Tweak/fix the attachment caching

app.py

@@ -225,8 +225,9 @@ def _get_file_url(url, size, kind):
         _GRIDFS_CACHE[k] = u
         return u
 
-    MEDIA_CACHE.cache(url, kind)
-    return _get_file_url(url, size, kind)
+    # MEDIA_CACHE.cache(url, kind)
+    app.logger.error("cache not available for {url}/{size}/{kind}")
+    return url
 
 
 @app.template_filter()

tasks.py

@@ -239,7 +239,10 @@ def cache_attachments(self, iri: str) -> None:
                     attachment.get("mediaType", "").startswith("image/")
                     or attachment.get("type") == ap.ActivityType.IMAGE.value
                 ):
-                    MEDIA_CACHE.cache(attachment["url"], Kind.ATTACHMENT)
+                    try:
+                        MEDIA_CACHE.cache(attachment["url"], Kind.ATTACHMENT)
+                    except ValueError:
+                        log.exception(f"failed to cache {attachment}")
 
         log.info(f"attachments cached for {iri}")
 

utils/media.py

@@ -15,6 +15,9 @@ def load(url, user_agent):
     """Initializes a `PIL.Image` from the URL."""
     with requests.get(url, stream=True, headers={"User-Agent": user_agent}) as resp:
         resp.raise_for_status()
+        if not resp.headers.get('content-type').startswith('image/'):
+            raise ValueError(f"bad content-type {resp.headers.get('content-type')}")
+
         resp.raw.decode_content = True
         return Image.open(BytesIO(resp.raw.read()))