns None.
import_module(package_name)
spec = importlib.util.find_spec(package_name)
assert spec is not None, "An import spec was not found for the package."
loader = spec.loader
assert loader is not None, "A loader was not found for the package."
self._loader = loader
self._archive = None
template_root = None
if isinstance(loader, zipimport.zipimporter):
self._archive = loader.archive
pkgdir = next(iter(spec.submodule_search_locations)) # type: ignore
template_root = os.path.join(pkgdir, package_path).rstrip(os.path.sep)
else:
roots: t.List[str] = []
# One element for regular packages, multiple for namespace
# packages, or None for single module file.
if spec.submodule_search_locations:
roots.extend(spec.submodule_search_locations)
# A single module file, use the parent directory instead.
elif spec.origin is not None:
roots.append(os.path.dirname(spec.origin))
for root in roots:
root = os.path.join(root, package_path)
if os.path.isdir(root):
template_root = root
break
if template_root is None:
raise ValueError(
f"The {package_name!r} package was not installed in a"
" way that PackageLoader understands."
)
self._template_root = template_root
def get_source(
self, environment: "Environment", template: str
) -> t.Tuple[str, str, t.Optional[t.Callable[[], bool]]]:
# Use posixpath even on Windows to avoid "drive:" or UNC
# segments breaking out of the search directory. Use normpath to
# convert Windows altsep to sep.
p = os.path.normpath(
posixpath.join(self._template_root, *split_template_path(template))
)
up_to_date: t.Optional[t.Callable[[], bool]]
if self._archive is None:
# Package is a directory.
if not os.path.isfile(p):
raise TemplateNotFound(template)
with open(p, "rb") as f:
source = f.read()
mtime = os.path.getmtime(p)
def up_to_date() -> bool:
return os.path.isfile(p) and os.path.getmtime(p) == mtime
else:
# Package is a zip file.
try:
source = self._loader.get_data(p) # type: ignore
except OSError as e:
raise TemplateNotFound(template) from e
# Could use the zip's mtime for all template mtimes, but
# would need to safely reload the module if it's out of
# date, so just report it as always current.
up_to_date = None
return source.decode(self.encoding), p, up_to_date
def list_templates(self) -> t.List[str]:
results: t.List[str] = []
if self._archive is None:
# Package is a directory.
offset = len(self._template_root)
for dirpath, _, filenames in os.walk(self._template_root):
dirpath = dirpath[offset:].lstrip(os.path.sep)
results.extend(
os.path.join(dirpath, name).replace(os.path.sep, "/")
for name in filenames
)
else:
if not hasattr(self._loader, "_files"):
raise TypeError(
"This zip import does not have the required"
" metadata to list templates."
)
# Package is a zip file.
prefix = (
self._template_root[len(self._archive) :].lstrip(os.path.sep)
+ os.path.sep
)
offset = len(prefix)
for name in self._loader._files.keys():
# Find names under the templates directory that aren't directories.
if name.startswith(prefix) and name[-1] != os.path.sep:
results.append(name[offset:].replace(os.path.sep, "/"))
results.sort()
return results
class DictLoader(BaseLoader):
"""Loads a template from a Python dict mapping template names to
template source. This loader is useful for unittesting:
>>> loader = DictLoader({'index.html': 'source here'})
Because auto reloading is rarely useful this is disabled per default.
"""
def __init__(self, mapping: t.Mapping[str, str]) -> None:
self.mapping = mapping
def get_source(
self, environment: "Environment", template: str
) -> t.Tuple[str, None, t.Callable[[], bool]]:
if template in self.mapping:
source = self.mapping[template]
return source, None, lambda: source == self.mapping.get(template)
raise TemplateNotFound(template)
def list_templates(self) -> t.List[str]:
return sorted(self.mapping)
class FunctionLoader(BaseLoader):
"""A loader that is passed a function which does the loading. The
function receives the name of the template and has to return either
a string with the template source, a tuple in the form ``(source,
filename, uptodatefunc)`` or `None` if the template does not exist.
>>> def load_template(name):
... if name == 'index.html':
... return '...'
...
>>> loader = FunctionLoader(load_template)
The `uptodatefunc` is a function that is called if autoreload is enabled
and has to return `True` if the template is still up to date. For more
details have a look at :meth:`BaseLoader.get_source` which has the same
return value.
"""
def __init__(
self,
load_func: t.Callable[
[str],
t.Optional[
t.Union[
str, t.Tuple[str, t.Optional[str], t.Optional[t.Callable[[], bool]]]
]
],
],
) -> None:
self.load_func = load_func
def get_source(
self, environment: "Environment", template: str
) -> t.Tuple[str, t.Optional[str], t.Optional[t.Callable[[], bool]]]:
rv = self.load_func(template)
if rv is None:
raise TemplateNotFound(template)
if isinstance(rv, str):
return rv, None, None
return rv
class PrefixLoader(BaseLoader):
"""A loader that is passed a dict of loaders where each loader is bound
to a prefix. The prefix is delimited from the template by a slash per
default, which can be changed by setting the `delimiter` argument to
something else::
loader = PrefixLoader({
'app1': PackageLoader('mypackage.app1'),
'app2': PackageLoader('mypackage.app2')
})
By loading ``'app1/index.html'`` the file from the app1 package is loaded,
by loading ``'app2/index.html'`` the file from the second.
"""
def __init__(
self, mapping: t.Mapping[str, BaseLoader], delimiter: str = "/"
) -> None:
self.mapping = mapping
self.delimiter = delimiter
def get_loader(self, template: str) -> t.Tuple[BaseLoader, str]:
try:
prefix, name = template.split(self.delimiter, 1)
loader = self.mapping[prefix]
except (ValueError, KeyError) as e:
raise TemplateNotFound(template) from e
return loader, name
def get_source(
self, environment: "Environment", template: str
) -> t.Tuple[str, t.Optional[str], t.Optional[t.Callable[[], bool]]]:
loader, name = self.get_loader(template)
try:
return loader.get_source(environment, name)
except TemplateNotFound as e:
# re-raise the exception with the correct filename here.
# (the one that includes the prefix)
raise TemplateNotFound(template) from e
@internalcode
def load(
self,
environment: "Environment",
name: str,
globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
) -> "Template":
loader, local_name = self.get_loader(name)
try:
return loader.load(environment, local_name, globals)
except TemplateNotFound as e:
# re-raise the exception with the correct filename here.
# (the one that includes the prefix)
raise TemplateNotFound(name) from e
def list_templates(self) -> t.List[str]:
result = []
for prefix, loader in self.mapping.items():
for template in loader.list_templates():
result.append(prefix + self.delimiter + template)
return result
class ChoiceLoader(BaseLoader):
"""This loader works like the `PrefixLoader` just that no prefix is
specified. If a template could not be found by one loader the next one
is tried.
>>> loader = ChoiceLoader([
... FileSystemLoader('/path/to/user/templates'),
... FileSystemLoader('/path/to/system/templates')
... ])
This is useful if you want to allow users to override builtin templates
from a different location.
"""
def __init__(self, loaders: t.Sequence[BaseLoader]) -> None:
self.loaders = loaders
def get_source(
self, environment: "Environment", template: str
) -> t.Tuple[str, t.Optional[str], t.Optional[t.Callable[[], bool]]]:
for loader in self.loaders:
try:
return loader.get_source(environment, template)
except TemplateNotFound:
pass
raise TemplateNotFound(template)
@internalcode
def load(
self,
environment: "Environment",
name: str,
globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
) -> "Template":
for loader in self.loaders:
try:
return loader.load(environment, name, globals)
except TemplateNotFound:
pass
raise TemplateNotFound(name)
def list_templates(self) -> t.List[str]:
found = set()
for loader in self.loaders:
found.update(loader.list_templates())
return sorted(found)
class _TemplateModule(ModuleType):
"""Like a normal module but with support for weak references"""
class ModuleLoader(BaseLoader):
"""This loader loads templates from precompiled templates.
Example usage:
>>> loader = ChoiceLoader([
... ModuleLoader('/path/to/compiled/templates'),
... FileSystemLoader('/path/to/templates')
... ])
Templates can be precompiled with :meth:`Environment.compile_templates`.
"""
has_source_access = False
def __init__(
self,
path: t.Union[
str, "os.PathLike[str]", t.Sequence[t.Union[str, "os.PathLike[str]"]]
],
) -> None:
package_name = f"_jinja2_module_templates_{id(self):x}"
# create a fake module that looks for the templates in the
# path given.
mod = _TemplateModule(package_name)
if not isinstance(path, abc.Iterable) or isinstance(path, str):
path = [path]
mod.__path__ = [os.fspath(p) for p in path]
sys.modules[package_name] = weakref.proxy(
mod, lambda x: sys.modules.pop(package_name, None)
)
# the only strong reference, the sys.modules entry is weak
# so that the garbage collector can remove it once the
# loader that created it goes out of business.
self.module = mod
self.package_name = package_name
@staticmethod
def get_template_key(name: str) -> str:
return "tmpl_" + sha1(name.encode("utf-8")).hexdigest()
@staticmethod
def get_module_filename(name: str) -> str:
return ModuleLoader.get_template_key(name) + ".py"
@internalcode
def load(
self,
environment: "Environment",
name: str,
globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
) -> "Template":
key = self.get_template_key(name)
module = f"{self.package_name}.{key}"
mod = getattr(self.module, module, None)
if mod is None:
try:
mod = __import__(module, None, None, ["root"])
except ImportError as e:
raise TemplateNotFound(name) from e
# remove the entry from sys.modules, we only want the attribute
# on the module object we have stored on the loader.
sys.modules.pop(module, None)
if globals is None:
globals = {}
return environment.template_class.from_module_dict(
environment, mod.__dict__, globals
)
"""Functions that expose information about templates that might be
interesting for introspection.
"""
import typing as t
from . import nodes
from .compiler import CodeGenerator
from .compiler import Frame
if t.TYPE_CHECKING:
from .environment import Environment
class TrackingCodeGenerator(CodeGenerator):
"""We abuse the code generator for introspection."""
def __init__(self, environment: "Environment") -> None:
super().__init__(environment, "<introspection>", "<introspection>")
self.undeclared_identifiers: t.Set[str] = set()
def write(self, x: str) -> None:
"""Don't write."""
def enter_frame(self, frame: Frame) -> None:
"""Remember all undeclared identifiers."""
super().enter_frame(frame)
for _, (action, param) in frame.symbols.loads.items():
if action == "resolve" and param not in self.environment.globals:
self.undeclared_identifiers.add(param)
def find_undeclared_variables(ast: nodes.Template) -> t.Set[str]:
"""Returns a set of all variables in the AST that will be looked up from
the context at runtime. Because at compile time it's not known which
variables will be used depending on the path the execution takes at
runtime, all variables are returned.
>>> from jinja2 import Environment, meta
>>> env = Environment()
>>> ast = env.parse('{% set foo = 42 %}{{ bar + foo }}')
>>> meta.find_undeclared_variables(ast) == {'bar'}
True
.. admonition:: Implementation
Internally the code generator is used for finding undeclared variables.
This is good to know because the code generator might raise a
:exc:`TemplateAssertionError` during compilation and as a matter of
fact this function can currently raise that exception as well.
"""
codegen = TrackingCodeGenerator(ast.environment) # type: ignore
codegen.visit(ast)
return codegen.undeclared_identifiers
_ref_types = (nodes.Extends, nodes.FromImport, nodes.Import, nodes.Include)
_RefType = t.Union[nodes.Extends, nodes.FromImport, nodes.Import, nodes.Include]
def find_referenced_templates(ast: nodes.Template) -> t.Iterator[t.Optional[str]]:
"""Finds all the referenced templates from the AST. This will return an
iterator over all the hardcoded template extensions, inclusions and
imports. If dynamic inheritance or inclusion is used, `None` will be
yielded.
>>> from jinja2 import Environment, meta
>>> env = Environment()
>>> ast = env.parse('{% extends "layout.html" %}{% include helper %}')
>>> list(meta.find_referenced_templates(ast))
['layout.html', None]
This function is useful for dependency tracking. For example if you want
to rebuild parts of the website after a layout template has changed.
"""
template_name: t.Any
for node in ast.find_all(_ref_types):
template: nodes.Expr = node.template # type: ignore
if not isinstance(template, nodes.Const):
# a tuple with some non consts in there
if isinstance(template, (nodes.Tuple, nodes.List)):
for template_name in template.items:
# something const, only yield the strings and ignore
# non-string consts that really just make no sense
if isinstance(template_name, nodes.Const):
if isinstance(template_name.value, str):
yield template_name.value
# something dynamic in there
else:
yield None
# something dynamic we don't know about here
else:
yield None
continue
# constant is a basestring, direct template name
if isinstance(template.value, str):
yield template.value
# a tuple or list (latter *should* not happen) made of consts,
# yield the consts that are strings. We could warn here for
# non string values
elif isinstance(node, nodes.Include) and isinstance(
template.value, (tuple, list)
):
for template_name in template.value:
if isinstance(template_name, str):
yield template_name
# something else we don't care about, we could warn here
else:
yield None
import typing as t
from ast import literal_eval
from ast import parse
from itertools import chain
from itertools import islice
from types import GeneratorType
from . import nodes
from .compiler import CodeGenerator
from .compiler import Frame
from .compiler import has_safe_repr
from .environment import Environment
from .environment import Template
def native_concat(values: t.Iterable[t.Any]) -> t.Optional[t.Any]:
"""Return a native Python type from the list of compiled nodes. If
the result is a single node, its value is returned. Otherwise, the
nodes are concatenated as strings. If the result can be parsed with
:func:`ast.literal_eval`, the parsed value is returned. Otherwise,
the string is returned.
:param values: Iterable of outputs to concatenate.
"""
head = list(islice(values, 2))
if not head:
return None
if len(head) == 1:
raw = head[0]
if not isinstance(raw, str):
return raw
else:
if isinstance(values, GeneratorType):
values = chain(head, values)
raw = "".join([str(v) for v in values])
try:
return literal_eval(
# In Python 3.10+ ast.literal_eval removes leading spaces/tabs
# from the given string. For backwards compatibility we need to
# parse the string ourselves without removing leading spaces/tabs.
parse(raw, mode="eval")
)
except (ValueError, SyntaxError, MemoryError):
return raw
class NativeCodeGenerator(CodeGenerator):
"""A code generator which renders Python types by not adding
``str()`` around output nodes.
"""
@staticmethod
def _default_finalize(value: t.Any) -> t.Any:
return value
def _output_const_repr(self, group: t.Iterable[t.Any]) -> str:
return repr("".join([str(v) for v in group]))
def _output_child_to_const(
self, node: nodes.Expr, frame: Frame, finalize: CodeGenerator._FinalizeInfo
) -> t.Any:
const = node.as_const(frame.eval_ctx)
if not has_safe_repr(const):
raise nodes.Impossible()
if isinstance(node, nodes.TemplateData):
return const
return finalize.const(const) # type: ignore
def _output_child_pre(
self, node: nodes.Expr, frame: Frame, finalize: CodeGenerator._FinalizeInfo
) -> None:
if finalize.src is not None:
self.write(finalize.src)
def _output_child_post(
self, node: nodes.Expr, frame: Frame, finalize: CodeGenerator._FinalizeInfo
) -> None:
if finalize.src is not None:
self.write(")")
class NativeEnvironment(Environment):
"""An environment that renders templates to native Python types."""
code_generator_class = NativeCodeGenerator
concat = staticmethod(native_concat) # type: ignore
class NativeTemplate(Template):
environment_class = NativeEnvironment
def render(self, *args: t.Any, **kwargs: t.Any) -> t.Any:
"""Render the template to produce a native Python type. If the
result is a single node, its value is returned. Otherwise, the
nodes are concatenated as strings. If the result can be parsed
with :func:`ast.literal_eval`, the parsed value is returned.
Otherwise, the string is returned.
"""
ctx = self.new_context(dict(*args, **kwargs))
try:
return self.environment_class.concat( # type: ignore
self.root_render_func(ctx)
)
except Exception:
return self.environment.handle_exception()
async def render_async(self, *args: t.Any, **kwargs: t.Any) -> t.Any:
if not self.environment.is_async:
raise RuntimeError(
"The environment was not created with async mode enabled."
)
ctx = self.new_context(dict(*args, **kwargs))
try:
return self.environment_class.concat( # type: ignore
[n async for n in self.root_render_func(ctx)] # type: ignore
)
except Exception:
return self.environment.handle_exception()
NativeEnvironment.template_class = NativeTemplate
"""AST nodes generated by the parser for the compiler. Also provides
some node tree helper functions used by the parser and compiler in order
to normalize nodes.
"""
import inspect
import operator
import typing as t
from collections import deque
from markupsafe import Markup
from .utils import _PassArg
if t.TYPE_CHECKING:
import typing_extensions as te
from .environment import Environment
_NodeBound = t.TypeVar("_NodeBound", bound="Node")
_binop_to_func: t.Dict[str, t.Callable[[t.Any, t.Any], t.Any]] = {
"*": operator.mul,
"/": operator.truediv,
"//": operator.floordiv,
"**": operator.pow,
"%": operator.mod,
"+": operator.add,
"-": operator.sub,
}
_uaop_to_func: t.Dict[str, t.Callable[[t.Any], t.Any]] = {
"not": operator.not_,
"+": operator.pos,
"-": operator.neg,
}
_cmpop_to_func: t.Dict[str, t.Callable[[t.Any, t.Any], t.Any]] = {
"eq": operator.eq,
"ne": operator.ne,
"gt": operator.gt,
"gteq": operator.ge,
"lt": operator.lt,
"lteq": operator.le,
"in": lambda a, b: a in b,
"notin": lambda a, b: a not in b,
}
class Impossible(Exception):
"""Raised if the node could not perform a requested action."""
class NodeType(type):
"""A metaclass for nodes that handles the field and attribute
inheritance. fields and attributes from the parent class are
automatically forwarded to the child."""
def __new__(mcs, name, bases, d): # type: ignore
for attr in "fields", "attributes":
storage: t.List[t.Tuple[str, ...]] = []
storage.extend(getattr(bases[0] if bases else object, attr, ()))
storage.extend(d.get(attr, ()))
assert len(bases) <= 1, "multiple inheritance not allowed"
assert len(storage) == len(set(storage)), "layout conflict"
d[attr] = tuple(storage)
d.setdefault("abstract", False)
return type.__new__(mcs, name, bases, d)
class EvalContext:
"""Holds evaluation time information. Custom attributes can be attached
to it in extensions.
"""
def __init__(
self, environment: "Environment", template_name: t.Optional[str] = None
) -> None:
self.environment = environment
if callable(environment.autoescape):
self.autoescape = environment.autoescape(template_name)
else:
self.autoescape = environment.autoescape
self.volatile = False
def save(self) -> t.Mapping[str, t.Any]:
return self.__dict__.copy()
def revert(self, old: t.Mapping[str, t.Any]) -> None:
self.__dict__.clear()
self.__dict__.update(old)
def get_eval_context(node: "Node", ctx: t.Optional[EvalContext]) -> EvalContext:
if ctx is None:
if node.environment is None:
raise RuntimeError(
"if no eval context is passed, the node must have an"
" attached environment."
)
return EvalContext(node.environment)
return ctx
class Node(metaclass=NodeType):
"""Baseclass for all Jinja nodes. There are a number of nodes available
of different types. There are four major types:
- :class:`Stmt`: statements
- :class:`Expr`: expressions
- :class:`Helper`: helper nodes
- :class:`Template`: the outermost wrapper node
All nodes have fields and attributes. Fields may be other nodes, lists,
or arbitrary values. Fields are passed to the constructor as regular
positional arguments, attributes as keyword arguments. Each node has
two attributes: `lineno` (the line number of the node) and `environment`.
The `environment` attribute is set at the end of the parsing process for
all nodes automatically.
"""
fields: t.Tuple[str, ...] = ()
attributes: t.Tuple[str, ...] = ("lineno", "environment")
abstract = True
lineno: int
environment: t.Optional["Environment"]
def __init__(self, *fields: t.Any, **attributes: t.Any) -> None:
if self.abstract:
raise TypeError("abstract nodes are not instantiable")
if fields:
if len(fields) != len(self.fields):
if not self.fields:
raise TypeError(f"{type(self).__name__!r} takes 0 arguments")
raise TypeError(
f"{type(self).__name__!r} takes 0 or {len(self.fields)}"
f" argument{'s' if len(self.fields) != 1 else ''}"
)
for name, arg in zip(self.fields, fields):
setattr(self, name, arg)
for attr in self.attributes:
setattr(self, attr, attributes.pop(attr, None))
if attributes:
raise TypeError(f"unknown attribute {next(iter(attributes))!r}")
def iter_fields(
self,
exclude: t.Optional[t.Container[str]] = None,
only: t.Optional[t.Container[str]] = None,
) -> t.Iterator[t.Tuple[str, t.Any]]:
"""This method iterates over all fields that are defined and yields
``(key, value)`` tuples. Per default all fields are returned, but
it's possible to limit that to some fields by providing the `only`
parameter or to exclude some using the `exclude` parameter. Both
should be sets or tuples of field names.
