Output OSM table

49ec3120 · Eric Duminil · 2f4477c0 · 49ec3120
Commit 49ec3120 authored Nov 20, 2024 by Eric Duminil
--- a/python_scripts/add_trees_to_open_street_map/Parse OSM data.ipynb
+++ b/python_scripts/add_trees_to_open_street_map/Parse OSM data.ipynb
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "4888de6c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from pathlib import Path"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "bb724cab",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pickle"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "b584ca38",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "e74c7f88",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from pyproj import Transformer\n",
+    "from shapely import LineString"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "9eac1a09",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "EPSG_ID = 25832\n",
+    "to_local = Transformer.from_crs(\"EPSG:4326\", f\"EPSG:{EPSG_ID}\", always_xy=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "0fed9052",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "pd.set_option('display.max_columns', None)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "89e9aff8",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "[WindowsPath('cache/48_864071__48_866928__9_196641__9_203522.pickle'),\n",
+       " WindowsPath('cache/49_80081__49_803063__9_947011__9_95511.pickle'),\n",
+       " WindowsPath('cache/49_802665__49_804083__9_942116__9_945581.pickle'),\n",
+       " WindowsPath('cache/49_802698__49_804055__9_942167__9_945563.pickle'),\n",
+       " WindowsPath('cache/49_802704__49_803486__9_944034__9_945488.pickle')]"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "list(Path('cache').glob('*.pickle'))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "8aac3681",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def load_cache(cache_path):\n",
+    "    with open(Path('cache') / '49_802698__49_804055__9_942167__9_945563.pickle', 'rb') as f:\n",
+    "        return pickle.load(f)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "d050a116",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def get_way_dict(way):\n",
+    "    # Length\n",
+    "    url = f'https://www.openstreetmap.org/way/{way.id}'\n",
+    "    xy_s = [to_local.transform(node.lon, node.lat) for node in way.nodes]\n",
+    "    path = LineString(xy_s)\n",
+    "    other_info = {'id': way.id,\n",
+    "                  'URL': f'=HYPERLINK(\"{url}\"; \"{url}\")',\n",
+    "                  'length [m]': round(path.length, 1)}\n",
+    "    return other_info | way.tags"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "7a9f108b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "numerical_columns = ['maxspeed [km/h]', 'width [m]', 'length [m]']"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "741fa825",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def create_table(ways):\n",
+    "    df = pd.DataFrame([get_way_dict(way) for way in ways]).set_index('id').rename(columns={'width': 'width [m]', 'maxspeed': 'maxspeed [km/h]'})\n",
+    "    for col in numerical_columns:\n",
+    "        df[col] = pd.to_numeric(df[col])\n",
+    "    df.fillna('')\n",
+    "    return df"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "a12d5684",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "for cache_path in Path('cache').glob('*.pickle'):\n",
+    "    df = create_table(load_cache(cache_path))\n",
+    "    df.to_csv(Path('output') / f'OSM_infos_{cache_path.stem}.csv',\n",
+    "          sep=';',\n",
+    "          decimal=',',\n",
+    "          encoding='utf-8-sig'\n",
+    "    )"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "328d048f",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.5"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
+%% Cell type:code id:4888de6c tags:
+
+``` python
+from pathlib import Path
+```
+
+%% Cell type:code id:bb724cab tags:
+
+``` python
+import pickle
+```
+
+%% Cell type:code id:b584ca38 tags:
+
+``` python
+import pandas as pd
+```
+
+%% Cell type:code id:e74c7f88 tags:
+
+``` python
+from pyproj import Transformer
+from shapely import LineString
+```
+
+%% Cell type:code id:9eac1a09 tags:
+
+``` python
+EPSG_ID = 25832
+to_local = Transformer.from_crs("EPSG:4326", f"EPSG:{EPSG_ID}", always_xy=True)
+```
+
+%% Cell type:code id:0fed9052 tags:
+
+``` python
+pd.set_option('display.max_columns', None)
+```
+
+%% Cell type:code id:89e9aff8 tags:
+
+``` python
+list(Path('cache').glob('*.pickle'))
+```
+
+%% Output
+
+    [WindowsPath('cache/48_864071__48_866928__9_196641__9_203522.pickle'),
+     WindowsPath('cache/49_80081__49_803063__9_947011__9_95511.pickle'),
+     WindowsPath('cache/49_802665__49_804083__9_942116__9_945581.pickle'),
+     WindowsPath('cache/49_802698__49_804055__9_942167__9_945563.pickle'),
+     WindowsPath('cache/49_802704__49_803486__9_944034__9_945488.pickle')]
+
+%% Cell type:code id:8aac3681 tags:
+
+``` python
+def load_cache(cache_path):
+    with open(Path('cache') / '49_802698__49_804055__9_942167__9_945563.pickle', 'rb') as f:
+        return pickle.load(f)
+```
+
+%% Cell type:code id:d050a116 tags:
+
+``` python
+def get_way_dict(way):
+    # Length
+    url = f'https://www.openstreetmap.org/way/{way.id}'
+    xy_s = [to_local.transform(node.lon, node.lat) for node in way.nodes]
+    path = LineString(xy_s)
+    other_info = {'id': way.id,
+                  'URL': f'=HYPERLINK("{url}"; "{url}")',
+                  'length [m]': round(path.length, 1)}
+    return other_info | way.tags
+```
+
+%% Cell type:code id:7a9f108b tags:
+
+``` python
+numerical_columns = ['maxspeed [km/h]', 'width [m]', 'length [m]']
+```
+
+%% Cell type:code id:741fa825 tags:
+
+``` python
+def create_table(ways):
+    df = pd.DataFrame([get_way_dict(way) for way in ways]).set_index('id').rename(columns={'width': 'width [m]', 'maxspeed': 'maxspeed [km/h]'})
+    for col in numerical_columns:
+        df[col] = pd.to_numeric(df[col])
+    df.fillna('')
+    return df
+```
+
+%% Cell type:code id:a12d5684 tags:
+
+``` python
+for cache_path in Path('cache').glob('*.pickle'):
+    df = create_table(load_cache(cache_path))
+    df.to_csv(Path('output') / f'OSM_infos_{cache_path.stem}.csv',
+          sep=';',
+          decimal=',',
+          encoding='utf-8-sig'
+    )
+```
+
+%% Cell type:code id:328d048f tags:
+
+``` python
+```