Merge pull request #48 from openstreetmap-polska/dev

Release to main

api/v1/countries.py

@@ -3,7 +3,7 @@
 
 import anyio
 from anyio.streams.memory import MemoryObjectSendStream
-from fastapi import APIRouter, Path, Request, Response
+from fastapi import APIRouter, Path, Response
 from shapely.geometry import mapping
 
 from middlewares.cache_middleware import configure_cache
@@ -22,7 +22,6 @@ async def _count_aed_in_country(country: Country, aed_state: AEDState, send_stre
 @router.get('/names')
 @configure_cache(timedelta(hours=1), stale=timedelta(days=7))
 async def get_names(
-    request: Request,
     country_state: CountryStateDep,
     aed_state: AEDStateDep,
     language: str | None = None,
@@ -66,7 +65,6 @@ def limit_country_names(names: dict[str, str]):
 @router.get('/{country_code}.geojson')
 @configure_cache(timedelta(hours=1), stale=timedelta(seconds=0))
 async def get_geojson(
-    request: Request,
     response: Response,
     country_code: Annotated[str, Path(min_length=2, max_length=5)],
     country_state: CountryStateDep,

api/v1/node.py

@@ -1,11 +1,12 @@
 import re
-from datetime import datetime
+from datetime import datetime, timedelta
 from urllib.parse import quote_plus
 
 from fastapi import APIRouter, HTTPException
 from pytz import timezone
 from tzfpy import get_tz
 
+from middlewares.cache_middleware import configure_cache
 from models.lonlat import LonLat
 from states.aed_state import AEDStateDep
 from states.photo_state import PhotoStateDep
@@ -30,42 +31,56 @@ def _get_timezone(lonlat: LonLat) -> tuple[str | None, str | None]:
     return timezone_name, timezone_offset
 
 
-@router.get('/node/{node_id}')
-async def get_node(node_id: str, aed_state: AEDStateDep, photo_state: PhotoStateDep):
-    aed = await aed_state.get_aed_by_id(node_id)
-
-    if aed is None:
-        raise HTTPException(404, f'Node {node_id!r} not found')
-
-    timezone_name, timezone_offset = _get_timezone(aed.position)
-    timezone_dict = {
-        '@timezone_name': timezone_name,
-        '@timezone_offset': timezone_offset,
-    }
-
-    image_url = aed.tags.get('image', '')
+async def _get_image_data(tags: dict[str, str], photo_state: PhotoStateDep) -> dict:
+    image_url: str = tags.get('image', '')
 
     if (
         image_url
         and (photo_id_match := photo_id_re.search(image_url))
         and (photo_id := photo_id_match.group('id'))
         and (photo_info := await photo_state.get_photo_by_id(photo_id))
     ):
-        photo_dict = {
+        return {
             '@photo_id': photo_info.id,
             '@photo_url': f'/api/v1/photos/view/{photo_info.id}.webp',
         }
-    elif image_url:
-        photo_dict = {
+
+    if image_url:
+        return {
             '@photo_id': None,
-            '@photo_url': f'/api/v1/photos/proxy/{quote_plus(image_url)}',
+            '@photo_url': f'/api/v1/photos/proxy/direct/{quote_plus(image_url)}',
         }
-    else:
-        photo_dict = {
+
+    wikimedia_commons: str = tags.get('wikimedia_commons', '')
+
+    if wikimedia_commons:
+        return {
             '@photo_id': None,
-            '@photo_url': None,
+            '@photo_url': f'/api/v1/photos/proxy/wikimedia-commons/{quote_plus(wikimedia_commons)}',
         }
 
+    return {
+        '@photo_id': None,
+        '@photo_url': None,
+    }
+
+
+@router.get('/node/{node_id}')
+@configure_cache(timedelta(minutes=1), stale=timedelta(minutes=5))
+async def get_node(node_id: str, aed_state: AEDStateDep, photo_state: PhotoStateDep):
+    aed = await aed_state.get_aed_by_id(node_id)
+
+    if aed is None:
+        raise HTTPException(404, f'Node {node_id!r} not found')
+
+    photo_dict = await _get_image_data(aed.tags, photo_state)
+
+    timezone_name, timezone_offset = _get_timezone(aed.position)
+    timezone_dict = {
+        '@timezone_name': timezone_name,
+        '@timezone_offset': timezone_offset,
+    }
+
     return {
         'version': 0.6,
         'copyright': 'OpenStreetMap and contributors',

api/v1/photos.py

@@ -5,6 +5,7 @@
 
 import magic
 import orjson
+from bs4 import BeautifulSoup
 from fastapi import APIRouter, File, Form, HTTPException, Request, Response, UploadFile
 from fastapi.responses import FileResponse
 from feedgen.feed import FeedGenerator
@@ -21,23 +22,10 @@
 router = APIRouter(prefix='/photos')
 