"""
for name in self.fields:
if (
(exclude is None and only is None)
or (exclude is not None and name not in exclude)
or (only is not None and name in only)
):
try:
yield name, getattr(self, name)
except AttributeError:
pass
def iter_child_nodes(
self,
exclude: t.Optional[t.Container[str]] = None,
only: t.Optional[t.Container[str]] = None,
) -> t.Iterator["Node"]:
"""Iterates over all direct child nodes of the node. This iterates
over all fields and yields the values of they are nodes. If the value
of a field is a list all the nodes in that list are returned.
"""
for _, item in self.iter_fields(exclude, only):
if isinstance(item, list):
for n in item:
if isinstance(n, Node):
yield n
elif isinstance(item, Node):
yield item
def find(self, node_type: t.Type[_NodeBound]) -> t.Optional[_NodeBound]:
"""Find the first node of a given type. If no such node exists the
return value is `None`.
"""
for result in self.find_all(node_type):
return result
return None
def find_all(
self, node_type: t.Union[t.Type[_NodeBound], t.Tuple[t.Type[_NodeBound], ...]]
) -> t.Iterator[_NodeBound]:
"""Find all the nodes of a given type. If the type is a tuple,
the check is performed for any of the tuple items.
"""
for child in self.iter_child_nodes():
if isinstance(child, node_type):
yield child # type: ignore
yield from child.find_all(node_type)
def set_ctx(self, ctx: str) -> "Node":
"""Reset the context of a node and all child nodes. Per default the
parser will all generate nodes that have a 'load' context as it's the
most common one. This method is used in the parser to set assignment
targets and other nodes to a store context.
"""
todo = deque([self])
while todo:
node = todo.popleft()
if "ctx" in node.fields:
node.ctx = ctx # type: ignore
todo.extend(node.iter_child_nodes())
return self
def set_lineno(self, lineno: int, override: bool = False) -> "Node":
"""Set the line numbers of the node and children."""
todo = deque([self])
while todo:
node = todo.popleft()
if "lineno" in node.attributes:
if node.lineno is None or override:
node.lineno = lineno
todo.extend(node.iter_child_nodes())
return self
def set_environment(self, environment: "Environment") -> "Node":
"""Set the environment for all nodes."""
todo = deque([self])
while todo:
node = todo.popleft()
node.environment = environment
todo.extend(node.iter_child_nodes())
return self
def __eq__(self, other: t.Any) -> bool:
if type(self) is not type(other):
return NotImplemented
return tuple(self.iter_fields()) == tuple(other.iter_fields())
__hash__ = object.__hash__
def __repr__(self) -> str:
args_str = ", ".join(f"{a}={getattr(self, a, None)!r}" for a in self.fields)
return f"{type(self).__name__}({args_str})"
def dump(self) -> str:
def _dump(node: t.Union[Node, t.Any]) -> None:
if not isinstance(node, Node):
buf.append(repr(node))
return
buf.append(f"nodes.{type(node).__name__}(")
if not node.fields:
buf.append(")")
return
for idx, field in enumerate(node.fields):
if idx:
buf.append(", ")
value = getattr(node, field)
if isinstance(value, list):
buf.append("[")
for idx, item in enumerate(value):
if idx:
buf.append(", ")
_dump(item)
buf.append("]")
else:
_dump(value)
buf.append(")")
buf: t.List[str] = []
_dump(self)
return "".join(buf)
class Stmt(Node):
"""Base node for all statements."""
abstract = True
class Helper(Node):
"""Nodes that exist in a specific context only."""
abstract = True
class Template(Node):
"""Node that represents a template. This must be the outermost node that
is passed to the compiler.
"""
fields = ("body",)
body: t.List[Node]
class Output(Stmt):
"""A node that holds multiple expressions which are then printed out.
This is used both for the `print` statement and the regular template data.
"""
fields = ("nodes",)
nodes: t.List["Expr"]
class Extends(Stmt):
"""Represents an extends statement."""
fields = ("template",)
template: "Expr"
class For(Stmt):
"""The for loop. `target` is the target for the iteration (usually a
:class:`Name` or :class:`Tuple`), `iter` the iterable. `body` is a list
of nodes that are used as loop-body, and `else_` a list of nodes for the
`else` block. If no else node exists it has to be an empty list.
For filtered nodes an expression can be stored as `test`, otherwise `None`.
"""
fields = ("target", "iter", "body", "else_", "test", "recursive")
target: Node
iter: Node
body: t.List[Node]
else_: t.List[Node]
test: t.Optional[Node]
recursive: bool
class If(Stmt):
"""If `test` is true, `body` is rendered, else `else_`."""
fields = ("test", "body", "elif_", "else_")
test: Node
body: t.List[Node]
elif_: t.List["If"]
else_: t.List[Node]
class Macro(Stmt):
"""A macro definition. `name` is the name of the macro, `args` a list of
arguments and `defaults` a list of defaults if there are any. `body` is
a list of nodes for the macro body.
"""
fields = ("name", "args", "defaults", "body")
name: str
args: t.List["Name"]
defaults: t.List["Expr"]
body: t.List[Node]
class CallBlock(Stmt):
"""Like a macro without a name but a call instead. `call` is called with
the unnamed macro as `caller` argument this node holds.
"""
fields = ("call", "args", "defaults", "body")
call: "Call"
args: t.List["Name"]
defaults: t.List["Expr"]
body: t.List[Node]
class FilterBlock(Stmt):
"""Node for filter sections."""
fields = ("body", "filter")
body: t.List[Node]
filter: "Filter"
class With(Stmt):
"""Specific node for with statements. In older versions of Jinja the
with statement was implemented on the base of the `Scope` node instead.
.. versionadded:: 2.9.3
"""
fields = ("targets", "values", "body")
targets: t.List["Expr"]
values: t.List["Expr"]
body: t.List[Node]
class Block(Stmt):
"""A node that represents a block.
.. versionchanged:: 3.0.0
the `required` field was added.
"""
fields = ("name", "body", "scoped", "required")
name: str
body: t.List[Node]
scoped: bool
required: bool
class Include(Stmt):
"""A node that represents the include tag."""
fields = ("template", "with_context", "ignore_missing")
template: "Expr"
with_context: bool
ignore_missing: bool
class Import(Stmt):
"""A node that represents the import tag."""
fields = ("template", "target", "with_context")
template: "Expr"
target: str
with_context: bool
class FromImport(Stmt):
"""A node that represents the from import tag. It's important to not
pass unsafe names to the name attribute. The compiler translates the
attribute lookups directly into getattr calls and does *not* use the
subscript callback of the interface. As exported variables may not
start with double underscores (which the parser asserts) this is not a
problem for regular Jinja code, but if this node is used in an extension
extra care must be taken.
The list of names may contain tuples if aliases are wanted.
"""
fields = ("template", "names", "with_context")
template: "Expr"
names: t.List[t.Union[str, t.Tuple[str, str]]]
with_context: bool
class ExprStmt(Stmt):
"""A statement that evaluates an expression and discards the result."""
fields = ("node",)
node: Node
class Assign(Stmt):
"""Assigns an expression to a target."""
fields = ("target", "node")
target: "Expr"
node: Node
class AssignBlock(Stmt):
"""Assigns a block to a target."""
fields = ("target", "filter", "body")
target: "Expr"
filter: t.Optional["Filter"]
body: t.List[Node]
class Expr(Node):
"""Baseclass for all expressions."""
abstract = True
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
"""Return the value of the expression as constant or raise
:exc:`Impossible` if this was not possible.
An :class:`EvalContext` can be provided, if none is given
a default context is created which requires the nodes to have
an attached environment.
.. versionchanged:: 2.4
the `eval_ctx` parameter was added.
"""
raise Impossible()
def can_assign(self) -> bool:
"""Check if it's possible to assign something to this node."""
return False
class BinExpr(Expr):
"""Baseclass for all binary expressions."""
fields = ("left", "right")
left: Expr
right: Expr
operator: str
abstract = True
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
eval_ctx = get_eval_context(self, eval_ctx)
# intercepted operators cannot be folded at compile time
if (
eval_ctx.environment.sandboxed
and self.operator in eval_ctx.environment.intercepted_binops # type: ignore
):
raise Impossible()
f = _binop_to_func[self.operator]
try:
return f(self.left.as_const(eval_ctx), self.right.as_const(eval_ctx))
except Exception as e:
raise Impossible() from e
class UnaryExpr(Expr):
"""Baseclass for all unary expressions."""
fields = ("node",)
node: Expr
operator: str
abstract = True
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
eval_ctx = get_eval_context(self, eval_ctx)
# intercepted operators cannot be folded at compile time
if (
eval_ctx.environment.sandboxed
and self.operator in eval_ctx.environment.intercepted_unops # type: ignore
):
raise Impossible()
f = _uaop_to_func[self.operator]
try:
return f(self.node.as_const(eval_ctx))
except Exception as e:
raise Impossible() from e
class Name(Expr):
"""Looks up a name or stores a value in a name.
The `ctx` of the node can be one of the following values:
- `store`: store a value in the name
- `load`: load that name
- `param`: like `store` but if the name was defined as function parameter.
"""
fields = ("name", "ctx")
name: str
ctx: str
def can_assign(self) -> bool:
return self.name not in {"true", "false", "none", "True", "False", "None"}
class NSRef(Expr):
"""Reference to a namespace value assignment"""
fields = ("name", "attr")
name: str
attr: str
def can_assign(self) -> bool:
# We don't need any special checks here; NSRef assignments have a
# runtime check to ensure the target is a namespace object which will
# have been checked already as it is created using a normal assignment
# which goes through a `Name` node.
return True
class Literal(Expr):
"""Baseclass for literals."""
abstract = True
class Const(Literal):
"""All constant values. The parser will return this node for simple
constants such as ``42`` or ``"foo"`` but it can be used to store more
complex values such as lists too. Only constants with a safe
representation (objects where ``eval(repr(x)) == x`` is true).
"""
fields = ("value",)
value: t.Any
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
return self.value
@classmethod
def from_untrusted(
cls,
value: t.Any,
lineno: t.Optional[int] = None,
environment: "t.Optional[Environment]" = None,
) -> "Const":
"""Return a const object if the value is representable as
constant value in the generated code, otherwise it will raise
an `Impossible` exception.
"""
from .compiler import has_safe_repr
if not has_safe_repr(value):
raise Impossible()
return cls(value, lineno=lineno, environment=environment)
class TemplateData(Literal):
"""A constant template string."""
fields = ("data",)
data: str
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> str:
eval_ctx = get_eval_context(self, eval_ctx)
if eval_ctx.volatile:
raise Impossible()
if eval_ctx.autoescape:
return Markup(self.data)
return self.data
class Tuple(Literal):
"""For loop unpacking and some other things like multiple arguments
for subscripts. Like for :class:`Name` `ctx` specifies if the tuple
is used for loading the names or storing.
"""
fields = ("items", "ctx")
items: t.List[Expr]
ctx: str
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Tuple[t.Any, ...]:
eval_ctx = get_eval_context(self, eval_ctx)
return tuple(x.as_const(eval_ctx) for x in self.items)
def can_assign(self) -> bool:
for item in self.items:
if not item.can_assign():
return False
return True
class List(Literal):
"""Any list literal such as ``[1, 2, 3]``"""
fields = ("items",)
items: t.List[Expr]
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.List[t.Any]:
eval_ctx = get_eval_context(self, eval_ctx)
return [x.as_const(eval_ctx) for x in self.items]
class Dict(Literal):
"""Any dict literal such as ``{1: 2, 3: 4}``. The items must be a list of
:class:`Pair` nodes.
"""
fields = ("items",)
items: t.List["Pair"]
def as_const(
self, eval_ctx: t.Optional[EvalContext] = None
) -> t.Dict[t.Any, t.Any]:
eval_ctx = get_eval_context(self, eval_ctx)
return dict(x.as_const(eval_ctx) for x in self.items)
class Pair(Helper):
"""A key, value pair for dicts."""
fields = ("key", "value")
key: Expr
value: Expr
def as_const(
self, eval_ctx: t.Optional[EvalContext] = None
) -> t.Tuple[t.Any, t.Any]:
eval_ctx = get_eval_context(self, eval_ctx)
return self.key.as_const(eval_ctx), self.value.as_const(eval_ctx)
class Keyword(Helper):
"""A key, value pair for keyword arguments where key is a string."""
fields = ("key", "value")
key: str
value: Expr
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Tuple[str, t.Any]:
eval_ctx = get_eval_context(self, eval_ctx)
return self.key, self.value.as_const(eval_ctx)
class CondExpr(Expr):
"""A conditional expression (inline if expression). (``{{
foo if bar else baz }}``)
"""
fields = ("test", "expr1", "expr2")
test: Expr
expr1: Expr
expr2: t.Optional[Expr]
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
eval_ctx = get_eval_context(self, eval_ctx)
if self.test.as_const(eval_ctx):
return self.expr1.as_const(eval_ctx)
# if we evaluate to an undefined object, we better do that at runtime
if self.expr2 is None:
raise Impossible()
return self.expr2.as_const(eval_ctx)
def args_as_const(
node: t.Union["_FilterTestCommon", "Call"], eval_ctx: t.Optional[EvalContext]
) -> t.Tuple[t.List[t.Any], t.Dict[t.Any, t.Any]]:
args = [x.as_const(eval_ctx) for x in node.args]
kwargs = dict(x.as_const(eval_ctx) for x in node.kwargs)
if node.dyn_args is not None:
try:
args.extend(node.dyn_args.as_const(eval_ctx))
except Exception as e:
raise Impossible() from e
if node.dyn_kwargs is not None:
try:
kwargs.update(node.dyn_kwargs.as_const(eval_ctx))
except Exception as e:
raise Impossible() from e
return args, kwargs
class _FilterTestCommon(Expr):
fields = ("node", "name", "args", "kwargs", "dyn_args", "dyn_kwargs")
node: Expr
name: str
args: t.List[Expr]
kwargs: t.List[Pair]
dyn_args: t.Optional[Expr]
dyn_kwargs: t.Optional[Expr]
abstract = True
_is_filter = True
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
eval_ctx = get_eval_context(self, eval_ctx)
if eval_ctx.volatile:
raise Impossible()
if self._is_filter:
env_map = eval_ctx.environment.filters
else:
env_map = eval_ctx.environment.tests
func = env_map.get(self.name)
pass_arg = _PassArg.from_obj(func) # type: ignore
if func is None or pass_arg is _PassArg.context:
raise Impossible()
if eval_ctx.environment.is_async and (
getattr(func, "jinja_async_variant", False) is True
or inspect.iscoroutinefunction(func)
):
raise Impossible()
args, kwargs = args_as_const(self, eval_ctx)
args.insert(0, self.node.as_const(eval_ctx))
if pass_arg is _PassArg.eval_context:
args.insert(0, eval_ctx)
elif pass_arg is _PassArg.environment:
args.insert(0, eval_ctx.environment)
try:
return func(*args, **kwargs)
except Exception as e:
raise Impossible() from e
class Filter(_FilterTestCommon):
"""Apply a filter to an expression. ``name`` is the name of the
filter, the other fields are the same as :class:`Call`.
If ``node`` is ``None``, the filter is being used in a filter block
and is applied to the content of the block.
"""
node: t.Optional[Expr] # type: ignore
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
if self.node is None:
raise Impossible()
return super().as_const(eval_ctx=eval_ctx)
class Test(_FilterTestCommon):
"""Apply a test to an expression. ``name`` is the name of the test,
the other field are the same as :class:`Call`.
.. versionchanged:: 3.0
``as_const`` shares the same logic for filters and tests. Tests
check for volatile, async, and ``@pass_context`` etc.
decorators.
"""
_is_filter = False
class Call(Expr):
"""Calls an expression. `args` is a list of arguments, `kwargs` a list
of keyword arguments (list of :class:`Keyword` nodes), and `dyn_args`
and `dyn_kwargs` has to be either `None` or a node that is used as
node for dynamic positional (``*args``) or keyword (``**kwargs``)
arguments.
"""
fields = ("node", "args", "kwargs", "dyn_args", "dyn_kwargs")
node: Expr
args: t.List[Expr]
kwargs: t.List[Keyword]
dyn_args: t.Optional[Expr]
dyn_kwargs: t.Optional[Expr]
class Getitem(Expr):
"""Get an attribute or item from an expression and prefer the item."""
fields = ("node", "arg", "ctx")
node: Expr
arg: Expr
ctx: str
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
if self.ctx != "load":
raise Impossible()
eval_ctx = get_eval_context(self, eval_ctx)
try:
return eval_ctx.environment.getitem(
self.node.as_const(eval_ctx), self.arg.as_const(eval_ctx)
)
except Exception as e:
raise Impossible() from e
class Getattr(Expr):
"""Get an attribute or item from an expression that is a ascii-only
bytestring and prefer the attribute.
"""
fields = ("node", "attr", "ctx")
node: Expr
attr: str
ctx: str
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
if self.ctx != "load":
raise Impossible()
eval_ctx = get_eval_context(self, eval_ctx)
try:
return eval_ctx.environment.getattr(self.node.as_const(eval_ctx), self.attr)
except Exception as e:
raise Impossible() from e
class Slice(Expr):
"""Represents a slice object. This must only be used as argument for
:class:`Subscript`.
"""
fields = ("start", "stop", "step")
start: t.Optional[Expr]
stop: t.Optional[Expr]
step: t.Optional[Expr]
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> slice:
eval_ctx = get_eval_context(self, eval_ctx)
def const(obj: t.Optional[Expr]) -> t.Optional[t.Any]:
if obj is None:
return None
return obj.as_const(eval_ctx)
return slice(const(self.start), const(self.stop), const(self.step))
class Concat(Expr):
"""Concatenates the list of expressions provided after converting
them to strings.
"""
fields = ("nodes",)
nodes: t.List[Expr]
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> str:
eval_ctx = get_eval_context(self, eval_ctx)
return "".join(str(x.as_const(eval_ctx)) for x in self.nodes)
class Compare(Expr):
"""Compares an expression with some other expressions. `ops` must be a
list of :class:`Operand`\\s.
"""
fields = ("expr", "ops")
expr: Expr
ops: t.List["Operand"]
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
eval_ctx = get_eval_context(self, eval_ctx)
result = value = self.expr.as_const(eval_ctx)
try:
for op in self.ops:
new_value = op.expr.as_const(eval_ctx)
result = _cmpop_to_func[op.op](value, new_value)
if not result:
return False
value = new_value
except Exception as e:
raise Impossible() from e
return result
class Operand(Helper):
"""Holds an operator and an expression."""
fields = ("op", "expr")
op: str
expr: Expr
class Mul(BinExpr):
"""Multiplies the left with the right node."""
operator = "*"
class Div(BinExpr):
"""Divides the left by the right node."""
operator = "/"
class FloorDiv(BinExpr):
"""Divides the left by the right node and converts the
result into an integer by truncating.
"""
operator = "//"
class Add(BinExpr):
"""Add the left to the right node."""
operator = "+"
class Sub(BinExpr):
"""Subtract the right from the left node."""
operator = "-"
class Mod(BinExpr):
"""Left modulo right."""
operator = "%"
class Pow(BinExpr):
"""Left to the power of right."""
operator = "**"
class And(BinExpr):
"""Short circuited AND."""
operator = "and"
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
eval_ctx = get_eval_context(self, eval_ctx)
return self.left.as_const(eval_ctx) and self.right.as_const(eval_ctx)
class Or(BinExpr):
"""Short circuited OR."""
operator = "or"
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> t.Any:
eval_ctx = get_eval_context(self, eval_ctx)
return self.left.as_const(eval_ctx) or self.right.as_const(eval_ctx)
class Not(UnaryExpr):
"""Negate the expression."""
operator = "not"
class Neg(UnaryExpr):
"""Make the expression negative."""
operator = "-"
class Pos(UnaryExpr):
"""Make the expression positive (noop for most expressions)"""
operator = "+"
# Helpers for extensions
class EnvironmentAttribute(Expr):
"""Loads an attribute from the environment object. This is useful for
extensions that want to call a callback stored on the environment.
"""
fields = ("name",)
name: str
class ExtensionAttribute(Expr):
"""Returns the attribute of an extension bound to the environment.
The identifier is the identifier of the :class:`Extension`.
This node is usually constructed by calling the
:meth:`~jinja2.ext.Extension.attr` method on an extension.
"""
fields = ("identifier", "name")
identifier: str
name: str
class ImportedName(Expr):
"""If created with an import name the import name is returned on node
access. For example ``ImportedName('cgi.escape')`` returns the `escape`
function from the cgi module on evaluation. Imports are optimized by the
compiler so there is no need to assign them to local variables.
