define Bundesland

02_bayern.py

@@ -13,9 +13,13 @@ import random
 import xml.etree.ElementTree as ET
 from pathlib import Path
 
-from citygml_download import CityGMLWithHash, download_all_files, download_file, SCRIPT_DIR, TMP_DIR
+from citygml_download import TMP_DIR, Bundesland, CityGMLWithHash, download_all_files, download_file
 
-BAYERN_METALINK = "https://geodaten.bayern.de/odd/a/lod2/citygml/meta/metalink/09.meta4"
+BAYERN = Bundesland(
+    "Bayern",
+    source="https://geodaten.bayern.de/odd/a/lod2/citygml/meta/metalink/09.meta4",
+    info="https://www.ldbv.bayern.de/",
+)
 
 
 def download_metalink(metalink_url: str, tmp_path: Path = TMP_DIR) -> Path:
@@ -27,7 +31,7 @@ def download_metalink(metalink_url: str, tmp_path: Path = TMP_DIR) -> Path:
     return metalink_path
 
 
-def parse_metalink(metalink_path: Path, download_dir: Path) -> list[CityGMLWithHash]:
+def parse_metalink(metalink_path: Path, bundesland: Bundesland) -> list[CityGMLWithHash]:
     """Parse metalink XML file and return CityGMLWithHash objects."""
     print(f"Parsing metalink file: {metalink_path}")
     tree = ET.parse(metalink_path)
@@ -58,10 +62,8 @@ def parse_metalink(metalink_path: Path, download_dir: Path) -> list[CityGMLWithH
         files.append(
             CityGMLWithHash(
                 url=url,
-                path=download_dir / filename,
                 coordinate_reference_system="EPSG:25832",
-                source="https://geodaten.bayern.de",
-                bundesland="Bayern",
+                bundesland=bundesland,
                 expected_sha256=hash_node.text,
             )
         )
@@ -70,13 +72,9 @@ def parse_metalink(metalink_path: Path, download_dir: Path) -> list[CityGMLWithH
 
 
 if __name__ == "__main__":
-    # Should be from CityGML
-    download_dir = SCRIPT_DIR / "citygml" / "bayern"
 
-    download_dir.mkdir(parents=True, exist_ok=True)
-
-    metalink_path = download_metalink(BAYERN_METALINK)
-    citygmls = parse_metalink(metalink_path, download_dir)
+    metalink_path = download_metalink(BAYERN.source)
+    citygmls = parse_metalink(metalink_path, BAYERN)
 
     # Download all files
     download_all_files(citygmls)

citygml_download.py

@@ -18,6 +18,7 @@ from urllib.request import urlopen
 
 SCRIPT_DIR = Path(__file__).resolve().parent
 TMP_DIR = SCRIPT_DIR / "tmp"
+DOWNLOAD_DIR = SCRIPT_DIR / "citygml"
 TMP_DIR.mkdir(parents=True, exist_ok=True)
 
 
@@ -74,6 +75,21 @@ def download_file(url: str, dest_path: Path, sleep: int = 0, chunk_size: int = 8
         return False
 
 
+@dataclass
+class Bundesland:
+    name: str
+    source: str ="Unknown"
+    info: str ="Unknown"
+    license: str = "CC BY 4.0."
+
+    def __post_init__(self):
+        self.download_folder.mkdir(parents=True, exist_ok=True)
+
+    @property
+    def download_folder(self) -> Path:
+        return DOWNLOAD_DIR / self.name.lower().replace("-", "_").replace("ü", "ue")
+
+
 @dataclass
 class CityGML(ABC):
     """
@@ -84,11 +100,16 @@ class CityGML(ABC):
     """
 
     url: str
-    path: Path
-    bundesland: str
+    bundesland: Bundesland
     coordinate_reference_system: str = "Unknown"
-    source: str = "Unknown"
-    license: str = "CC BY 4.0."
+
+    @property
+    def download_folder(self) -> Path:
+        return self.bundesland.download_folder
+
+    @property
+    def path(self) -> Path:
+        return self.download_folder / self.filename
 
     @property
     def filename(self) -> str:
@@ -180,7 +201,6 @@ class CityGML(ABC):
 
             if self.verify(tmp_path):
                 # Move to final location
-                self.path.parent.mkdir(parents=True, exist_ok=True)
                 tmp_path.rename(self.path)
                 print(f"✓ {self.filename} downloaded and verified successfully")
                 return True