 
-@router.get('/view/{id}.webp')
-@configure_cache(timedelta(days=365), stale=timedelta(days=365))
-async def view(request: Request, id: str, photo_state: PhotoStateDep) -> FileResponse:
-    info = await photo_state.get_photo_by_id(id)
-
-    if info is None:
-        raise HTTPException(404, f'Photo {id!r} not found')
-
-    return FileResponse(info.path)
-
-
-@router.get('/proxy/{url_encoded:path}')
-@configure_cache(timedelta(days=7), stale=timedelta(days=7))
-async def proxy(request: Request, url_encoded: str) -> FileResponse:
+async def _fetch_image(url: str) -> tuple[bytes, str]:
     # NOTE: ideally we would verify whether url is not a private resource
     async with get_http_client() as http:
-        r = await http.get(unquote_plus(url_encoded))
+        r = await http.get(url)
         r.raise_for_status()
 
     # Early detection of unsupported types
@@ -58,7 +46,43 @@ async def proxy(request: Request, url_encoded: str) -> FileResponse:
     if content_type not in IMAGE_CONTENT_TYPES:
         raise HTTPException(500, f'Unsupported file type {content_type!r}, must be one of {IMAGE_CONTENT_TYPES}')
 
-    return Response(content=file, media_type=content_type)
+    return file, content_type
+
+
+@router.get('/view/{id}.webp')
+@configure_cache(timedelta(days=365), stale=timedelta(days=365))
+async def view(id: str, photo_state: PhotoStateDep) -> FileResponse:
+    info = await photo_state.get_photo_by_id(id)
+
+    if info is None:
+        raise HTTPException(404, f'Photo {id!r} not found')
+
+    return FileResponse(info.path)
+
+
+@router.get('/proxy/direct/{url_encoded:path}')
+@configure_cache(timedelta(days=7), stale=timedelta(days=7))
+async def proxy_direct(url_encoded: str) -> FileResponse:
+    file, content_type = await _fetch_image(unquote_plus(url_encoded))
+    return Response(file, media_type=content_type)
+
+
+@router.get('/proxy/wikimedia-commons/{path_encoded:path}')
+@configure_cache(timedelta(days=7), stale=timedelta(days=7))
+async def proxy_wikimedia_commons(path_encoded: str) -> FileResponse:
+    async with get_http_client() as http:
+        url = f'https://commons.wikimedia.org/wiki/{unquote_plus(path_encoded)}'
+        r = await http.get(url)
+        r.raise_for_status()
+
+    bs = BeautifulSoup(r.text, 'lxml')
+    og_image = bs.find('meta', property='og:image')
+    if not og_image:
+        raise HTTPException(404, 'Missing og:image meta tag')
+
+    image_url = og_image['content']
+    file, content_type = await _fetch_image(image_url)
+    return Response(file, media_type=content_type)
 
 
 @router.post('/upload')
@@ -76,12 +100,10 @@ async def upload(
 
     if file_license not in accept_licenses:
         raise HTTPException(400, f'Unsupported license {file_license!r}, must be one of {accept_licenses}')
-
     if file.size <= 0:
         raise HTTPException(400, 'File must not be empty')
 
     content_type = magic.from_buffer(file.file.read(2048), mime=True)
-
     if content_type not in IMAGE_CONTENT_TYPES:
         raise HTTPException(400, f'Unsupported file type {content_type!r}, must be one of {IMAGE_CONTENT_TYPES}')
 
@@ -94,13 +116,11 @@ async def upload(
         raise HTTPException(400, 'OAuth2 credentials must contain an access_token field')
 
     aed = await aed_state.get_aed_by_id(node_id)
-
     if aed is None:
         raise HTTPException(404, f'Node {node_id!r} not found, perhaps it is not an AED?')
 
     osm = OpenStreetMap(oauth2_credentials_)
     osm_user = await osm.get_authorized_user()
-
     if osm_user is None:
         raise HTTPException(401, 'OAuth2 credentials are invalid')
 

middlewares/cache_middleware.py

@@ -1,10 +1,13 @@
 import functools
+from contextvars import ContextVar
 from datetime import timedelta
 
 from fastapi import Request
 from starlette.middleware.base import BaseHTTPMiddleware
 from starlette.types import ASGIApp
 
+_request_context = ContextVar('Request_context')
+
 
 def make_cache_control(max_age: timedelta, stale: timedelta):
     return f'public, max-age={int(max_age.total_seconds())}, stale-while-revalidate={int(stale.total_seconds())}'
@@ -17,7 +20,11 @@ def __init__(self, app: ASGIApp, max_age: timedelta, stale: timedelta):
         self.stale = stale
 
     async def dispatch(self, request: Request, call_next):
-        response = await call_next(request)
+        token = _request_context.set(request)
+        try:
+            response = await call_next(request)
+        finally:
+            _request_context.reset(token)
 
         if request.method in ('GET', 'HEAD') and 200 <= response.status_code < 300:
             try:
@@ -40,7 +47,7 @@ def configure_cache(max_age: timedelta, stale: timedelta):
     def decorator(func):
         @functools.wraps(func)
         async def wrapper(*args, **kwargs):
-            request: Request = kwargs['request']
+            request: Request = _request_context.get()
             request.state.max_age = max_age
             request.state.stale = stale
             return await func(*args, **kwargs)