"""
fields = ("importname",)
importname: str
class InternalName(Expr):
"""An internal name in the compiler. You cannot create these nodes
yourself but the parser provides a
:meth:`~jinja2.parser.Parser.free_identifier` method that creates
a new identifier for you. This identifier is not available from the
template and is not treated specially by the compiler.
"""
fields = ("name",)
name: str
def __init__(self) -> None:
raise TypeError(
"Can't create internal names. Use the "
"`free_identifier` method on a parser."
)
class MarkSafe(Expr):
"""Mark the wrapped expression as safe (wrap it as `Markup`)."""
fields = ("expr",)
expr: Expr
def as_const(self, eval_ctx: t.Optional[EvalContext] = None) -> Markup:
eval_ctx = get_eval_context(self, eval_ctx)
return Markup(self.expr.as_const(eval_ctx))
class MarkSafeIfAutoescape(Expr):
"""Mark the wrapped expression as safe (wrap it as `Markup`) but
only if autoescaping is active.
.. versionadded:: 2.5
"""
fields = ("expr",)
expr: Expr
def as_const(
self, eval_ctx: t.Optional[EvalContext] = None
) -> t.Union[Markup, t.Any]:
eval_ctx = get_eval_context(self, eval_ctx)
if eval_ctx.volatile:
raise Impossible()
expr = self.expr.as_const(eval_ctx)
if eval_ctx.autoescape:
return Markup(expr)
return expr
class ContextReference(Expr):
"""Returns the current template context. It can be used like a
:class:`Name` node, with a ``'load'`` ctx and will return the
current :class:`~jinja2.runtime.Context` object.
Here an example that assigns the current template name to a
variable named `foo`::
Assign(Name('foo', ctx='store'),
Getattr(ContextReference(), 'name'))
This is basically equivalent to using the
:func:`~jinja2.pass_context` decorator when using the high-level
API, which causes a reference to the context to be passed as the
first argument to a function.
"""
class DerivedContextReference(Expr):
"""Return the current template context including locals. Behaves
exactly like :class:`ContextReference`, but includes local
variables, such as from a ``for`` loop.
.. versionadded:: 2.11
"""
class Continue(Stmt):
"""Continue a loop."""
class Break(Stmt):
"""Break a loop."""
class Scope(Stmt):
"""An artificial scope."""
fields = ("body",)
body: t.List[Node]
class OverlayScope(Stmt):
"""An overlay scope for extensions. This is a largely unoptimized scope
that however can be used to introduce completely arbitrary variables into
a sub scope from a dictionary or dictionary like object. The `context`
field has to evaluate to a dictionary object.
Example usage::
OverlayScope(context=self.call_method('get_context'),
body=[...])
.. versionadded:: 2.10
"""
fields = ("context", "body")
context: Expr
body: t.List[Node]
class EvalContextModifier(Stmt):
"""Modifies the eval context. For each option that should be modified,
a :class:`Keyword` has to be added to the :attr:`options` list.
Example to change the `autoescape` setting::
EvalContextModifier(options=[Keyword('autoescape', Const(True))])
"""
fields = ("options",)
options: t.List[Keyword]
class ScopedEvalContextModifier(EvalContextModifier):
"""Modifies the eval context and reverts it later. Works exactly like
:class:`EvalContextModifier` but will only modify the
:class:`~jinja2.nodes.EvalContext` for nodes in the :attr:`body`.
"""
fields = ("body",)
body: t.List[Node]
# make sure nobody creates custom nodes
def _failing_new(*args: t.Any, **kwargs: t.Any) -> "te.NoReturn":
raise TypeError("can't create custom node types")
NodeType.__new__ = staticmethod(_failing_new) # type: ignore
del _failing_new
"""The optimizer tries to constant fold expressions and modify the AST
in place so that it should be faster to evaluate.
Because the AST does not contain all the scoping information and the
compiler has to find that out, we cannot do all the optimizations we
want. For example, loop unrolling doesn't work because unrolled loops
would have a different scope. The solution would be a second syntax tree
that stored the scoping rules.
"""
import typing as t
from . import nodes
from .visitor import NodeTransformer
if t.TYPE_CHECKING:
from .environment import Environment
def optimize(node: nodes.Node, environment: "Environment") -> nodes.Node:
"""The context hint can be used to perform an static optimization
based on the context given."""
optimizer = Optimizer(environment)
return t.cast(nodes.Node, optimizer.visit(node))
class Optimizer(NodeTransformer):
def __init__(self, environment: "t.Optional[Environment]") -> None:
self.environment = environment
def generic_visit(
self, node: nodes.Node, *args: t.Any, **kwargs: t.Any
) -> nodes.Node:
node = super().generic_visit(node, *args, **kwargs)
# Do constant folding. Some other nodes besides Expr have
# as_const, but folding them causes errors later on.
if isinstance(node, nodes.Expr):
try:
return nodes.Const.from_untrusted(
node.as_const(args[0] if args else None),
lineno=node.lineno,
environment=self.environment,
)
except nodes.Impossible:
pass
return node
"""Parse tokens from the lexer into nodes for the compiler."""
import typing
import typing as t
from . import nodes
from .exceptions import TemplateAssertionError
from .exceptions import TemplateSyntaxError
from .lexer import describe_token
from .lexer import describe_token_expr
if t.TYPE_CHECKING:
import typing_extensions as te
from .environment import Environment
_ImportInclude = t.TypeVar("_ImportInclude", nodes.Import, nodes.Include)
_MacroCall = t.TypeVar("_MacroCall", nodes.Macro, nodes.CallBlock)
_statement_keywords = frozenset(
[
"for",
"if",
"block",
"extends",
"print",
"macro",
"include",
"from",
"import",
"set",
"with",
"autoescape",
]
)
_compare_operators = frozenset(["eq", "ne", "lt", "lteq", "gt", "gteq"])
_math_nodes: t.Dict[str, t.Type[nodes.Expr]] = {
"add": nodes.Add,
"sub": nodes.Sub,
"mul": nodes.Mul,
"div": nodes.Div,
"floordiv": nodes.FloorDiv,
"mod": nodes.Mod,
}
class Parser:
"""This is the central parsing class Jinja uses. It's passed to
extensions and can be used to parse expressions or statements.
"""
def __init__(
self,
environment: "Environment",
source: str,
name: t.Optional[str] = None,
filename: t.Optional[str] = None,
state: t.Optional[str] = None,
) -> None:
self.environment = environment
self.stream = environment._tokenize(source, name, filename, state)
self.name = name
self.filename = filename
self.closed = False
self.extensions: t.Dict[
str, t.Callable[["Parser"], t.Union[nodes.Node, t.List[nodes.Node]]]
] = {}
for extension in environment.iter_extensions():
for tag in extension.tags:
self.extensions[tag] = extension.parse
self._last_identifier = 0
self._tag_stack: t.List[str] = []
self._end_token_stack: t.List[t.Tuple[str, ...]] = []
def fail(
self,
msg: str,
lineno: t.Optional[int] = None,
exc: t.Type[TemplateSyntaxError] = TemplateSyntaxError,
) -> "te.NoReturn":
"""Convenience method that raises `exc` with the message, passed
line number or last line number as well as the current name and
filename.
"""
if lineno is None:
lineno = self.stream.current.lineno
raise exc(msg, lineno, self.name, self.filename)
def _fail_ut_eof(
self,
name: t.Optional[str],
end_token_stack: t.List[t.Tuple[str, ...]],
lineno: t.Optional[int],
) -> "te.NoReturn":
expected: t.Set[str] = set()
for exprs in end_token_stack:
expected.update(map(describe_token_expr, exprs))
if end_token_stack:
currently_looking: t.Optional[str] = " or ".join(
map(repr, map(describe_token_expr, end_token_stack[-1]))
)
else:
currently_looking = None
if name is None:
message = ["Unexpected end of template."]
else:
message = [f"Encountered unknown tag {name!r}."]
if currently_looking:
if name is not None and name in expected:
message.append(
"You probably made a nesting mistake. Jinja is expecting this tag,"
f" but currently looking for {currently_looking}."
)
else:
message.append(
f"Jinja was looking for the following tags: {currently_looking}."
)
if self._tag_stack:
message.append(
"The innermost block that needs to be closed is"
f" {self._tag_stack[-1]!r}."
)
self.fail(" ".join(message), lineno)
def fail_unknown_tag(
self, name: str, lineno: t.Optional[int] = None
) -> "te.NoReturn":
"""Called if the parser encounters an unknown tag. Tries to fail
with a human readable error message that could help to identify
the problem.
"""
self._fail_ut_eof(name, self._end_token_stack, lineno)
def fail_eof(
self,
end_tokens: t.Optional[t.Tuple[str, ...]] = None,
lineno: t.Optional[int] = None,
) -> "te.NoReturn":
"""Like fail_unknown_tag but for end of template situations."""
stack = list(self._end_token_stack)
if end_tokens is not None:
stack.append(end_tokens)
self._fail_ut_eof(None, stack, lineno)
def is_tuple_end(
self, extra_end_rules: t.Optional[t.Tuple[str, ...]] = None
) -> bool:
"""Are we at the end of a tuple?"""
if self.stream.current.type in ("variable_end", "block_end", "rparen"):
return True
elif extra_end_rules is not None:
return self.stream.current.test_any(extra_end_rules) # type: ignore
return False
def free_identifier(self, lineno: t.Optional[int] = None) -> nodes.InternalName:
"""Return a new free identifier as :class:`~jinja2.nodes.InternalName`."""
self._last_identifier += 1
rv = object.__new__(nodes.InternalName)
nodes.Node.__init__(rv, f"fi{self._last_identifier}", lineno=lineno)
return rv
def parse_statement(self) -> t.Union[nodes.Node, t.List[nodes.Node]]:
"""Parse a single statement."""
token = self.stream.current
if token.type != "name":
self.fail("tag name expected", token.lineno)
self._tag_stack.append(token.value)
pop_tag = True
try:
if token.value in _statement_keywords:
f = getattr(self, f"parse_{self.stream.current.value}")
return f() # type: ignore
if token.value == "call":
return self.parse_call_block()
if token.value == "filter":
return self.parse_filter_block()
ext = self.extensions.get(token.value)
if ext is not None:
return ext(self)
# did not work out, remove the token we pushed by accident
# from the stack so that the unknown tag fail function can
# produce a proper error message.
self._tag_stack.pop()
pop_tag = False
self.fail_unknown_tag(token.value, token.lineno)
finally:
if pop_tag:
self._tag_stack.pop()
def parse_statements(
self, end_tokens: t.Tuple[str, ...], drop_needle: bool = False
) -> t.List[nodes.Node]:
"""Parse multiple statements into a list until one of the end tokens
is reached. This is used to parse the body of statements as it also
parses template data if appropriate. The parser checks first if the
current token is a colon and skips it if there is one. Then it checks
for the block end and parses until if one of the `end_tokens` is
reached. Per default the active token in the stream at the end of
the call is the matched end token. If this is not wanted `drop_needle`
can be set to `True` and the end token is removed.
"""
# the first token may be a colon for python compatibility
self.stream.skip_if("colon")
# in the future it would be possible to add whole code sections
# by adding some sort of end of statement token and parsing those here.
self.stream.expect("block_end")
result = self.subparse(end_tokens)
# we reached the end of the template too early, the subparser
# does not check for this, so we do that now
if self.stream.current.type == "eof":
self.fail_eof(end_tokens)
if drop_needle:
next(self.stream)
return result
def parse_set(self) -> t.Union[nodes.Assign, nodes.AssignBlock]:
"""Parse an assign statement."""
lineno = next(self.stream).lineno
target = self.parse_assign_target(with_namespace=True)
if self.stream.skip_if("assign"):
expr = self.parse_tuple()
return nodes.Assign(target, expr, lineno=lineno)
filter_node = self.parse_filter(None)
body = self.parse_statements(("name:endset",), drop_needle=True)
return nodes.AssignBlock(target, filter_node, body, lineno=lineno)
def parse_for(self) -> nodes.For:
"""Parse a for loop."""
lineno = self.stream.expect("name:for").lineno
target = self.parse_assign_target(extra_end_rules=("name:in",))
self.stream.expect("name:in")
iter = self.parse_tuple(
with_condexpr=False, extra_end_rules=("name:recursive",)
)
test = None
if self.stream.skip_if("name:if"):
test = self.parse_expression()
recursive = self.stream.skip_if("name:recursive")
body = self.parse_statements(("name:endfor", "name:else"))
if next(self.stream).value == "endfor":
else_ = []
else:
else_ = self.parse_statements(("name:endfor",), drop_needle=True)
return nodes.For(target, iter, body, else_, test, recursive, lineno=lineno)
def parse_if(self) -> nodes.If:
"""Parse an if construct."""
node = result = nodes.If(lineno=self.stream.expect("name:if").lineno)
while True:
node.test = self.parse_tuple(with_condexpr=False)
node.body = self.parse_statements(("name:elif", "name:else", "name:endif"))
node.elif_ = []
node.else_ = []
token = next(self.stream)
if token.test("name:elif"):
node = nodes.If(lineno=self.stream.current.lineno)
result.elif_.append(node)
continue
elif token.test("name:else"):
result.else_ = self.parse_statements(("name:endif",), drop_needle=True)
break
return result
def parse_with(self) -> nodes.With:
node = nodes.With(lineno=next(self.stream).lineno)
targets: t.List[nodes.Expr] = []
values: t.List[nodes.Expr] = []
while self.stream.current.type != "block_end":
if targets:
self.stream.expect("comma")
target = self.parse_assign_target()
target.set_ctx("param")
targets.append(target)
self.stream.expect("assign")
values.append(self.parse_expression())
node.targets = targets
node.values = values
node.body = self.parse_statements(("name:endwith",), drop_needle=True)
return node
def parse_autoescape(self) -> nodes.Scope:
node = nodes.ScopedEvalContextModifier(lineno=next(self.stream).lineno)
node.options = [nodes.Keyword("autoescape", self.parse_expression())]
node.body = self.parse_statements(("name:endautoescape",), drop_needle=True)
return nodes.Scope([node])
def parse_block(self) -> nodes.Block:
node = nodes.Block(lineno=next(self.stream).lineno)
node.name = self.stream.expect("name").value
node.scoped = self.stream.skip_if("name:scoped")
node.required = self.stream.skip_if("name:required")
# common problem people encounter when switching from django
# to jinja. we do not support hyphens in block names, so let's
# raise a nicer error message in that case.
if self.stream.current.type == "sub":
self.fail(
"Block names in Jinja have to be valid Python identifiers and may not"
" contain hyphens, use an underscore instead."
)
node.body = self.parse_statements(("name:endblock",), drop_needle=True)
# enforce that required blocks only contain whitespace or comments
# by asserting that the body, if not empty, is just TemplateData nodes
# with whitespace data
if node.required:
for body_node in node.body:
if not isinstance(body_node, nodes.Output) or any(
not isinstance(output_node, nodes.TemplateData)
or not output_node.data.isspace()
for output_node in body_node.nodes
):
self.fail("Required blocks can only contain comments or whitespace")
self.stream.skip_if("name:" + node.name)
return node
def parse_extends(self) -> nodes.Extends:
node = nodes.Extends(lineno=next(self.stream).lineno)
node.template = self.parse_expression()
return node
def parse_import_context(
self, node: _ImportInclude, default: bool
) -> _ImportInclude:
if self.stream.current.test_any(
"name:with", "name:without"
) and self.stream.look().test("name:context"):
node.with_context = next(self.stream).value == "with"
self.stream.skip()
else:
node.with_context = default
return node
def parse_include(self) -> nodes.Include:
node = nodes.Include(lineno=next(self.stream).lineno)
node.template = self.parse_expression()
if self.stream.current.test("name:ignore") and self.stream.look().test(
"name:missing"
):
node.ignore_missing = True
self.stream.skip(2)
else:
node.ignore_missing = False
return self.parse_import_context(node, True)
def parse_import(self) -> nodes.Import:
node = nodes.Import(lineno=next(self.stream).lineno)
node.template = self.parse_expression()
self.stream.expect("name:as")
node.target = self.parse_assign_target(name_only=True).name
return self.parse_import_context(node, False)
def parse_from(self) -> nodes.FromImport:
node = nodes.FromImport(lineno=next(self.stream).lineno)
node.template = self.parse_expression()
self.stream.expect("name:import")
node.names = []
def parse_context() -> bool:
if self.stream.current.value in {
"with",
"without",
} and self.stream.look().test("name:context"):
node.with_context = next(self.stream).value == "with"
self.stream.skip()
return True
return False
while True:
if node.names:
self.stream.expect("comma")
if self.stream.current.type == "name":
if parse_context():
break
target = self.parse_assign_target(name_only=True)
if target.name.startswith("_"):
self.fail(
"names starting with an underline can not be imported",
target.lineno,
exc=TemplateAssertionError,
)
if self.stream.skip_if("name:as"):
alias = self.parse_assign_target(name_only=True)
node.names.append((target.name, alias.name))
else:
node.names.append(target.name)
if parse_context() or self.stream.current.type != "comma":
break
else:
self.stream.expect("name")
if not hasattr(node, "with_context"):
node.with_context = False
return node
def parse_signature(self, node: _MacroCall) -> None:
args = node.args = []
defaults = node.defaults = []
self.stream.expect("lparen")
while self.stream.current.type != "rparen":
if args:
self.stream.expect("comma")
arg = self.parse_assign_target(name_only=True)
arg.set_ctx("param")
if self.stream.skip_if("assign"):
defaults.append(self.parse_expression())
elif defaults:
self.fail("non-default argument follows default argument")
args.append(arg)
self.stream.expect("rparen")
def parse_call_block(self) -> nodes.CallBlock:
node = nodes.CallBlock(lineno=next(self.stream).lineno)
if self.stream.current.type == "lparen":
self.parse_signature(node)
else:
node.args = []
node.defaults = []
call_node = self.parse_expression()
if not isinstance(call_node, nodes.Call):
self.fail("expected call", node.lineno)
node.call = call_node
node.body = self.parse_statements(("name:endcall",), drop_needle=True)
return node
def parse_filter_block(self) -> nodes.FilterBlock:
node = nodes.FilterBlock(lineno=next(self.stream).lineno)
node.filter = self.parse_filter(None, start_inline=True) # type: ignore
node.body = self.parse_statements(("name:endfilter",), drop_needle=True)
return node
def parse_macro(self) -> nodes.Macro:
node = nodes.Macro(lineno=next(self.stream).lineno)
node.name = self.parse_assign_target(name_only=True).name
self.parse_signature(node)
node.body = self.parse_statements(("name:endmacro",), drop_needle=True)
return node
def parse_print(self) -> nodes.Output:
node = nodes.Output(lineno=next(self.stream).lineno)
node.nodes = []
while self.stream.current.type != "block_end":
if node.nodes:
self.stream.expect("comma")
node.nodes.append(self.parse_expression())
return node
@typing.overload
def parse_assign_target(
self, with_tuple: bool = ..., name_only: "te.Literal[True]" = ...
) -> nodes.Name: ...
@typing.overload
def parse_assign_target(
self,
with_tuple: bool = True,
name_only: bool = False,
extra_end_rules: t.Optional[t.Tuple[str, ...]] = None,
with_namespace: bool = False,
) -> t.Union[nodes.NSRef, nodes.Name, nodes.Tuple]: ...
def parse_assign_target(
self,
with_tuple: bool = True,
name_only: bool = False,
extra_end_rules: t.Optional[t.Tuple[str, ...]] = None,
with_namespace: bool = False,
) -> t.Union[nodes.NSRef, nodes.Name, nodes.Tuple]:
"""Parse an assignment target. As Jinja allows assignments to
tuples, this function can parse all allowed assignment targets. Per
default assignments to tuples are parsed, that can be disable however
by setting `with_tuple` to `False`. If only assignments to names are
wanted `name_only` can be set to `True`. The `extra_end_rules`
parameter is forwarded to the tuple parsing function. If
`with_namespace` is enabled, a namespace assignment may be parsed.
"""
target: nodes.Expr
if with_namespace and self.stream.look().type == "dot":
token = self.stream.expect("name")
next(self.stream) # dot
attr = self.stream.expect("name")
target = nodes.NSRef(token.value, attr.value, lineno=token.lineno)
elif name_only:
token = self.stream.expect("name")
target = nodes.Name(token.value, "store", lineno=token.lineno)
else:
if with_tuple:
target = self.parse_tuple(
simplified=True, extra_end_rules=extra_end_rules
)
else:
target = self.parse_primary()
target.set_ctx("store")
if not target.can_assign():
self.fail(
f"can't assign to {type(target).__name__.lower()!r}", target.lineno
)
return target # type: ignore
def parse_expression(self, with_condexpr: bool = True) -> nodes.Expr:
"""Parse an expression. Per default all expressions are parsed, if
the optional `with_condexpr` parameter is set to `False` conditional
expressions are not parsed.
"""
if with_condexpr:
return self.parse_condexpr()
return self.parse_or()
def parse_condexpr(self) -> nodes.Expr:
lineno = self.stream.current.lineno
expr1 = self.parse_or()
expr3: t.Optional[nodes.Expr]
while self.stream.skip_if("name:if"):
expr2 = self.parse_or()
if self.stream.skip_if("name:else"):
expr3 = self.parse_condexpr()
else:
expr3 = None
expr1 = nodes.CondExpr(expr2, expr1, expr3, lineno=lineno)
lineno = self.stream.current.lineno
return expr1
def parse_or(self) -> nodes.Expr:
lineno = self.stream.current.lineno
left = self.parse_and()
while self.stream.skip_if("name:or"):
right = self.parse_and()
left = nodes.Or(left, right, lineno=lineno)
lineno = self.stream.current.lineno
return left
def parse_and(self) -> nodes.Expr:
lineno = self.stream.current.lineno
left = self.parse_not()
while self.stream.skip_if("name:and"):
right = self.parse_not()
left = nodes.And(left, right, lineno=lineno)
lineno = self.stream.current.lineno
return left
def parse_not(self) -> nodes.Expr:
if self.stream.current.test("name:not"):
lineno = next(self.stream).lineno
return nodes.Not(self.parse_not(), lineno=lineno)
return self.parse_compare()
def parse_compare(self) -> nodes.Expr:
lineno = self.stream.current.lineno
expr = self.parse_math1()
ops = []
while True:
token_type = self.stream.current.type
if token_type in _compare_operators:
next(self.stream)
ops.append(nodes.Operand(token_type, self.parse_math1()))
elif self.stream.skip_if("name:in"):
ops.append(nodes.Operand("in", self.parse_math1()))
elif self.stream.current.test("name:not") and self.stream.look().test(
"name:in"
):
self.stream.skip(2)
ops.append(nodes.Operand("notin", self.parse_math1()))
else:
break
lineno = self.stream.current.lineno
if not ops:
return expr
return nodes.Compare(expr, ops, lineno=lineno)
def parse_math1(self) -> nodes.Expr:
lineno = self.stream.current.lineno
left = self.parse_concat()
while self.stream.current.type in ("add", "sub"):
cls = _math_nodes[self.stream.current.type]
next(self.stream)
right = self.parse_concat()
left = cls(left, right, lineno=lineno)
lineno = self.stream.current.lineno
return left
def parse_concat(self) -> nodes.Expr:
lineno = self.stream.current.lineno
args = [self.parse_math2()]
while self.stream.current.type == "tilde":
next(self.stream)
args.append(self.parse_math2())
if len(args) == 1:
return args[0]
return nodes.Concat(args, lineno=lineno)
def parse_math2(self) -> nodes.Expr:
lineno = self.stream.current.lineno
left = self.parse_pow()
while self.stream.current.type in ("mul", "div", "floordiv", "mod"):
cls = _math_nodes[self.stream.current.type]
next(self.stream)
right = self.parse_pow()
left = cls(left, right, lineno=lineno)
lineno = self.stream.current.lineno
return left
def parse_pow(self) -> nodes.Expr:
lineno = self.stream.current.lineno
left = self.parse_unary()
while self.stream.current.type == "pow":
next(self.stream)
right = self.parse_unary()
left = nodes.Pow(left, right, lineno=lineno)
lineno = self.stream.current.lineno
return left
def parse_unary(self, with_filter: bool = True) -> nodes.Expr:
token_type = self.stream.current.type
lineno = self.stream.current.lineno
node: nodes.Expr
if token_type == "sub":
next(self.stream)
node = nodes.Neg(self.parse_unary(False), lineno=lineno)
elif token_type == "add":
next(self.stream)
node = nodes.Pos(self.parse_unary(False), lineno=lineno)
else:
node = self.parse_primary()
node = self.parse_postfix(node)
if with_filter:
node = self.parse_filter_expr(node)
return node
def parse_primary(self) -> nodes.Expr:
token = self.stream.current
node: nodes.Expr
if token.type == "name":
if token.value in ("true", "false", "True", "False"):
node = nodes.Const(token.value in ("true", "True"), lineno=token.lineno)
elif token.value in ("none", "None"):
node = nodes.Const(None, lineno=token.lineno)
else:
node = nodes.Name(token.value, "load", lineno=token.lineno)
next(self.stream)
elif token.type == "string":
next(self.stream)
buf = [token.value]
lineno = token.lineno
while self.stream.current.type == "string":
buf.append(self.stream.current.value)
next(self.stream)
node = nodes.Const("".join(buf), lineno=lineno)
elif token.type in ("integer", "float"):
next(self.stream)
node = nodes.Const(token.value, lineno=token.lineno)
elif token.type == "lparen":
next(self.stream)
node = self.parse_tuple(explicit_parentheses=True)
self.stream.expect("rparen")
elif token.type == "lbracket":
node = self.parse_list()
elif token.type == "lbrace":
node = self.parse_dict()
else:
self.fail(f"unexpected {describe_token(token)!r}", token.lineno)
return node
def parse_tuple(
self,
simplified: bool = False,
with_condexpr: bool = True,
extra_end_rules: t.Optional[t.Tuple[str, ...]] = None,
explicit_parentheses: bool = False,
) -> t.Union[nodes.Tuple, nodes.Expr]:
"""Works like `parse_expression` but if multiple expressions are
delimited by a comma a :class:`~jinja2.nodes.Tuple` node is created.
This method could also return a regular expression instead of a tuple
if no commas where found.
The default parsing mode is a full tuple. If `simplified` is `True`
only names and literals are parsed. The `no_condexpr` parameter is
forwarded to :meth:`parse_expression`.
Because tuples do not require delimiters and may end in a bogus comma
an extra hint is needed that marks the end of a tuple. For example
for loops support tuples between `for` and `in`. In that case the
`extra_end_rules` is set to ``['name:in']``.
`explicit_parentheses` is true if the parsing was triggered by an
expression in parentheses. This is used to figure out if an empty
tuple is a valid expression or not.
"""
lineno = self.stream.current.lineno
if simplified:
parse = self.parse_primary
elif with_condexpr:
parse = self.parse_expression
else:
def parse() -> nodes.Expr:
return self.parse_expression(with_condexpr=False)
args: t.List[nodes.Expr] = []
is_tuple = False
while True:
if args:
self.stream.expect("comma")
if self.is_tuple_end(extra_end_rules):
break
args.append(parse())
if self.stream.current.type == "comma":
is_tuple = True
else:
break
lineno = self.stream.current.lineno
if not is_tuple:
if args:
return args[0]
# if we don't have explicit parentheses, an empty tuple is
# not a valid expression. This would mean nothing (literally
# nothing) in the spot of an expression would be an empty
# tuple.
if not explicit_parentheses:
self.fail(
"Expected an expression,"
f" got {describe_token(self.stream.current)!r}"
)
return nodes.Tuple(args, "load", lineno=lineno)
def parse_list(self) -> nodes.List:
token = self.stream.expect("lbracket")
items: t.List[nodes.Expr] = []
while self.stream.current.type != "rbracket":
if items:
self.stream.expect("comma")
if self.stream.current.type == "rbracket":
break
items.append(self.parse_expression())
self.stream.expect("rbracket")
return nodes.List(items, lineno=token.lineno)
def parse_dict(self) -> nodes.Dict:
token = self.stream.expect("lbrace")
items: t.List[nodes.Pair] = []
while self.stream.current.type != "rbrace":
if items:
self.stream.expect("comma")
if self.stream.current.type == "rbrace":
break
key = self.parse_expression()
self.stream.expect("colon")
value = self.parse_expression()
items.append(nodes.Pair(key, value, lineno=key.lineno))
self.stream.expect("rbrace")
return nodes.Dict(items, lineno=token.lineno)
def parse_postfix(self, node: nodes.Expr) -> nodes.Expr:
while True:
token_type = self.stream.current.type
if token_type == "dot" or token_type == "lbracket":
node = self.parse_subscript(node)
# calls are valid both after postfix expressions (getattr
# and getitem) as well as filters and tests
elif token_type == "lparen":
node = self.parse_call(node)
else:
break
return node
def parse_filter_expr(self, node: nodes.Expr) -> nodes.Expr:
while True:
token_type = self.stream.current.type
if token_type == "pipe":
node = self.parse_filter(node) # type: ignore
elif token_type == "name" and self.stream.current.value == "is":
node = self.parse_test(node)
# calls are valid both after postfix expressions (getattr
# and getitem) as well as filters and tests
elif token_type == "lparen":
node = self.parse_call(node)
else:
break
return node
def parse_subscript(
self, node: nodes.Expr
) -> t.Union[nodes.Getattr, nodes.Getitem]:
token = next(self.stream)
arg: nodes.Expr
if token.type == "dot":
attr_token = self.stream.current
next(self.stream)
if attr_token.type == "name":
return nodes.Getattr(
node, attr_token.value, "load", lineno=token.lineno
)
elif attr_token.type != "integer":
self.fail("expected name or number", attr_token.lineno)
arg = nodes.Const(attr_token.value, lineno=attr_token.lineno)
return nodes.Getitem(node, arg, "load", lineno=token.lineno)
if token.type == "lbracket":
args: t.List[nodes.Expr] = []
while self.stream.current.type != "rbracket":
if args:
self.stream.expect("comma")
args.append(self.parse_subscribed())
self.stream.expect("rbracket")
if len(args) == 1:
arg = args[0]
else:
arg = nodes.Tuple(args, "load", lineno=token.lineno)
return nodes.Getitem(node, arg, "load", lineno=token.lineno)
self.fail("expected subscript expression", token.lineno)
def parse_subscribed(self) -> nodes.Expr:
lineno = self.stream.current.lineno
args: t.List[t.Optional[nodes.Expr]]
if self.stream.current.type == "colon":
next(self.stream)
args = [None]
else:
node = self.parse_expression()
if self.stream.current.type != "colon":
return node
next(self.stream)
args = [node]
if self.stream.current.type == "colon":
args.append(None)
elif self.stream.current.type not in ("rbracket", "comma"):
args.append(self.parse_expression())
else:
args.append(None)
if self.stream.current.type == "colon":
next(self.stream)
if self.stream.current.type not in ("rbracket", "comma"):
args.append(self.parse_expression())
else:
args.append(None)
else:
args.append(None)
return nodes.Slice(lineno=lineno, *args) # noqa: B026
def parse_call_args(
self,
) -> t.Tuple[
t.List[nodes.Expr],
t.List[nodes.Keyword],
t.Optional[nodes.Expr],
t.Optional[nodes.Expr],
]:
token = self.stream.expect("lparen")
args = []
kwargs = []
dyn_args = None
dyn_kwargs = None
require_comma = False
def ensure(expr: bool) -> None:
if not expr:
self.fail("invalid syntax for function call expression", token.lineno)
while self.stream.current.type != "rparen":
if require_comma:
self.stream.expect("comma")
# support for trailing comma
if self.stream.current.type == "rparen":
break
if self.stream.current.type == "mul":
ensure(dyn_args is None and dyn_kwargs is None)
next(self.stream)
dyn_args = self.parse_expression()
elif self.stream.current.type == "pow":
ensure(dyn_kwargs is None)
next(self.stream)
dyn_kwargs = self.parse_expression()
else:
if (
self.stream.current.type == "name"
and self.stream.look().type == "assign"
):
# Parsing a kwarg
ensure(dyn_kwargs is None)
key = self.stream.current.value
self.stream.skip(2)
value = self.parse_expression()
kwargs.append(nodes.Keyword(key, value, lineno=value.lineno))
else:
# Parsing an arg
ensure(dyn_args is None and dyn_kwargs is None and not kwargs)
args.append(self.parse_expression())
require_comma = True
self.stream.expect("rparen")
return args, kwargs, dyn_args, dyn_kwargs
def parse_call(self, node: nodes.Expr) -> nodes.Call:
# The lparen will be expected in parse_call_args, but the lineno
# needs to be recorded before the stream is advanced.
token = self.stream.current
args, kwargs, dyn_args, dyn_kwargs = self.parse_call_args()
return nodes.Call(node, args, kwargs, dyn_args, dyn_kwargs, lineno=token.lineno)
def parse_filter(
self, node: t.Optional[nodes.Expr], start_inline: bool = False
) -> t.Optional[nodes.Expr]:
while self.stream.current.type == "pipe" or start_inline:
if not start_inline:
next(self.stream)
token = self.stream.expect("name")
name = token.value
while self.stream.current.type == "dot":
next(self.stream)
name += "." + self.stream.expect("name").value
if self.stream.current.type == "lparen":
args, kwargs, dyn_args, dyn_kwargs = self.parse_call_args()
else:
args = []
kwargs = []
dyn_args = dyn_kwargs = None
node = nodes.Filter(
node, name, args, kwargs, dyn_args, dyn_kwargs, lineno=token.lineno
)
start_inline = False
return node
def parse_test(self, node: nodes.Expr) -> nodes.Expr:
token = next(self.stream)
if self.stream.current.test("name:not"):
next(self.stream)
negated = True
else:
negated = False
name = self.stream.expect("name").value
while self.stream.current.type == "dot":
next(self.stream)
name += "." + self.stream.expect("name").value
dyn_args = dyn_kwargs = None
kwargs: t.List[nodes.Keyword] = []
if self.stream.current.type == "lparen":
args, kwargs, dyn_args, dyn_kwargs = self.parse_call_args()
elif self.stream.current.type in {
"name",
"string",
"integer",
"float",
"lparen",
"lbracket",
"lbrace",
} and not self.stream.current.test_any("name:else", "name:or", "name:and"):
if self.stream.current.test("name:is"):
self.fail("You cannot chain multiple tests with is")
arg_node = self.parse_primary()
arg_node = self.parse_postfix(arg_node)
args = [arg_node]
else:
args = []
node = nodes.Test(
node, name, args, kwargs, dyn_args, dyn_kwargs, lineno=token.lineno
)
if negated:
node = nodes.Not(node, lineno=token.lineno)
return node
def subparse(
self, end_tokens: t.Optional[t.Tuple[str, ...]] = None
) -> t.List[nodes.Node]:
body: t.List[nodes.Node] = []
data_buffer: t.List[nodes.Node] = []
add_data = data_buffer.append
if end_tokens is not None:
self._end_token_stack.append(end_tokens)
def flush_data() -> None:
if data_buffer:
lineno = data_buffer[0].lineno
body.append(nodes.Output(data_buffer[:], lineno=lineno))
del data_buffer[:]
try:
while self.stream:
token = self.stream.current
if token.type == "data":
if token.value:
add_data(nodes.TemplateData(token.value, lineno=token.lineno))
next(self.stream)
elif token.type == "variable_begin":
next(self.stream)
add_data(self.parse_tuple(with_condexpr=True))
self.stream.expect("variable_end")
elif token.type == "block_begin":
flush_data()
next(self.stream)
if end_tokens is not None and self.stream.current.test_any(
*end_tokens
):
return body
rv = self.parse_statement()
if isinstance(rv, list):
body.extend(rv)
else:
body.append(rv)
self.stream.expect("block_end")
else:
raise AssertionError("internal parsing error")
flush_data()
finally:
if end_tokens is not None:
self._end_token_stack.pop()
return body
def parse(self) -> nodes.Template:
"""Parse the whole template into a `Template` node."""
result = nodes.Template(self.subparse(), lineno=1)
result.set_environment(self.environment)
return result
"""The runtime functions and state used by compiled templates."""
import functools
import sys
import typing as t
from collections import abc
from itertools import chain
from markupsafe import escape # noqa: F401
from markupsafe import Markup
from markupsafe import soft_str
from .async_utils import auto_aiter
from .async_utils import auto_await # noqa: F401
from .exceptions import TemplateNotFound # noqa: F401
from .exceptions import TemplateRuntimeError # noqa: F401
from .exceptions import UndefinedError
from .nodes import EvalContext
from .utils import _PassArg
from .utils import concat
from .utils import internalcode
from .utils import missing
from .utils import Namespace # noqa: F401
from .utils import object_type_repr
from .utils import pass_eval_context
V = t.TypeVar("V")
F = t.TypeVar("F", bound=t.Callable[..., t.Any])
if t.TYPE_CHECKING:
import logging
import typing_extensions as te
from .environment import Environment
class LoopRenderFunc(te.Protocol):
def __call__(
self,
reciter: t.Iterable[V],
loop_render_func: "LoopRenderFunc",
depth: int = 0,
) -> str: ...
# these variables are exported to the template runtime
exported = [
"LoopContext",
"TemplateReference",
"Macro",
"Markup",
"TemplateRuntimeError",
"missing",
"escape",
"markup_join",
"str_join",
"identity",
"TemplateNotFound",
"Namespace",
"Undefined",
"internalcode",
]
async_exported = [
"AsyncLoopContext",
"auto_aiter",
"auto_await",
]
def identity(x: V) -> V:
"""Returns its argument. Useful for certain things in the
environment.
"""
return x
def markup_join(seq: t.Iterable[t.Any]) -> str:
"""Concatenation that escapes if necessary and converts to string."""
buf = []
iterator = map(soft_str, seq)
for arg in iterator:
buf.append(arg)
if hasattr(arg, "__html__"):
return Markup("").join(chain(buf, iterator))
return concat(buf)
def str_join(seq: t.Iterable[t.Any]) -> str:
"""Simple args to string conversion and concatenation."""
return concat(map(str, seq))
def new_context(
environment: "Environment",
template_name: t.Optional[str],
blocks: t.Dict[str, t.Callable[["Context"], t.Iterator[str]]],
vars: t.Optional[t.Dict[str, t.Any]] = None,
shared: bool = False,
globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
locals: t.Optional[t.Mapping[str, t.Any]] = None,
) -> "Context":
"""Internal helper for context creation."""
if vars is None:
vars = {}
if shared:
parent = vars
else:
parent = dict(globals or (), **vars)
if locals:
# if the parent is shared a copy should be created because
# we don't want to modify the dict passed
if shared:
parent = dict(parent)
for key, value in locals.items():
if value is not missing:
parent[key] = value
return environment.context_class(
environment, parent, template_name, blocks, globals=globals
)
class TemplateReference:
"""The `self` in templates."""
def __init__(self, context: "Context") -> None:
self.__context = context
def __getitem__(self, name: str) -> t.Any:
blocks = self.__context.blocks[name]
return BlockReference(name, self.__context, blocks, 0)
def __repr__(self) -> str:
return f"<{type(self).__name__} {self.__context.name!r}>"
def _dict_method_all(dict_method: F) -> F:
@functools.wraps(dict_method)
def f_all(self: "Context") -> t.Any:
return dict_method(self.get_all())
return t.cast(F, f_all)
@abc.Mapping.register
class Context:
"""The template context holds the variables of a template. It stores the
values passed to the template and also the names the template exports.
Creating instances is neither supported nor useful as it's created
automatically at various stages of the template evaluation and should not
be created by hand.
The context is immutable. Modifications on :attr:`parent` **must not**
happen and modifications on :attr:`vars` are allowed from generated
template code only. Template filters and global functions marked as
:func:`pass_context` get the active context passed as first argument
and are allowed to access the context read-only.
The template context supports read only dict operations (`get`,
`keys`, `values`, `items`, `iterkeys`, `itervalues`, `iteritems`,
`__getitem__`, `__contains__`). Additionally there is a :meth:`resolve`
method that doesn't fail with a `KeyError` but returns an
:class:`Undefined` object for missing variables.
"""
def __init__(
self,
environment: "Environment",
parent: t.Dict[str, t.Any],
name: t.Optional[str],
blocks: t.Dict[str, t.Callable[["Context"], t.Iterator[str]]],
globals: t.Optional[t.MutableMapping[str, t.Any]] = None,
):
self.parent = parent
self.vars: t.Dict[str, t.Any] = {}
self.environment: "Environment" = environment
self.eval_ctx = EvalContext(self.environment, name)
self.exported_vars: t.Set[str] = set()
self.name = name
self.globals_keys = set() if globals is None else set(globals)
# create the initial mapping of blocks. Whenever template inheritance
# takes place the runtime will update this mapping with the new blocks
# from the template.
self.blocks = {k: [v] for k, v in blocks.items()}
def super(
self, name: str, current: t.Callable[["Context"], t.Iterator[str]]
) -> t.Union["BlockReference", "Undefined"]:
"""Render a parent block."""
try:
blocks = self.blocks[name]
index = blocks.index(current) + 1
blocks[index]
except LookupError:
return self.environment.undefined(
f"there is no parent block called {name!r}.", name="super"
)
return BlockReference(name, self, blocks, index)
def get(self, key: str, default: t.Any = None) -> t.Any:
"""Look up a variable by name, or return a default if the key is
not found.
:param key: The variable name to look up.
:param default: The value to return if the key is not found.
"""
try:
return self[key]
except KeyError:
return default
def resolve(self, key: str) -> t.Union[t.Any, "Undefined"]:
"""Look up a variable by name, or return an :class:`Undefined`
object if the key is not found.
If you need to add custom behavior, override
:meth:`resolve_or_missing`, not this method. The various lookup
functions use that method, not this one.
:param key: The variable name to look up.
"""
rv = self.resolve_or_missing(key)
if rv is missing:
return self.environment.undefined(name=key)
return rv
def resolve_or_missing(self, key: str) -> t.Any:
"""Look up a variable by name, or return a ``missing`` sentinel
if the key is not found.
Override this method to add custom lookup behavior.
:meth:`resolve`, :meth:`get`, and :meth:`__getitem__` use this
method. Don't call this method directly.
:param key: The variable name to look up.
"""
if key in self.vars:
return self.vars[key]
if key in self.parent:
return self.parent[key]
return missing
def get_exported(self) -> t.Dict[str, t.Any]:
"""Get a new dict with the exported variables."""
return {k: self.vars[k] for k in self.exported_vars}
def get_all(self) -> t.Dict[str, t.Any]:
"""Return the complete context as dict including the exported
variables. For optimizations reasons this might not return an
actual copy so be careful with using it.
"""
if not self.vars:
return self.parent
if not self.parent:
return self.vars
return dict(self.parent, **self.vars)
@internalcode
def call(
__self,
__obj: t.Callable[..., t.Any],
*args: t.Any,
**kwargs: t.Any, # noqa: B902
) -> t.Union[t.Any, "Undefined"]:
"""Call the callable with the arguments and keyword arguments
provided but inject the active context or environment as first
argument if the callable has :func:`pass_context` or
:func:`pass_environment`.
"""
if __debug__:
__traceback_hide__ = True # noqa
# Allow callable classes to take a context
if (
hasattr(__obj, "__call__") # noqa: B004
and _PassArg.from_obj(__obj.__call__) is not None
):
__obj = __obj.__call__
pass_arg = _PassArg.from_obj(__obj)
if pass_arg is _PassArg.context:
# the active context should have access to variables set in
# loops and blocks without mutating the context itself
if kwargs.get("_loop_vars"):
__self = __self.derived(kwargs["_loop_vars"])
if kwargs.get("_block_vars"):
__self = __self.derived(kwargs["_block_vars"])
args = (__self,) + args
elif pass_arg is _PassArg.eval_context:
args = (__self.eval_ctx,) + args
elif pass_arg is _PassArg.environment:
args = (__self.environment,) + args
kwargs.pop("_block_vars", None)
kwargs.pop("_loop_vars", None)
try:
return __obj(*args, **kwargs)
except StopIteration:
return __self.environment.undefined(
"value was undefined because a callable raised a"
" StopIteration exception"
)
def derived(self, locals: t.Optional[t.Dict[str, t.Any]] = None) -> "Context":
"""Internal helper function to create a derived context. This is
used in situations where the system needs a new context in the same
template that is independent.
"""
context = new_context(
self.environment, self.name, {}, self.get_all(), True, None, locals
)
context.eval_ctx = self.eval_ctx
context.blocks.update((k, list(v)) for k, v in self.blocks.items())
return context
keys = _dict_method_all(dict.keys)
values = _dict_method_all(dict.values)
items = _dict_method_all(dict.items)
def __contains__(self, name: str) -> bool:
return name in self.vars or name in self.parent
def __getitem__(self, key: str) -> t.Any:
"""Look up a variable by name with ``[]`` syntax, or raise a
``KeyError`` if the key is not found.
"""
item = self.resolve_or_missing(key)
if item is missing:
raise KeyError(key)
return item
def __repr__(self) -> str:
return f"<{type(self).__name__} {self.get_all()!r} of {self.name!r}>"
class BlockReference:
"""One block on a template reference."""
def __init__(
self,
name: str,
context: "Context",
stack: t.List[t.Callable[["Context"], t.Iterator[str]]],
depth: int,
) -> None:
self.name = name
self._context = context
self._stack = stack
self._depth = depth
@property
def super(self) -> t.Union["BlockReference", "Undefined"]:
"""Super the block."""
if self._depth + 1 >= len(self._stack):
return self._context.environment.undefined(
f"there is no parent block called {self.name!r}.", name="super"
)
return BlockReference(self.name, self._context, self._stack, self._depth + 1)
@internalcode
async def _async_call(self) -> str:
rv = concat(
[x async for x in self._stack[self._depth](self._context)] # type: ignore
)
if self._context.eval_ctx.autoescape:
return Markup(rv)
return rv
@internalcode
def __call__(self) -> str:
if self._context.environment.is_async:
return self._async_call() # type: ignore
rv = concat(self._stack[self._depth](self._context))
if self._context.eval_ctx.autoescape:
return Markup(rv)
return rv
class LoopContext:
"""A wrapper iterable for dynamic ``for`` loops, with information
about the loop and iteration.
"""
#: Current iteration of the loop, starting at 0.
index0 = -1
_length: t.Optional[int] = None
_after: t.Any = missing
_current: t.Any = missing
_before: t.Any = missing
_last_changed_value: t.Any = missing
def __init__(
self,
iterable: t.Iterable[V],
undefined: t.Type["Undefined"],
recurse: t.Optional["LoopRenderFunc"] = None,
depth0: int = 0,
) -> None:
"""
:param iterable: Iterable to wrap.
:param undefined: :class:`Undefined` class to use for next and
previous items.
:param recurse: The function to render the loop body when the
loop is marked recursive.
:param depth0: Incremented when looping recursively.
"""
self._iterable = iterable
self._iterator = self._to_iterator(iterable)
self._undefined = undefined
self._recurse = recurse
#: How many levels deep a recursive loop currently is, starting at 0.
self.depth0 = depth0
@staticmethod
def _to_iterator(iterable: t.Iterable[V]) -> t.Iterator[V]:
return iter(iterable)
@property
def length(self) -> int:
"""Length of the iterable.
If the iterable is a generator or otherwise does not have a
size, it is eagerly evaluated to get a size.
"""
if self._length is not None:
return self._length
try:
self._length = len(self._iterable) # type: ignore
except TypeError:
iterable = list(self._iterator)
self._iterator = self._to_iterator(iterable)
self._length = len(iterable) + self.index + (self._after is not missing)
return self._length
def __len__(self) -> int:
return self.length
@property
def depth(self) -> int:
"""How many levels deep a recursive loop currently is, starting at 1."""
return self.depth0 + 1
@property
def index(self) -> int:
"""Current iteration of the loop, starting at 1."""
return self.index0 + 1
@property
def revindex0(self) -> int:
"""Number of iterations from the end of the loop, ending at 0.
Requires calculating :attr:`length`.
"""
return self.length - self.index
@property
def revindex(self) -> int:
"""Number of iterations from the end of the loop, ending at 1.
Requires calculating :attr:`length`.
"""
return self.length - self.index0
@property
def first(self) -> bool:
"""Whether this is the first iteration of the loop."""
return self.index0 == 0
def _peek_next(self) -> t.Any:
"""Return the next element in the iterable, or :data:`missing`
if the iterable is exhausted. Only peeks one item ahead, caching
the result in :attr:`_last` for use in subsequent checks. The
cache is reset when :meth:`__next__` is called.
"""
if self._after is not missing:
return self._after
self._after = next(self._iterator, missing)
return self._after
@property
def last(self) -> bool:
"""Whether this is the last iteration of the loop.
Causes the iterable to advance early. See
:func:`itertools.groupby` for issues this can cause.
The :func:`groupby` filter avoids that issue.
"""
return self._peek_next() is missing
@property
def previtem(self) -> t.Union[t.Any, "Undefined"]:
"""The item in the previous iteration. Undefined during the
first iteration.
"""
if self.first:
return self._undefined("there is no previous item")
return self._before
@property
def nextitem(self) -> t.Union[t.Any, "Undefined"]:
"""The item in the next iteration. Undefined during the last
iteration.
Causes the iterable to advance early. See
:func:`itertools.groupby` for issues this can cause.
The :func:`jinja-filters.groupby` filter avoids that issue.
"""
rv = self._peek_next()
if rv is missing:
return self._undefined("there is no next item")
return rv
def cycle(self, *args: V) -> V:
"""Return a value from the given args, cycling through based on
the current :attr:`index0`.
:param args: One or more values to cycle through.
"""
if not args:
raise TypeError("no items for cycling given")
return args[self.index0 % len(args)]
def changed(self, *value: t.Any) -> bool:
"""Return ``True`` if previously called with a different value
(including when called for the first time).
:param value: One or more values to compare to the last call.
"""
if self._last_changed_value != value:
self._last_changed_value = value
return True
return False
def __iter__(self) -> "LoopContext":
return self
def __next__(self) -> t.Tuple[t.Any, "LoopContext"]:
if self._after is not missing:
rv = self._after
self._after = missing
else:
rv = next(self._iterator)
self.index0 += 1
self._before = self._current
self._current = rv
return rv, self
@internalcode
def __call__(self, iterable: t.Iterable[V]) -> str:
"""When iterating over nested data, render the body of the loop
recursively with the given inner iterable data.
The loop must have the ``recursive`` marker for this to work.
"""
if self._recurse is None:
raise TypeError(
"The loop must have the 'recursive' marker to be called recursively."
)
return self._recurse(iterable, self._recurse, depth=self.depth)
def __repr__(self) -> str:
return f"<{type(self).__name__} {self.index}/{self.length}>"
class AsyncLoopContext(LoopContext):
_iterator: t.AsyncIterator[t.Any] # type: ignore
@staticmethod
def _to_iterator( # type: ignore
iterable: t.Union[t.Iterable[V], t.AsyncIterable[V]],
) -> t.AsyncIterator[V]:
return auto_aiter(iterable)
@property
async def length(self) -> int: # type: ignore
if self._length is not None:
return self._length
try:
self._length = len(self._iterable) # type: ignore
except TypeError:
iterable = [x async for x in self._iterator]
self._iterator = self._to_iterator(iterable)
self._length = len(iterable) + self.index + (self._after is not missing)
return self._length
@property
async def revindex0(self) -> int: # type: ignore
return await self.length - self.index
@property
async def revindex(self) -> int: # type: ignore
return await self.length - self.index0
async def _peek_next(self) -> t.Any:
if self._after is not missing:
return self._after
try:
self._after = await self._iterator.__anext__()
except StopAsyncIteration:
self._after = missing
return self._after
@property
async def last(self) -> bool: # type: ignore
return await self._peek_next() is missing
@property
async def nextitem(self) -> t.Union[t.Any, "Undefined"]:
rv = await self._peek_next()
if rv is missing:
return self._undefined("there is no next item")
return rv
def __aiter__(self) -> "AsyncLoopContext":
return self
async def __anext__(self) -> t.Tuple[t.Any, "AsyncLoopContext"]:
if self._after is not missing:
rv = self._after
self._after = missing
else:
rv = await self._iterator.__anext__()
self.index0 += 1
self._before = self._current
self._current = rv
return rv, self
class Macro:
"""Wraps a macro function."""
def __init__(
self,
environment: "Environment",
func: t.Callable[..., str],
name: str,
arguments: t.List[str],
catch_kwargs: bool,
catch_varargs: bool,
caller: bool,
default_autoescape: t.Optional[bool] = None,
):
self._environment = environment
self._func = func
self._argument_count = len(arguments)
self.name = name
self.arguments = arguments
self.catch_kwargs = catch_kwargs
self.catch_varargs = catch_varargs
self.caller = caller
self.explicit_caller = "caller" in arguments
if default_autoescape is None:
if callable(environment.autoescape):
default_autoescape = environment.autoescape(None)
else:
default_autoescape = environment.autoescape
self._default_autoescape = default_autoescape
@internalcode
@pass_eval_context
def __call__(self, *args: t.Any, **kwargs: t.Any) -> str:
# This requires a bit of explanation, In the past we used to
# decide largely based on compile-time information if a macro is
# safe or unsafe. While there was a volatile mode it was largely
# unused for deciding on escaping. This turns out to be
# problematic for macros because whether a macro is safe depends not
# on the escape mode when it was defined, but rather when it was used.
#
# Because however we export macros from the module system and
# there are historic callers that do not pass an eval context (and
# will continue to not pass one), we need to perform an instance
# check here.
#
# This is considered safe because an eval context is not a valid
# argument to callables otherwise anyway. Worst case here is
# that if no eval context is passed we fall back to the compile
# time autoescape flag.
if args and isinstance(args[0], EvalContext):
autoescape = args[0].autoescape
args = args[1:]
else:
autoescape = self._default_autoescape
# try to consume the positional arguments
arguments = list(args[: self._argument_count])
off = len(arguments)
# For information why this is necessary refer to the handling
# of caller in the `macro_body` handler in the compiler.
found_caller = False
# if the number of arguments consumed is not the number of
# arguments expected we start filling in keyword arguments
# and defaults.
if off != self._argument_count:
for name in self.arguments[len(arguments) :]:
try:
value = kwargs.pop(name)
except KeyError:
value = missing
if name == "caller":
found_caller = True
arguments.append(value)
else:
found_caller = self.explicit_caller
# it's important that the order of these arguments does not change
# if not also changed in the compiler's `function_scoping` method.
# the order is caller, keyword arguments, positional arguments!
if self.caller and not found_caller:
caller = kwargs.pop("caller", None)
if caller is None:
caller = self._environment.undefined("No caller defined", name="caller")
arguments.append(caller)
if self.catch_kwargs:
arguments.append(kwargs)
elif kwargs:
if "caller" in kwargs:
raise TypeError(
f"macro {self.name!r} was invoked with two values for the special"
" caller argument. This is most likely a bug."
)
raise TypeError(
f"macro {self.name!r} takes no keyword argument {next(iter(kwargs))!r}"
)
if self.catch_varargs:
arguments.append(args[self._argument_count :])
elif len(args) > self._argument_count:
raise TypeError(
f"macro {self.name!r} takes not more than"
f" {len(self.arguments)} argument(s)"
)
return self._invoke(arguments, autoescape)
async def _async_invoke(self, arguments: t.List[t.Any], autoescape: bool) -> str:
rv = await self._func(*arguments) # type: ignore
if autoescape:
return Markup(rv)
return rv # type: ignore
def _invoke(self, arguments: t.List[t.Any], autoescape: bool) -> str:
if self._environment.is_async:
return self._async_invoke(arguments, autoescape) # type: ignore
rv = self._func(*arguments)
if autoescape:
rv = Markup(rv)
return rv
def __repr__(self) -> str:
name = "anonymous" if self.name is None else repr(self.name)
return f"<{type(self).__name__} {name}>"
class Undefined:
"""The default undefined type. This undefined type can be printed and
iterated over, but every other access will raise an :exc:`UndefinedError`:
>>> foo = Undefined(name='foo')
>>> str(foo)
''
>>> not foo
True
>>> foo + 42
Traceback (most recent call last):
...
jinja2.exceptions.UndefinedError: 'foo' is undefined
"""
__slots__ = (
"_undefined_hint",
"_undefined_obj",
"_undefined_name",
"_undefined_exception",
)
def __init__(
self,
hint: t.Optional[str] = None,
obj: t.Any = missing,
name: t.Optional[str] = None,
exc: t.Type[TemplateRuntimeError] = UndefinedError,
) -> None:
self._undefined_hint = hint
self._undefined_obj = obj
self._undefined_name = name
self._undefined_exception = exc
@property
def _undefined_message(self) -> str:
"""Build a message about the undefined value based on how it was
accessed.
"""
if self._undefined_hint:
return self._undefined_hint
if self._undefined_obj is missing:
return f"{self._undefined_name!r} is undefined"
if not isinstance(self._undefined_name, str):
return (
f"{object_type_repr(self._undefined_obj)} has no"
f" element {self._undefined_name!r}"
)
return (
f"{object_type_repr(self._undefined_obj)!r} has no"
f" attribute {self._undefined_name!r}"
)
@internalcode
def _fail_with_undefined_error(
self, *args: t.Any, **kwargs: t.Any
) -> "te.NoReturn":
"""Raise an :exc:`UndefinedError` when operations are performed
on the undefined value.
"""
raise self._undefined_exception(self._undefined_message)
@internalcode
def __getattr__(self, name: str) -> t.Any:
if name[:2] == "__":
raise AttributeError(name)
return self._fail_with_undefined_error()
__add__ = __radd__ = __sub__ = __rsub__ = _fail_with_undefined_error
__mul__ = __rmul__ = __div__ = __rdiv__ = _fail_with_undefined_error
__truediv__ = __rtruediv__ = _fail_with_undefined_error
__floordiv__ = __rfloordiv__ = _fail_with_undefined_error
__mod__ = __rmod__ = _fail_with_undefined_error
__pos__ = __neg__ = _fail_with_undefined_error
__call__ = __getitem__ = _fail_with_undefined_error
__lt__ = __le__ = __gt__ = __ge__ = _fail_with_undefined_error
__int__ = __float__ = __complex__ = _fail_with_undefined_error
__pow__ = __rpow__ = _fail_with_undefined_error
def __eq__(self, other: t.Any) -> bool:
return type(self) is type(other)
def __ne__(self, other: t.Any) -> bool:
return not self.__eq__(other)
def __hash__(self) -> int:
return id(type(self))
def __str__(self) -> str:
return ""
def __len__(self) -> int:
return 0
def __iter__(self) -> t.Iterator[t.Any]:
yield from ()
async def __aiter__(self) -> t.AsyncIterator[t.Any]:
for _ in ():
yield
def __bool__(self) -> bool:
return False
def __repr__(self) -> str:
return "Undefined"
def make_logging_undefined(
logger: t.Optional["logging.Logger"] = None, base: t.Type[Undefined] = Undefined
) -> t.Type[Undefined]:
"""Given a logger object this returns a new undefined class that will
log certain failures. It will log iterations and printing. If no
logger is given a default logger is created.
Example::
logger = logging.getLogger(__name__)
LoggingUndefined = make_logging_undefined(
logger=logger,
base=Undefined
)
.. versionadded:: 2.8
:param logger: the logger to use. If not provided, a default logger
is created.
:param base: the base class to add logging functionality to. This
defaults to :class:`Undefined`.
"""
if logger is None:
import logging
logger = logging.getLogger(__name__)
logger.addHandler(logging.StreamHandler(sys.stderr))
def _log_message(undef: Undefined) -> None:
logger.warning("Template variable warning: %s", undef._undefined_message)
class LoggingUndefined(base): # type: ignore
__slots__ = ()
def _fail_with_undefined_error( # type: ignore
self, *args: t.Any, **kwargs: t.Any
) -> "te.NoReturn":
try:
super()._fail_with_undefined_error(*args, **kwargs)
except self._undefined_exception as e:
logger.error("Template variable error: %s", e) # type: ignore
raise e
def __str__(self) -> str:
_log_message(self)
return super().__str__() # type: ignore
def __iter__(self) -> t.Iterator[t.Any]:
_log_message(self)
return super().__iter__() # type: ignore
def __bool__(self) -> bool:
_log_message(self)
return super().__bool__() # type: ignore
return LoggingUndefined
class ChainableUndefined(Undefined):
"""An undefined that is chainable, where both ``__getattr__`` and
``__getitem__`` return itself rather than raising an
:exc:`UndefinedError`.
>>> foo = ChainableUndefined(name='foo')
>>> str(foo.bar['baz'])
''
>>> foo.bar['baz'] + 42
Traceback (most recent call last):
...
jinja2.exceptions.UndefinedError: 'foo' is undefined
.. versionadded:: 2.11.0
"""
__slots__ = ()
def __html__(self) -> str:
return str(self)
def __getattr__(self, _: str) -> "ChainableUndefined":
return self
__getitem__ = __getattr__ # type: ignore
class DebugUndefined(Undefined):
"""An undefined that returns the debug info when printed.
>>> foo = DebugUndefined(name='foo')
>>> str(foo)
'{{ foo }}'
>>> not foo
True
>>> foo + 42
Traceback (most recent call last):
...
jinja2.exceptions.UndefinedError: 'foo' is undefined
"""
__slots__ = ()
def __str__(self) -> str:
if self._undefined_hint:
message = f"undefined value printed: {self._undefined_hint}"
elif self._undefined_obj is missing:
message = self._undefined_name # type: ignore
else:
message = (
f"no such element: {object_type_repr(self._undefined_obj)}"
f"[{self._undefined_name!r}]"
)
return f"{{{{ {message} }}}}"
class StrictUndefined(Undefined):
"""An undefined that barks on print and iteration as well as boolean
tests and all kinds of comparisons. In other words: you can do nothing
with it except checking if it's defined using the `defined` test.
>>> foo = StrictUndefined(name='foo')
>>> str(foo)
Traceback (most recent call last):
...
jinja2.exceptions.UndefinedError: 'foo' is undefined
>>> not foo
Traceback (most recent call last):
...
jinja2.exceptions.UndefinedError: 'foo' is undefined
>>> foo + 42
Traceback (most recent call last):
...
jinja2.exceptions.UndefinedError: 'foo' is undefined
"""
__slots__ = ()
__iter__ = __str__ = __len__ = Undefined._fail_with_undefined_error
__eq__ = __ne__ = __bool__ = __hash__ = Undefined._fail_with_undefined_error
__contains__ = Undefined._fail_with_undefined_error
# Remove slots attributes, after the metaclass is applied they are
# unneeded and contain wrong data for subclasses.
del (
Undefined.__slots__,
ChainableUndefined.__slots__,
DebugUndefined.__slots__,
StrictUndefined.__slots__,
)
"""A sandbox layer that ensures unsafe operations cannot be performed.
Useful when the template itself comes from an untrusted source.
"""
import operator
import types
import typing as t
from collections import abc
from collections import deque
from string import Formatter
from _string import formatter_field_name_split # type: ignore
from markupsafe import EscapeFormatter
from markupsafe import Markup
from .environment import Environment
from .exceptions import SecurityError
from .runtime import Context
from .runtime import Undefined
F = t.TypeVar("F", bound=t.Callable[..., t.Any])
#: maximum number of items a range may produce
MAX_RANGE = 100000
#: Unsafe function attributes.
UNSAFE_FUNCTION_ATTRIBUTES: t.Set[str] = set()
#: Unsafe method attributes. Function attributes are unsafe for methods too.
UNSAFE_METHOD_ATTRIBUTES: t.Set[str] = set()
#: unsafe generator attributes.
UNSAFE_GENERATOR_ATTRIBUTES = {"gi_frame", "gi_code"}
#: unsafe attributes on coroutines
UNSAFE_COROUTINE_ATTRIBUTES = {"cr_frame", "cr_code"}
#: unsafe attributes on async generators
UNSAFE_ASYNC_GENERATOR_ATTRIBUTES = {"ag_code", "ag_frame"}
_mutable_spec: t.Tuple[t.Tuple[t.Type[t.Any], t.FrozenSet[str]], ...] = (
(
abc.MutableSet,
frozenset(
[
"add",
"clear",
"difference_update",
"discard",
"pop",
"remove",
"symmetric_difference_update",
"update",
]
),
),
(
abc.MutableMapping,
frozenset(["clear", "pop", "popitem", "setdefault", "update"]),
),
(
abc.MutableSequence,
frozenset(["append", "reverse", "insert", "sort", "extend", "remove"]),
),
(
deque,
frozenset(
[
"append",
"appendleft",
"clear",
"extend",
"extendleft",
"pop",
"popleft",
"remove",
"rotate",
]
),
),
)
def inspect_format_method(callable: t.Callable[..., t.Any]) -> t.Optional[str]:
if not isinstance(
callable, (types.MethodType, types.BuiltinMethodType)
) or callable.__name__ not in ("format", "format_map"):
return None
obj = callable.__self__
if isinstance(obj, str):
return obj
return None
def safe_range(*args: int) -> range:
"""A range that can't generate ranges with a length of more than
MAX_RANGE items.
"""
rng = range(*args)
if len(rng) > MAX_RANGE:
raise OverflowError(
"Range too big. The sandbox blocks ranges larger than"
f" MAX_RANGE ({MAX_RANGE})."
)
return rng
def unsafe(f: F) -> F:
"""Marks a function or method as unsafe.
.. code-block: python
@unsafe
def delete(self):
pass
"""
f.unsafe_callable = True # type: ignore
return f
def is_internal_attribute(obj: t.Any, attr: str) -> bool:
"""Test if the attribute given is an internal python attribute. For
example this function returns `True` for the `func_code` attribute of
python objects. This is useful if the environment method
:meth:`~SandboxedEnvironment.is_safe_attribute` is overridden.
>>> from jinja2.sandbox import is_internal_attribute
>>> is_internal_attribute(str, "mro")
True
>>> is_internal_attribute(str, "upper")
False
"""
if isinstance(obj, types.FunctionType):
if attr in UNSAFE_FUNCTION_ATTRIBUTES:
return True
elif isinstance(obj, types.MethodType):
if attr in UNSAFE_FUNCTION_ATTRIBUTES or attr in UNSAFE_METHOD_ATTRIBUTES:
return True
elif isinstance(obj, type):
if attr == "mro":
return True
elif isinstance(obj, (types.CodeType, types.TracebackType, types.FrameType)):
return True
elif isinstance(obj, types.GeneratorType):
if attr in UNSAFE_GENERATOR_ATTRIBUTES:
return True
elif hasattr(types, "CoroutineType") and isinstance(obj, types.CoroutineType):
if attr in UNSAFE_COROUTINE_ATTRIBUTES:
return True
elif hasattr(types, "AsyncGeneratorType") and isinstance(
obj, types.AsyncGeneratorType
):
if attr in UNSAFE_ASYNC_GENERATOR_ATTRIBUTES:
return True
return attr.startswith("__")
def modifies_known_mutable(obj: t.Any, attr: str) -> bool:
"""This function checks if an attribute on a builtin mutable object
(list, dict, set or deque) or the corresponding ABCs would modify it
if called.
>>> modifies_known_mutable({}, "clear")
True
>>> modifies_known_mutable({}, "keys")
False
>>> modifies_known_mutable([], "append")
True
>>> modifies_known_mutable([], "index")
False
If called with an unsupported object, ``False`` is returned.
>>> modifies_known_mutable("foo", "upper")
False
"""
for typespec, unsafe in _mutable_spec:
if isinstance(obj, typespec):
return attr in unsafe
return False
class SandboxedEnvironment(Environment):
"""The sandboxed environment. It works like the regular environment but
tells the compiler to generate sandboxed code. Additionally subclasses of
this environment may override the methods that tell the runtime what
attributes or functions are safe to access.
If the template tries to access insecure code a :exc:`SecurityError` is
raised. However also other exceptions may occur during the rendering so
the caller has to ensure that all exceptions are caught.
"""
sandboxed = True
#: default callback table for the binary operators. A copy of this is
#: available on each instance of a sandboxed environment as
#: :attr:`binop_table`
default_binop_table: t.Dict[str, t.Callable[[t.Any, t.Any], t.Any]] = {
"+": operator.add,
"-": operator.sub,
"*": operator.mul,
"/": operator.truediv,
"//": operator.floordiv,
"**": operator.pow,
"%": operator.mod,
}
#: default callback table for the unary operators. A copy of this is
#: available on each instance of a sandboxed environment as
#: :attr:`unop_table`
default_unop_table: t.Dict[str, t.Callable[[t.Any], t.Any]] = {
"+": operator.pos,
"-": operator.neg,
}
#: a set of binary operators that should be intercepted. Each operator
#: that is added to this set (empty by default) is delegated to the
#: :meth:`call_binop` method that will perform the operator. The default
#: operator callback is specified by :attr:`binop_table`.
#:
#: The following binary operators are interceptable:
#: ``//``, ``%``, ``+``, ``*``, ``-``, ``/``, and ``**``
#:
#: The default operation form the operator table corresponds to the
#: builtin function. Intercepted calls are always slower than the native
#: operator call, so make sure only to intercept the ones you are
#: interested in.
#:
#: .. versionadded:: 2.6
intercepted_binops: t.FrozenSet[str] = frozenset()
#: a set of unary operators that should be intercepted. Each operator
#: that is added to this set (empty by default) is delegated to the
#: :meth:`call_unop` method that will perform the operator. The default
#: operator callback is specified by :attr:`unop_table`.
#:
#: The following unary operators are interceptable: ``+``, ``-``
#:
#: The default operation form the operator table corresponds to the
#: builtin function. Intercepted calls are always slower than the native
#: operator call, so make sure only to intercept the ones you are
#: interested in.
#:
#: .. versionadded:: 2.6
intercepted_unops: t.FrozenSet[str] = frozenset()
def __init__(self, *args: t.Any, **kwargs: t.Any) -> None:
super().__init__(*args, **kwargs)
self.globals["range"] = safe_range
self.binop_table = self.default_binop_table.copy()
self.unop_table = self.default_unop_table.copy()
def is_safe_attribute(self, obj: t.Any, attr: str, value: t.Any) -> bool:
"""The sandboxed environment will call this method to check if the
attribute of an object is safe to access. Per default all attributes
starting with an underscore are considered private as well as the
special attributes of internal python objects as returned by the
:func:`is_internal_attribute` function.
"""
return not (attr.startswith("_") or is_internal_attribute(obj, attr))
def is_safe_callable(self, obj: t.Any) -> bool:
"""Check if an object is safely callable. By default callables
are considered safe unless decorated with :func:`unsafe`.
This also recognizes the Django convention of setting
``func.alters_data = True``.
"""
return not (
getattr(obj, "unsafe_callable", False) or getattr(obj, "alters_data", False)
)
def call_binop(
self, context: Context, operator: str, left: t.Any, right: t.Any
) -> t.Any:
"""For intercepted binary operator calls (:meth:`intercepted_binops`)
this function is executed instead of the builtin operator. This can
be used to fine tune the behavior of certain operators.
.. versionadded:: 2.6
"""
return self.binop_table[operator](left, right)
def call_unop(self, context: Context, operator: str, arg: t.Any) -> t.Any:
"""For intercepted unary operator calls (:meth:`intercepted_unops`)
this function is executed instead of the builtin operator. This can
be used to fine tune the behavior of certain operators.
.. versionadded:: 2.6
"""
return self.unop_table[operator](arg)
def getitem(
self, obj: t.Any, argument: t.Union[str, t.Any]
) -> t.Union[t.Any, Undefined]:
"""Subscribe an object from sandboxed code."""
try:
return obj[argument]
except (TypeError, LookupError):
if isinstance(argument, str):
try:
attr = str(argument)
except Exception:
pass
else:
try:
value = getattr(obj, attr)
except AttributeError:
pass
else:
if self.is_safe_attribute(obj, argument, value):
return value
return self.unsafe_undefined(obj, argument)
return self.undefined(obj=obj, name=argument)
def getattr(self, obj: t.Any, attribute: str) -> t.Union[t.Any, Undefined]:
"""Subscribe an object from sandboxed code and prefer the
attribute. The attribute passed *must* be a bytestring.
"""
try:
value = getattr(obj, attribute)
except AttributeError:
try:
return obj[attribute]
except (TypeError, LookupError):
pass
else:
if self.is_safe_attribute(obj, attribute, value):
return value
return self.unsafe_undefined(obj, attribute)
return self.undefined(obj=obj, name=attribute)
def unsafe_undefined(self, obj: t.Any, attribute: str) -> Undefined:
"""Return an undefined object for unsafe attributes."""
return self.undefined(
f"access to attribute {attribute!r} of"
f" {type(obj).__name__!r} object is unsafe.",
name=attribute,
obj=obj,
exc=SecurityError,
)
def format_string(
self,
s: str,
args: t.Tuple[t.Any, ...],
kwargs: t.Dict[str, t.Any],
format_func: t.Optional[t.Callable[..., t.Any]] = None,
) -> str:
"""If a format call is detected, then this is routed through this
method so that our safety sandbox can be used for it.
"""
formatter: SandboxedFormatter
if isinstance(s, Markup):
formatter = SandboxedEscapeFormatter(self, escape=s.escape)
else:
formatter = SandboxedFormatter(self)
if format_func is not None and format_func.__name__ == "format_map":
if len(args) != 1 or kwargs:
raise TypeError(
"format_map() takes exactly one argument"
f" {len(args) + (kwargs is not None)} given"
)
kwargs = args[0]
args = ()
rv = formatter.vformat(s, args, kwargs)
return type(s)(rv)
def call(
__self, # noqa: B902
__context: Context,
__obj: t.Any,
*args: t.Any,
**kwargs: t.Any,
) -> t.Any:
"""Call an object from sandboxed code."""
fmt = inspect_format_method(__obj)
if fmt is not None:
return __self.format_string(fmt, args, kwargs, __obj)
# the double prefixes are to avoid double keyword argument
# errors when proxying the call.
if not __self.is_safe_callable(__obj):
raise SecurityError(f"{__obj!r} is not safely callable")
return __context.call(__obj, *args, **kwargs)
class ImmutableSandboxedEnvironment(SandboxedEnvironment):
"""Works exactly like the regular `SandboxedEnvironment` but does not
permit modifications on the builtin mutable objects `list`, `set`, and
`dict` by using the :func:`modifies_known_mutable` function.
"""
def is_safe_attribute(self, obj: t.Any, attr: str, value: t.Any) -> bool:
if not super().is_safe_attribute(obj, attr, value):
return False
return not modifies_known_mutable(obj, attr)
class SandboxedFormatter(Formatter):
def __init__(self, env: Environment, **kwargs: t.Any) -> None:
self._env = env
super().__init__(**kwargs)
def get_field(
self, field_name: str, args: t.Sequence[t.Any], kwargs: t.Mapping[str, t.Any]
) -> t.Tuple[t.Any, str]:
first, rest = formatter_field_name_split(field_name)
obj = self.get_value(first, args, kwargs)
for is_attr, i in rest:
if is_attr:
obj = self._env.getattr(obj, i)
else:
obj = self._env.getitem(obj, i)
return obj, first
class SandboxedEscapeFormatter(SandboxedFormatter, EscapeFormatter):
pass
"""Built-in template tests used with the ``is`` operator."""
import operator
import typing as t
from collections import abc
from numbers import Number
from .runtime import Undefined
from .utils import pass_environment
if t.TYPE_CHECKING:
from .environment import Environment
def test_odd(value: int) -> bool:
"""Return true if the variable is odd."""
return value % 2 == 1
def test_even(value: int) -> bool:
"""Return true if the variable is even."""
return value % 2 == 0
def test_divisibleby(value: int, num: int) -> bool:
"""Check if a variable is divisible by a number."""
return value % num == 0
def test_defined(value: t.Any) -> bool:
"""Return true if the variable is defined:
.. sourcecode:: jinja
{% if variable is defined %}
value of variable: {{ variable }}
{% else %}
variable is not defined
{% endif %}
See the :func:`default` filter for a simple way to set undefined
variables.
"""
return not isinstance(value, Undefined)
def test_undefined(value: t.Any) -> bool:
"""Like :func:`defined` but the other way round."""
return isinstance(value, Undefined)
@pass_environment
def test_filter(env: "Environment", value: str) -> bool:
"""Check if a filter exists by name. Useful if a filter may be
optionally available.
.. code-block:: jinja
{% if 'markdown' is filter %}
{{ value | markdown }}
{% else %}
{{ value }}
{% endif %}
.. versionadded:: 3.0
"""
return value in env.filters
@pass_environment
def test_test(env: "Environment", value: str) -> bool:
"""Check if a test exists by name. Useful if a test may be
optionally available.
.. code-block:: jinja
{% if 'loud' is test %}
{% if value is loud %}
{{ value|upper }}
{% else %}
{{ value|lower }}
{% endif %}
{% else %}
{{ value }}
{% endif %}
.. versionadded:: 3.0
"""
return value in env.tests
def test_none(value: t.Any) -> bool:
"""Return true if the variable is none."""
return value is None
def test_boolean(value: t.Any) -> bool:
"""Return true if the object is a boolean value.
.. versionadded:: 2.11
"""
return value is True or value is False
def test_false(value: t.Any) -> bool:
"""Return true if the object is False.
.. versionadded:: 2.11
"""
return value is False
def test_true(value: t.Any) -> bool:
"""Return true if the object is True.
.. versionadded:: 2.11
"""
return value is True
# NOTE: The existing 'number' test matches booleans and floats
def test_integer(value: t.Any) -> bool:
"""Return true if the object is an integer.
.. versionadded:: 2.11
"""
return isinstance(value, int) and value is not True and value is not False
# NOTE: The existing 'number' test matches booleans and integers
def test_float(value: t.Any) -> bool:
"""Return true if the object is a float.
.. versionadded:: 2.11
"""
return isinstance(value, float)
def test_lower(value: str) -> bool:
"""Return true if the variable is lowercased."""
return str(value).islower()
def test_upper(value: str) -> bool:
"""Return true if the variable is uppercased."""
return str(value).isupper()
def test_string(value: t.Any) -> bool:
"""Return true if the object is a string."""
return isinstance(value, str)
def test_mapping(value: t.Any) -> bool:
"""Return true if the object is a mapping (dict etc.).
.. versionadded:: 2.6
"""
return isinstance(value, abc.Mapping)
def test_number(value: t.Any) -> bool:
"""Return true if the variable is a number."""
return isinstance(value, Number)
def test_sequence(value: t.Any) -> bool:
"""Return true if the variable is a sequence. Sequences are variables
that are iterable.
"""
try:
len(value)
value.__getitem__ # noqa B018
except Exception:
return False
return True
def test_sameas(value: t.Any, other: t.Any) -> bool:
"""Check if an object points to the same memory address than another
object:
.. sourcecode:: jinja
{% if foo.attribute is sameas false %}
the foo attribute really is the `False` singleton
{% endif %}
"""
return value is other
def test_iterable(value: t.Any) -> bool:
"""Check if it's possible to iterate over an object."""
try:
iter(value)
except TypeError:
return False
return True
def test_escaped(value: t.Any) -> bool:
"""Check if the value is escaped."""
return hasattr(value, "__html__")
def test_in(value: t.Any, seq: t.Container[t.Any]) -> bool:
"""Check if value is in seq.
.. versionadded:: 2.10
"""
return value in seq
TESTS = {
"odd": test_odd,
"even": test_even,
"divisibleby": test_divisibleby,
"defined": test_defined,
"undefined": test_undefined,
"filter": test_filter,
"test": test_test,
"none": test_none,
"boolean": test_boolean,
"false": test_false,
"true": test_true,
"integer": test_integer,
"float": test_float,
"lower": test_lower,
"upper": test_upper,
"string": test_string,
"mapping": test_mapping,
"number": test_number,
"sequence": test_sequence,
"iterable": test_iterable,
"callable": callable,
"sameas": test_sameas,
"escaped": test_escaped,
"in": test_in,
"==": operator.eq,
"eq": operator.eq,
"equalto": operator.eq,
"!=": operator.ne,
"ne": operator.ne,
">": operator.gt,
"gt": operator.gt,
"greaterthan": operator.gt,
"ge": operator.ge,
">=": operator.ge,
"<": operator.lt,
"lt": operator.lt,
"lessthan": operator.lt,
"<=": operator.le,
"le": operator.le,
}
import enum
import json
import os
import re
import typing as t
from collections import abc
from collections import deque
from random import choice
from random import randrange
from threading import Lock
from types import CodeType
from urllib.parse import quote_from_bytes
import markupsafe
if t.TYPE_CHECKING:
import typing_extensions as te
F = t.TypeVar("F", bound=t.Callable[..., t.Any])
# special singleton representing missing values for the runtime
missing: t.Any = type("MissingType", (), {"__repr__": lambda x: "missing"})()
internal_code: t.MutableSet[CodeType] = set()
concat = "".join
def pass_context(f: F) -> F:
"""Pass the :class:`~jinja2.runtime.Context` as the first argument
to the decorated function when called while rendering a template.
Can be used on functions, filters, and tests.
If only ``Context.eval_context`` is needed, use
:func:`pass_eval_context`. If only ``Context.environment`` is
needed, use :func:`pass_environment`.
.. versionadded:: 3.0.0
Replaces ``contextfunction`` and ``contextfilter``.
"""
f.jinja_pass_arg = _PassArg.context # type: ignore
return f
def pass_eval_context(f: F) -> F:
"""Pass the :class:`~jinja2.nodes.EvalContext` as the first argument
to the decorated function when called while rendering a template.
See :ref:`eval-context`.
Can be used on functions, filters, and tests.
If only ``EvalContext.environment`` is needed, use
:func:`pass_environment`.
.. versionadded:: 3.0.0
Replaces ``evalcontextfunction`` and ``evalcontextfilter``.
"""
f.jinja_pass_arg = _PassArg.eval_context # type: ignore
return f
def pass_environment(f: F) -> F:
"""Pass the :class:`~jinja2.Environment` as the first argument to
the decorated function when called while rendering a template.
Can be used on functions, filters, and tests.
.. versionadded:: 3.0.0
Replaces ``environmentfunction`` and ``environmentfilter``.
"""
f.jinja_pass_arg = _PassArg.environment # type: ignore
return f
class _PassArg(enum.Enum):
context = enum.auto()
eval_context = enum.auto()
environment = enum.auto()
@classmethod
def from_obj(cls, obj: F) -> t.Optional["_PassArg"]:
if hasattr(obj, "jinja_pass_arg"):
return obj.jinja_pass_arg # type: ignore
return None
def internalcode(f: F) -> F:
"""Marks the function as internally used"""
internal_code.add(f.__code__)
return f
def is_undefined(obj: t.Any) -> bool:
"""Check if the object passed is undefined. This does nothing more than
performing an instance check against :class:`Undefined` but looks nicer.
This can be used for custom filters or tests that want to react to
undefined variables. For example a custom default filter can look like
this::
def default(var, default=''):
if is_undefined(var):
return default
return var
"""
from .runtime import Undefined
return isinstance(obj, Undefined)
def consume(iterable: t.Iterable[t.Any]) -> None:
"""Consumes an iterable without doing anything with it."""
for _ in iterable:
pass
def clear_caches() -> None:
"""Jinja keeps internal caches for environments and lexers. These are
used so that Jinja doesn't have to recreate environments and lexers all
the time. Normally you don't have to care about that but if you are
measuring memory consumption you may want to clean the caches.
"""
from .environment import get_spontaneous_environment
from .lexer import _lexer_cache
get_spontaneous_environment.cache_clear()
_lexer_cache.clear()
def import_string(import_name: str, silent: bool = False) -> t.Any:
"""Imports an object based on a string. This is useful if you want to
use import paths as endpoints or something similar. An import path can
be specified either in dotted notation (``xml.sax.saxutils.escape``)
or with a colon as object delimiter (``xml.sax.saxutils:escape``).
If the `silent` is True the return value will be `None` if the import
fails.
:return: imported object
"""
try:
if ":" in import_name:
module, obj = import_name.split(":", 1)
elif "." in import_name:
module, _, obj = import_name.rpartition(".")
else:
return __import__(import_name)
return getattr(__import__(module, None, None, [obj]), obj)
except (ImportError, AttributeError):
if not silent:
raise
def open_if_exists(filename: str, mode: str = "rb") -> t.Optional[t.IO[t.Any]]:
"""Returns a file descriptor for the filename if that file exists,
otherwise ``None``.
"""
if not os.path.isfile(filename):
return None
return open(filename, mode)
def object_type_repr(obj: t.Any) -> str:
"""Returns the name of the object's type. For some recognized
singletons the name of the object is returned instead. (For
example for `None` and `Ellipsis`).
"""
if obj is None:
return "None"
elif obj is Ellipsis:
return "Ellipsis"
cls = type(obj)
if cls.__module__ == "builtins":
return f"{cls.__name__} object"
return f"{cls.__module__}.{cls.__name__} object"
def pformat(obj: t.Any) -> str:
"""Format an object using :func:`pprint.pformat`."""
from pprint import pformat
return pformat(obj)
_http_re = re.compile(
r"""
^
(
(https?://|www\.) # scheme or www
(([\w%-]+\.)+)? # subdomain
(
[a-z]{2,63} # basic tld
|
xn--[\w%]{2,59} # idna tld
)
|
([\w%-]{2,63}\.)+ # basic domain
(com|net|int|edu|gov|org|info|mil) # basic tld
|
(https?://) # scheme
(
(([\d]{1,3})(\.[\d]{1,3}){3}) # IPv4
|
(\[([\da-f]{0,4}:){2}([\da-f]{0,4}:?){1,6}]) # IPv6
)
)
(?::[\d]{1,5})? # port
(?:[/?#]\S*)? # path, query, and fragment
$
""",
re.IGNORECASE | re.VERBOSE,
)
_email_re = re.compile(r"^\S+@\w[\w.-]*\.\w+$")
def urlize(
text: str,
trim_url_limit: t.Optional[int] = None,
rel: t.Optional[str] = None,
target: t.Optional[str] = None,
extra_schemes: t.Optional[t.Iterable[str]] = None,
) -> str:
"""Convert URLs in text into clickable links.
This may not recognize links in some situations. Usually, a more
comprehensive formatter, such as a Markdown library, is a better
choice.
Works on ``http://``, ``https://``, ``www.``, ``mailto:``, and email
addresses. Links with trailing punctuation (periods, commas, closing
parentheses) and leading punctuation (opening parentheses) are
recognized excluding the punctuation. Email addresses that include
header fields are not recognized (for example,
``mailto:address@example.com?cc=copy@example.com``).
:param text: Original text containing URLs to link.
:param trim_url_limit: Shorten displayed URL values to this length.
:param target: Add the ``target`` attribute to links.
:param rel: Add the ``rel`` attribute to links.
:param extra_schemes: Recognize URLs that start with these schemes
in addition to the default behavior.
.. versionchanged:: 3.0
The ``extra_schemes`` parameter was added.
.. versionchanged:: 3.0
Generate ``https://`` links for URLs without a scheme.
.. versionchanged:: 3.0
The parsing rules were updated. Recognize email addresses with
or without the ``mailto:`` scheme. Validate IP addresses. Ignore
parentheses and brackets in more cases.
"""
if trim_url_limit is not None:
def trim_url(x: str) -> str:
if len(x) > trim_url_limit:
return f"{x[:trim_url_limit]}..."
return x
else:
def trim_url(x: str) -> str:
return x
words = re.split(r"(\s+)", str(markupsafe.escape(text)))
rel_attr = f' rel="{markupsafe.escape(rel)}"' if rel else ""
target_attr = f' target="{markupsafe.escape(target)}"' if target else ""
for i, word in enumerate(words):
head, middle, tail = "", word, ""
match = re.match(r"^([(<]|&lt;)+", middle)
if match:
head = match.group()
middle = middle[match.end() :]
# Unlike lead, which is anchored to the start of the string,
# need to check that the string ends with any of the characters
# before trying to match all of them, to avoid backtracking.
if middle.endswith((")", ">", ".", ",", "\n", "&gt;")):
match = re.search(r"([)>.,\n]|&gt;)+$", middle)
if match:
tail = match.group()
middle = middle[: match.start()]
# Prefer balancing parentheses in URLs instead of ignoring a
# trailing character.
for start_char, end_char in ("(", ")"), ("<", ">"), ("&lt;", "&gt;"):
start_count = middle.count(start_char)
if start_count <= middle.count(end_char):
# Balanced, or lighter on the left
continue
# Move as many as possible from the tail to balance
for _ in range(min(start_count, tail.count(end_char))):
end_index = tail.index(end_char) + len(end_char)
# Move anything in the tail before the end char too
middle += tail[:end_index]
tail = tail[end_index:]
if _http_re.match(middle):
if middle.startswith("https://") or middle.startswith("http://"):
middle = (
f'<a href="{middle}"{rel_attr}{target_attr}>{trim_url(middle)}</a>'
)
else:
middle = (
f'<a href="https://{middle}"{rel_attr}{target_attr}>'
f"{trim_url(middle)}</a>"
)
elif middle.startswith("mailto:") and _email_re.match(middle[7:]):
middle = f'<a href="{middle}">{middle[7:]}</a>'
elif (
"@" in middle
and not middle.startswith("www.")
and ":" not in middle
and _email_re.match(middle)
):
middle = f'<a href="mailto:{middle}">{middle}</a>'
elif extra_schemes is not None:
for scheme in extra_schemes:
if middle != scheme and middle.startswith(scheme):
middle = f'<a href="{middle}"{rel_attr}{target_attr}>{middle}</a>'
words[i] = f"{head}{middle}{tail}"
return "".join(words)
def generate_lorem_ipsum(
n: int = 5, html: bool = True, min: int = 20, max: int = 100
) -> str:
"""Generate some lorem ipsum for the template."""
from .constants import LOREM_IPSUM_WORDS
words = LOREM_IPSUM_WORDS.split()
result = []
for _ in range(n):
next_capitalized = True
last_comma = last_fullstop = 0
word = None
last = None
p = []
# each paragraph contains out of 20 to 100 words.
for idx, _ in enumerate(range(randrange(min, max))):
while True:
word = choice(words)
if word != last:
last = word
break
if next_capitalized:
word = word.capitalize()
next_capitalized = False
# add commas
if idx - randrange(3, 8) > last_comma:
last_comma = idx
last_fullstop += 2
word += ","
# add end of sentences
if idx - randrange(10, 20) > last_fullstop:
last_comma = last_fullstop = idx
word += "."
next_capitalized = True
p.append(word)
# ensure that the paragraph ends with a dot.
p_str = " ".join(p)
if p_str.endswith(","):
p_str = p_str[:-1] + "."
elif not p_str.endswith("."):
p_str += "."
result.append(p_str)
if not html:
return "\n\n".join(result)
return markupsafe.Markup(
"\n".join(f"<p>{markupsafe.escape(x)}</p>" for x in result)
)
def url_quote(obj: t.Any, charset: str = "utf-8", for_qs: bool = False) -> str:
"""Quote a string for use in a URL using the given charset.
:param obj: String or bytes to quote. Other types are converted to
string then encoded to bytes using the given charset.
:param charset: Encode text to bytes using this charset.
:param for_qs: Quote "/" and use "+" for spaces.
"""
if not isinstance(obj, bytes):
if not isinstance(obj, str):
obj = str(obj)
obj = obj.encode(charset)
safe = b"" if for_qs else b"/"
rv = quote_from_bytes(obj, safe)
if for_qs:
rv = rv.replace("%20", "+")
return rv
@abc.MutableMapping.register
class LRUCache:
"""A simple LRU Cache implementation."""
# this is fast for small capacities (something below 1000) but doesn't
# scale. But as long as it's only used as storage for templates this
# won't do any harm.
def __init__(self, capacity: int) -> None:
self.capacity = capacity
self._mapping: t.Dict[t.Any, t.Any] = {}
self._queue: "te.Deque[t.Any]" = deque()
self._postinit()
def _postinit(self) -> None:
# alias all queue methods for faster lookup
self._popleft = self._queue.popleft
self._pop = self._queue.pop
self._remove = self._queue.remove
self._wlock = Lock()
self._append = self._queue.append
def __getstate__(self) -> t.Mapping[str, t.Any]:
return {
"capacity": self.capacity,
"_mapping": self._mapping,
"_queue": self._queue,
}
def __setstate__(self, d: t.Mapping[str, t.Any]) -> None:
self.__dict__.update(d)
self._postinit()
def __getnewargs__(self) -> t.Tuple[t.Any, ...]:
return (self.capacity,)
def copy(self) -> "LRUCache":
"""Return a shallow copy of the instance."""
rv = self.__class__(self.capacity)
rv._mapping.update(self._mapping)
rv._queue.extend(self._queue)
return rv
def get(self, key: t.Any, default: t.Any = None) -> t.Any:
"""Return an item from the cache dict or `default`"""
try:
return self[key]
except KeyError:
return default
def setdefault(self, key: t.Any, default: t.Any = None) -> t.Any:
"""Set `default` if the key is not in the cache otherwise
leave unchanged. Return the value of this key.
"""
try:
return self[key]
except KeyError:
self[key] = default
return default
def clear(self) -> None:
"""Clear the cache."""
with self._wlock:
self._mapping.clear()
self._queue.clear()
def __contains__(self, key: t.Any) -> bool:
"""Check if a key exists in this cache."""
return key in self._mapping
def __len__(self) -> int:
"""Return the current size of the cache."""
return len(self._mapping)
def __repr__(self) -> str:
return f"<{type(self).__name__} {self._mapping!r}>"
def __getitem__(self, key: t.Any) -> t.Any:
"""Get an item from the cache. Moves the item up so that it has the
highest priority then.
Raise a `KeyError` if it does not exist.
"""
with self._wlock:
rv = self._mapping[key]
if self._queue[-1] != key:
try:
self._remove(key)
except ValueError:
# if something removed the key from the container
# when we read, ignore the ValueError that we would
# get otherwise.
pass
self._append(key)
return rv
def __setitem__(self, key: t.Any, value: t.Any) -> None:
"""Sets the value for an item. Moves the item up so that it
has the highest priority then.
"""
with self._wlock:
if key in self._mapping:
self._remove(key)
elif len(self._mapping) == self.capacity:
del self._mapping[self._popleft()]
self._append(key)
self._mapping[key] = value
def __delitem__(self, key: t.Any) -> None:
"""Remove an item from the cache dict.
Raise a `KeyError` if it does not exist.
"""
with self._wlock:
del self._mapping[key]
try:
self._remove(key)
except ValueError:
pass
def items(self) -> t.Iterable[t.Tuple[t.Any, t.Any]]:
"""Return a list of items."""
result = [(key, self._mapping[key]) for key in list(self._queue)]
result.reverse()
return result
def values(self) -> t.Iterable[t.Any]:
"""Return a list of all values."""
return [x[1] for x in self.items()]
def keys(self) -> t.Iterable[t.Any]:
"""Return a list of all keys ordered by most recent usage."""
return list(self)
def __iter__(self) -> t.Iterator[t.Any]:
return reversed(tuple(self._queue))
def __reversed__(self) -> t.Iterator[t.Any]:
"""Iterate over the keys in the cache dict, oldest items
coming first.
"""
return iter(tuple(self._queue))
__copy__ = copy
def select_autoescape(
enabled_extensions: t.Collection[str] = ("html", "htm", "xml"),
disabled_extensions: t.Collection[str] = (),
default_for_string: bool = True,
default: bool = False,
) -> t.Callable[[t.Optional[str]], bool]:
"""Intelligently sets the initial value of autoescaping based on the
filename of the template. This is the recommended way to configure
autoescaping if you do not want to write a custom function yourself.
If you want to enable it for all templates created from strings or
for all templates with `.html` and `.xml` extensions::
from jinja2 import Environment, select_autoescape
env = Environment(autoescape=select_autoescape(
enabled_extensions=('html', 'xml'),
default_for_string=True,
))
Example configuration to turn it on at all times except if the template
ends with `.txt`::
from jinja2 import Environment, select_autoescape
env = Environment(autoescape=select_autoescape(
disabled_extensions=('txt',),
default_for_string=True,
default=True,
))
The `enabled_extensions` is an iterable of all the extensions that
autoescaping should be enabled for. Likewise `disabled_extensions` is
a list of all templates it should be disabled for. If a template is
loaded from a string then the default from `default_for_string` is used.
If nothing matches then the initial value of autoescaping is set to the
value of `default`.
For security reasons this function operates case insensitive.
.. versionadded:: 2.9
"""
enabled_patterns = tuple(f".{x.lstrip('.').lower()}" for x in enabled_extensions)
disabled_patterns = tuple(f".{x.lstrip('.').lower()}" for x in disabled_extensions)
def autoescape(template_name: t.Optional[str]) -> bool:
if template_name is None:
return default_for_string
template_name = template_name.lower()
if template_name.endswith(enabled_patterns):
return True
if template_name.endswith(disabled_patterns):
return False
return default
return autoescape
def htmlsafe_json_dumps(
obj: t.Any, dumps: t.Optional[t.Callable[..., str]] = None, **kwargs: t.Any
) -> markupsafe.Markup:
"""Serialize an object to a string of JSON with :func:`json.dumps`,
then replace HTML-unsafe characters with Unicode escapes and mark
the result safe with :class:`~markupsafe.Markup`.
This is available in templates as the ``|tojson`` filter.
The following characters are escaped: ``<``, ``>``, ``&``, ``'``.
The returned string is safe to render in HTML documents and
``<script>`` tags. The exception is in HTML attributes that are
double quoted; either use single quotes or the ``|forceescape``
filter.
:param obj: The object to serialize to JSON.
:param dumps: The ``dumps`` function to use. Defaults to
``env.policies["json.dumps_function"]``, which defaults to
:func:`json.dumps`.
:param kwargs: Extra arguments to pass to ``dumps``. Merged onto
``env.policies["json.dumps_kwargs"]``.
.. versionchanged:: 3.0
The ``dumper`` parameter is renamed to ``dumps``.
.. versionadded:: 2.9
"""
if dumps is None:
dumps = json.dumps
return markupsafe.Markup(
dumps(obj, **kwargs)
.replace("<", "\\u003c")
.replace(">", "\\u003e")
.replace("&", "\\u0026")
.replace("'", "\\u0027")
)
class Cycler:
"""Cycle through values by yield them one at a time, then restarting
once the end is reached. Available as ``cycler`` in templates.
Similar to ``loop.cycle``, but can be used outside loops or across
multiple loops. For example, render a list of folders and files in a
list, alternating giving them "odd" and "even" classes.
.. code-block:: html+jinja
{% set row_class = cycler("odd", "even") %}
<ul class="browser">
{% for folder in folders %}
<li class="folder {{ row_class.next() }}">{{ folder }}
{% endfor %}
{% for file in files %}
<li class="file {{ row_class.next() }}">{{ file }}
{% endfor %}
</ul>
:param items: Each positional argument will be yielded in the order
given for each cycle.
.. versionadded:: 2.1
"""
def __init__(self, *items: t.Any) -> None:
if not items:
raise RuntimeError("at least one item has to be provided")
self.items = items
self.pos = 0
def reset(self) -> None:
"""Resets the current item to the first item."""
self.pos = 0
@property
def current(self) -> t.Any:
"""Return the current item. Equivalent to the item that will be
returned next time :meth:`next` is called.
"""
return self.items[self.pos]
def next(self) -> t.Any:
"""Return the current item, then advance :attr:`current` to the
next item.
"""
rv = self.current
self.pos = (self.pos + 1) % len(self.items)
return rv
__next__ = next
class Joiner:
"""A joining helper for templates."""
def __init__(self, sep: str = ", ") -> None:
self.sep = sep
self.used = False
def __call__(self) -> str:
if not self.used:
self.used = True
return ""
return self.sep
class Namespace:
"""A namespace object that can hold arbitrary attributes. It may be
initialized from a dictionary or with keyword arguments."""
def __init__(*args: t.Any, **kwargs: t.Any) -> None: # noqa: B902
self, args = args[0], args[1:]
self.__attrs = dict(*args, **kwargs)
def __getattribute__(self, name: str) -> t.Any:
# __class__ is needed for the awaitable check in async mode
if name in {"_Namespace__attrs", "__class__"}:
return object.__getattribute__(self, name)
try:
return self.__attrs[name]
except KeyError:
raise AttributeError(name) from None
def __setitem__(self, name: str, value: t.Any) -> None:
self.__attrs[name] = value
def __repr__(self) -> str:
return f"<Namespace {self.__attrs!r}>"
"""API for traversing the AST nodes. Implemented by the compiler and
meta introspection.
"""
import typing as t
from .nodes import Node
if t.TYPE_CHECKING:
import typing_extensions as te
class VisitCallable(te.Protocol):
def __call__(self, node: Node, *args: t.Any, **kwargs: t.Any) -> t.Any: ...
class NodeVisitor:
"""Walks the abstract syntax tree and call visitor functions for every
node found. The visitor functions may return values which will be
forwarded by the `visit` method.
Per default the visitor functions for the nodes are ``'visit_'`` +
class name of the node. So a `TryFinally` node visit function would
be `visit_TryFinally`. This behavior can be changed by overriding
the `get_visitor` function. If no visitor function exists for a node
(return value `None`) the `generic_visit` visitor is used instead.
"""
def get_visitor(self, node: Node) -> "t.Optional[VisitCallable]":
"""Return the visitor function for this node or `None` if no visitor
exists for this node. In that case the generic visit function is
used instead.
"""
return getattr(self, f"visit_{type(node).__name__}", None)
def visit(self, node: Node, *args: t.Any, **kwargs: t.Any) -> t.Any:
"""Visit a node."""
f = self.get_visitor(node)
if f is not None:
return f(node, *args, **kwargs)
return self.generic_visit(node, *args, **kwargs)
def generic_visit(self, node: Node, *args: t.Any, **kwargs: t.Any) -> t.Any:
"""Called if no explicit visitor function exists for a node."""
for child_node in node.iter_child_nodes():
self.visit(child_node, *args, **kwargs)
class NodeTransformer(NodeVisitor):
"""Walks the abstract syntax tree and allows modifications of nodes.
The `NodeTransformer` will walk the AST and use the return value of the
visitor functions to replace or remove the old node. If the return
value of the visitor function is `None` the node will be removed
from the previous location otherwise it's replaced with the return
value. The return value may be the original node in which case no
replacement takes place.
"""
def generic_visit(self, node: Node, *args: t.Any, **kwargs: t.Any) -> Node:
for field, old_value in node.iter_fields():
if isinstance(old_value, list):
new_values = []
for value in old_value:
if isinstance(value, Node):
value = self.visit(value, *args, **kwargs)
if value is None:
continue
elif not isinstance(value, Node):
new_values.extend(value)
continue
new_values.append(value)
old_value[:] = new_values
elif isinstance(old_value, Node):
new_node = self.visit(old_value, *args, **kwargs)
if new_node is None:
delattr(node, field)
else:
setattr(node, field, new_node)
return node
def visit_list(self, node: Node, *args: t.Any, **kwargs: t.Any) -> t.List[Node]:
"""As transformers may return lists in some places this method
can be used to enforce a list as return value.
"""
rv = self.visit(node, *args, **kwargs)
if not isinstance(rv, list):
return [rv]
return rv
import functools
import string
import sys
import typing as t
if t.TYPE_CHECKING:
import typing_extensions as te
class HasHTML(te.Protocol):
def __html__(self) -> str:
pass
_P = te.ParamSpec("_P")
__version__ = "2.1.5"
def _simple_escaping_wrapper(func: "t.Callable[_P, str]") -> "t.Callable[_P, Markup]":
@functools.wraps(func)
def wrapped(self: "Markup", *args: "_P.args", **kwargs: "_P.kwargs") -> "Markup":
arg_list = _escape_argspec(list(args), enumerate(args), self.escape)
_escape_argspec(kwargs, kwargs.items(), self.escape)
return self.__class__(func(self, *arg_list, **kwargs)) # type: ignore[arg-type]
return wrapped # type: ignore[return-value]
class Markup(str):
"""A string that is ready to be safely inserted into an HTML or XML
document, either because it was escaped or because it was marked
safe.
Passing an object to the constructor converts it to text and wraps
it to mark it safe without escaping. To escape the text, use the
:meth:`escape` class method instead.
>>> Markup("Hello, <em>World</em>!")
Markup('Hello, <em>World</em>!')
>>> Markup(42)
Markup('42')
>>> Markup.escape("Hello, <em>World</em>!")
Markup('Hello &lt;em&gt;World&lt;/em&gt;!')
This implements the ``__html__()`` interface that some frameworks
use. Passing an object that implements ``__html__()`` will wrap the
output of that method, marking it safe.
>>> class Foo:
... def __html__(self):
... return '<a href="/foo">foo</a>'
...
>>> Markup(Foo())
Markup('<a href="/foo">foo</a>')
This is a subclass of :class:`str`. It has the same methods, but
escapes their arguments and returns a ``Markup`` instance.
>>> Markup("<em>%s</em>") % ("foo & bar",)
Markup('<em>foo &amp; bar</em>')
>>> Markup("<em>Hello</em> ") + "<foo>"
Markup('<em>Hello</em> &lt;foo&gt;')
"""
__slots__ = ()
def __new__(
cls, base: t.Any = "", encoding: t.Optional[str] = None, errors: str = "strict"
) -> "te.Self":
if hasattr(base, "__html__"):
base = base.__html__()
if encoding is None:
return super().__new__(cls, base)
return super().__new__(cls, base, encoding, errors)
def __html__(self) -> "te.Self":
return self
def __add__(self, other: t.Union[str, "HasHTML"]) -> "te.Self":
if isinstance(other, str) or hasattr(other, "__html__"):
return self.__class__(super().__add__(self.escape(other)))
return NotImplemented
def __radd__(self, other: t.Union[str, "HasHTML"]) -> "te.Self":
if isinstance(other, str) or hasattr(other, "__html__"):
return self.escape(other).__add__(self)
return NotImplemented
def __mul__(self, num: "te.SupportsIndex") -> "te.Self":
if isinstance(num, int):
return self.__class__(super().__mul__(num))
return NotImplemented
__rmul__ = __mul__
def __mod__(self, arg: t.Any) -> "te.Self":
if isinstance(arg, tuple):
# a tuple of arguments, each wrapped
arg = tuple(_MarkupEscapeHelper(x, self.escape) for x in arg)
elif hasattr(type(arg), "__getitem__") and not isinstance(arg, str):
# a mapping of arguments, wrapped
arg = _MarkupEscapeHelper(arg, self.escape)
else:
# a single argument, wrapped with the helper and a tuple
arg = (_MarkupEscapeHelper(arg, self.escape),)
return self.__class__(super().__mod__(arg))
def __repr__(self) -> str:
return f"{self.__class__.__name__}({super().__repr__()})"
def join(self, seq: t.Iterable[t.Union[str, "HasHTML"]]) -> "te.Self":
return self.__class__(super().join(map(self.escape, seq)))
join.__doc__ = str.join.__doc__
def split( # type: ignore[override]
self, sep: t.Optional[str] = None, maxsplit: int = -1
) -> t.List["te.Self"]:
return [self.__class__(v) for v in super().split(sep, maxsplit)]
split.__doc__ = str.split.__doc__
def rsplit( # type: ignore[override]
self, sep: t.Optional[str] = None, maxsplit: int = -1
) -> t.List["te.Self"]:
return [self.__class__(v) for v in super().rsplit(sep, maxsplit)]
rsplit.__doc__ = str.rsplit.__doc__
def splitlines( # type: ignore[override]
self, keepends: bool = False
) -> t.List["te.Self"]:
return [self.__class__(v) for v in super().splitlines(keepends)]
splitlines.__doc__ = str.splitlines.__doc__
def unescape(self) -> str:
"""Convert escaped markup back into a text string. This replaces
HTML entities with the characters they represent.
>>> Markup("Main &raquo; <em>About</em>").unescape()
'Main » <em>About</em>'
"""
from html import unescape
return unescape(str(self))
def striptags(self) -> str:
""":meth:`unescape` the markup, remove tags, and normalize
whitespace to single spaces.
>>> Markup("Main &raquo;\t<em>About</em>").striptags()
'Main » About'
"""
value = str(self)
# Look for comments then tags separately. Otherwise, a comment that
# contains a tag would end early, leaving some of the comment behind.
while True:
# keep finding comment start marks
start = value.find("<!--")
if start == -1:
break
# find a comment end mark beyond the start, otherwise stop
end = value.find("-->", start)
if end == -1:
break
value = f"{value[:start]}{value[end + 3:]}"
# remove tags using the same method
while True:
start = value.find("<")
if start == -1:
break
end = value.find(">", start)
if end == -1:
break
value = f"{value[:start]}{value[end + 1:]}"
# collapse spaces
value = " ".join(value.split())
return self.__class__(value).unescape()
@classmethod
def escape(cls, s: t.Any) -> "te.Self":
"""Escape a string. Calls :func:`escape` and ensures that for
subclasses the correct type is returned.
"""
rv = escape(s)
if rv.__class__ is not cls:
return cls(rv)
return rv # type: ignore[return-value]
__getitem__ = _simple_escaping_wrapper(str.__getitem__)
capitalize = _simple_escaping_wrapper(str.capitalize)
title = _simple_escaping_wrapper(str.title)
lower = _simple_escaping_wrapper(str.lower)
upper = _simple_escaping_wrapper(str.upper)
replace = _simple_escaping_wrapper(str.replace)
ljust = _simple_escaping_wrapper(str.ljust)
rjust = _simple_escaping_wrapper(str.rjust)
lstrip = _simple_escaping_wrapper(str.lstrip)
rstrip = _simple_escaping_wrapper(str.rstrip)
center = _simple_escaping_wrapper(str.center)
strip = _simple_escaping_wrapper(str.strip)
translate = _simple_escaping_wrapper(str.translate)
expandtabs = _simple_escaping_wrapper(str.expandtabs)
swapcase = _simple_escaping_wrapper(str.swapcase)
zfill = _simple_escaping_wrapper(str.zfill)
casefold = _simple_escaping_wrapper(str.casefold)
if sys.version_info >= (3, 9):
removeprefix = _simple_escaping_wrapper(str.removeprefix)
removesuffix = _simple_escaping_wrapper(str.removesuffix)
def partition(self, sep: str) -> t.Tuple["te.Self", "te.Self", "te.Self"]:
l, s, r = super().partition(self.escape(sep))
cls = self.__class__
return cls(l), cls(s), cls(r)
def rpartition(self, sep: str) -> t.Tuple["te.Self", "te.Self", "te.Self"]:
l, s, r = super().rpartition(self.escape(sep))
cls = self.__class__
return cls(l), cls(s), cls(r)
def format(self, *args: t.Any, **kwargs: t.Any) -> "te.Self":
formatter = EscapeFormatter(self.escape)
return self.__class__(formatter.vformat(self, args, kwargs))
def format_map( # type: ignore[override]
self, map: t.Mapping[str, t.Any]
) -> "te.Self":
formatter = EscapeFormatter(self.escape)
return self.__class__(formatter.vformat(self, (), map))
def __html_format__(self, format_spec: str) -> "te.Self":
if format_spec:
raise ValueError("Unsupported format specification for Markup.")
return self
class EscapeFormatter(string.Formatter):
__slots__ = ("escape",)
def __init__(self, escape: t.Callable[[t.Any], Markup]) -> None:
self.escape = escape
super().__init__()
def format_field(self, value: t.Any, format_spec: str) -> str:
if hasattr(value, "__html_format__"):
rv = value.__html_format__(format_spec)
elif hasattr(value, "__html__"):
if format_spec:
raise ValueError(
f"Format specifier {format_spec} given, but {type(value)} does not"
" define __html_format__. A class that defines __html__ must define"
" __html_format__ to work with format specifiers."
)
rv = value.__html__()
else:
# We need to make sure the format spec is str here as
# otherwise the wrong callback methods are invoked.
rv = string.Formatter.format_field(self, value, str(format_spec))
return str(self.escape(rv))
_ListOrDict = t.TypeVar("_ListOrDict", list, dict)
def _escape_argspec(
obj: _ListOrDict, iterable: t.Iterable[t.Any], escape: t.Callable[[t.Any], Markup]
) -> _ListOrDict:
"""Helper for various string-wrapped functions."""
for key, value in iterable:
if isinstance(value, str) or hasattr(value, "__html__"):
obj[key] = escape(value)
return obj
class _MarkupEscapeHelper:
"""Helper for :meth:`Markup.__mod__`."""
__slots__ = ("obj", "escape")
def __init__(self, obj: t.Any, escape: t.Callable[[t.Any], Markup]) -> None:
self.obj = obj
self.escape = escape
def __getitem__(self, item: t.Any) -> "te.Self":
return self.__class__(self.obj[item], self.escape)
def __str__(self) -> str:
return str(self.escape(self.obj))
def __repr__(self) -> str:
return str(self.escape(repr(self.obj)))
def __int__(self) -> int:
return int(self.obj)
def __float__(self) -> float:
return float(self.obj)
# circular import
try:
from ._speedups import escape as escape
from ._speedups import escape_silent as escape_silent
from ._speedups import soft_str as soft_str
except ImportError:
from ._native import escape as escape
from ._native import escape_silent as escape_silent # noqa: F401
from ._native import soft_str as soft_str # noqa: F401
import typing as t
from . import Markup
def escape(s: t.Any) -> Markup:
"""Replace the characters ``&``, ``<``, ``>``, ``'``, and ``"`` in
the string with HTML-safe sequences. Use this if you need to display
text that might contain such characters in HTML.
If the object has an ``__html__`` method, it is called and the
return value is assumed to already be safe for HTML.
:param s: An object to be converted to a string and escaped.
:return: A :class:`Markup` string with the escaped text.
"""
if hasattr(s, "__html__"):
return Markup(s.__html__())
return Markup(
str(s)
.replace("&", "&amp;")
.replace(">", "&gt;")
.replace("<", "&lt;")
.replace("'", "&#39;")
.replace('"', "&#34;")
)
def escape_silent(s: t.Optional[t.Any]) -> Markup:
"""Like :func:`escape` but treats ``None`` as the empty string.
Useful with optional values, as otherwise you get the string
``'None'`` when the value is ``None``.
>>> escape(None)
Markup('None')
>>> escape_silent(None)
Markup('')
"""
if s is None:
return Markup()
return escape(s)
def soft_str(s: t.Any) -> str:
"""Convert an object to a string if it isn't already. This preserves
a :class:`Markup` string rather than converting it back to a basic
string, so it will still be marked as safe and won't be escaped
again.
>>> value = escape("<User 1>")
>>> value
Markup('&lt;User 1&gt;')
>>> escape(str(value))
Markup('&amp;lt;User 1&amp;gt;')
>>> escape(soft_str(value))
Markup('&lt;User 1&gt;')
"""
if not isinstance(s, str):
return str(s)
return s
#include <Python.h>
static PyObject* markup;
static int
init_constants(void)
{
PyObject *module;
/* import markup type so that we can mark the return value */
module = PyImport_ImportModule("markupsafe");
if (!module)
return 0;
markup = PyObject_GetAttrString(module, "Markup");
Py_DECREF(module);
return 1;
}
#define GET_DELTA(inp, inp_end, delta) \
while (inp < inp_end) { \
switch (*inp++) { \
case '"': \
case '\'': \
case '&': \
delta += 4; \
break; \
case '<': \
case '>': \
delta += 3; \
break; \
} \
}
#define DO_ESCAPE(inp, inp_end, outp) \
{ \
Py_ssize_t ncopy = 0; \
while (inp < inp_end) { \
switch (*inp) { \
case '"': \
memcpy(outp, inp-ncopy, sizeof(*outp)*ncopy); \
outp += ncopy; ncopy = 0; \
*outp++ = '&'; \
*outp++ = '#'; \
*outp++ = '3'; \
*outp++ = '4'; \
*outp++ = ';'; \
break; \
case '\'': \
memcpy(outp, inp-ncopy, sizeof(*outp)*ncopy); \
outp += ncopy; ncopy = 0; \
*outp++ = '&'; \
*outp++ = '#'; \
*outp++ = '3'; \
*outp++ = '9'; \
*outp++ = ';'; \
break; \
case '&': \
memcpy(outp, inp-ncopy, sizeof(*outp)*ncopy); \
outp += ncopy; ncopy = 0; \
*outp++ = '&'; \
*outp++ = 'a'; \
*outp++ = 'm'; \
*outp++ = 'p'; \
*outp++ = ';'; \
break; \
case '<': \
memcpy(outp, inp-ncopy, sizeof(*outp)*ncopy); \
outp += ncopy; ncopy = 0; \
*outp++ = '&'; \
*outp++ = 'l'; \
*outp++ = 't'; \
*outp++ = ';'; \
break; \
case '>': \
memcpy(outp, inp-ncopy, sizeof(*outp)*ncopy); \
outp += ncopy; ncopy = 0; \
*outp++ = '&'; \
*outp++ = 'g'; \
*outp++ = 't'; \
*outp++ = ';'; \
break; \
default: \
ncopy++; \
} \
inp++; \
} \
memcpy(outp, inp-ncopy, sizeof(*outp)*ncopy); \
}
static PyObject*
escape_unicode_kind1(PyUnicodeObject *in)
{
Py_UCS1 *inp = PyUnicode_1BYTE_DATA(in);
Py_UCS1 *inp_end = inp + PyUnicode_GET_LENGTH(in);
Py_UCS1 *outp;
PyObject *out;
Py_ssize_t delta = 0;
GET_DELTA(inp, inp_end, delta);
if (!delta) {
Py_INCREF(in);
return (PyObject*)in;
}
out = PyUnicode_New(PyUnicode_GET_LENGTH(in) + delta,
PyUnicode_IS_ASCII(in) ? 127 : 255);
if (!out)
return NULL;
inp = PyUnicode_1BYTE_DATA(in);
outp = PyUnicode_1BYTE_DATA(out);
DO_ESCAPE(inp, inp_end, outp);
return out;
}
static PyObject*
escape_unicode_kind2(PyUnicodeObject *in)
{
Py_UCS2 *inp = PyUnicode_2BYTE_DATA(in);
Py_UCS2 *inp_end = inp + PyUnicode_GET_LENGTH(in);
Py_UCS2 *outp;
PyObject *out;
Py_ssize_t delta = 0;
GET_DELTA(inp, inp_end, delta);
if (!delta) {
Py_INCREF(in);
return (PyObject*)in;
}
out = PyUnicode_New(PyUnicode_GET_LENGTH(in) + delta, 65535);
if (!out)
return NULL;
inp = PyUnicode_2BYTE_DATA(in);
outp = PyUnicode_2BYTE_DATA(out);
DO_ESCAPE(inp, inp_end, outp);
return out;
}
static PyObject*
escape_unicode_kind4(PyUnicodeObject *in)
{
Py_UCS4 *inp = PyUnicode_4BYTE_DATA(in);
Py_UCS4 *inp_end = inp + PyUnicode_GET_LENGTH(in);
Py_UCS4 *outp;
PyObject *out;
Py_ssize_t delta = 0;
GET_DELTA(inp, inp_end, delta);
if (!delta) {
Py_INCREF(in);
return (PyObject*)in;
}
out = PyUnicode_New(PyUnicode_GET_LENGTH(in) + delta, 1114111);
if (!out)
return NULL;
inp = PyUnicode_4BYTE_DATA(in);
outp = PyUnicode_4BYTE_DATA(out);
DO_ESCAPE(inp, inp_end, outp);
return out;
}
static PyObject*
escape_unicode(PyUnicodeObject *in)
{
if (PyUnicode_READY(in))
return NULL;
switch (PyUnicode_KIND(in)) {
case PyUnicode_1BYTE_KIND:
return escape_unicode_kind1(in);
case PyUnicode_2BYTE_KIND:
return escape_unicode_kind2(in);
case PyUnicode_4BYTE_KIND:
return escape_unicode_kind4(in);
}
assert(0); /* shouldn't happen */
return NULL;
}
static PyObject*
escape(PyObject *self, PyObject *text)
{
static PyObject *id_html;
PyObject *s = NULL, *rv = NULL, *html;
if (id_html == NULL) {
id_html = PyUnicode_InternFromString("__html__");
if (id_html == NULL) {
return NULL;
}
}
/* we don't have to escape integers, bools or floats */
if (PyLong_CheckExact(text) ||
PyFloat_CheckExact(text) || PyBool_Check(text) ||
text == Py_None)
return PyObject_CallFunctionObjArgs(markup, text, NULL);
/* if the object has an __html__ method that performs the escaping */
html = PyObject_GetAttr(text ,id_html);
if (html) {
s = PyObject_CallObject(html, NULL);
Py_DECREF(html);
if (s == NULL) {
return NULL;
}
/* Convert to Markup object */
rv = PyObject_CallFunctionObjArgs(markup, (PyObject*)s, NULL);
Py_DECREF(s);
return rv;
}
/* otherwise make the object unicode if it isn't, then escape */
PyErr_Clear();
if (!PyUnicode_Check(text)) {
PyObject *unicode = PyObject_Str(text);
if (!unicode)
return NULL;
s = escape_unicode((PyUnicodeObject*)unicode);
Py_DECREF(unicode);
}
else
s = escape_unicode((PyUnicodeObject*)text);
/* convert the unicode string into a markup object. */
rv = PyObject_CallFunctionObjArgs(markup, (PyObject*)s, NULL);
Py_DECREF(s);
return rv;
}
static PyObject*
escape_silent(PyObject *self, PyObject *text)
{
if (text != Py_None)
return escape(self, text);
return PyObject_CallFunctionObjArgs(markup, NULL);
}
static PyObject*
soft_str(PyObject *self, PyObject *s)
{
if (!PyUnicode_Check(s))
return PyObject_Str(s);
Py_INCREF(s);
return s;
}
static PyMethodDef module_methods[] = {
{
"escape",
(PyCFunction)escape,
METH_O,
"Replace the characters ``&``, ``<``, ``>``, ``'``, and ``\"`` in"
" the string with HTML-safe sequences. Use this if you need to display"
" text that might contain such characters in HTML.\n\n"
"If the object has an ``__html__`` method, it is called and the"
" return value is assumed to already be safe for HTML.\n\n"
":param s: An object to be converted to a string and escaped.\n"
":return: A :class:`Markup` string with the escaped text.\n"
},
{
"escape_silent",
(PyCFunction)escape_silent,
METH_O,
"Like :func:`escape` but treats ``None`` as the empty string."
" Useful with optional values, as otherwise you get the string"
" ``'None'`` when the value is ``None``.\n\n"
">>> escape(None)\n"
"Markup('None')\n"
">>> escape_silent(None)\n"
"Markup('')\n"
},
{
"soft_str",
(PyCFunction)soft_str,
METH_O,
"Convert an object to a string if it isn't already. This preserves"
" a :class:`Markup` string rather than converting it back to a basic"
" string, so it will still be marked as safe and won't be escaped"
" again.\n\n"
">>> value = escape(\"<User 1>\")\n"
">>> value\n"
"Markup('&lt;User 1&gt;')\n"
">>> escape(str(value))\n"
"Markup('&amp;lt;User 1&amp;gt;')\n"
">>> escape(soft_str(value))\n"
"Markup('&lt;User 1&gt;')\n"
},
{NULL, NULL, 0, NULL} /* Sentinel */
};
static struct PyModuleDef module_definition = {
PyModuleDef_HEAD_INIT,
"markupsafe._speedups",
NULL,
-1,
module_methods,
NULL,
NULL,
NULL,
NULL
};
PyMODINIT_FUNC
PyInit__speedups(void)
{
if (!init_constants())
return NULL;
return PyModule_Create(&module_definition);
}
from typing import Any
from typing import Optional
from . import Markup
def escape(s: Any) -> Markup: ...
def escape_silent(s: Optional[Any]) -> Markup: ...
def soft_str(s: Any) -> str: ...
def soft_unicode(s: Any) -> str: ...
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment