feat: fix ios aspect ratio

.gitignore

@@ -275,4 +275,5 @@ conf/*
 apps/worker/celerybeat-schedule.db
 /apps/api/conf/*
 !/apps/api/conf/
-!/apps/api/conf/.gitkeep
+!/apps/api/conf/.gitkeep
+.idea/FastFetchBot.iml

AGENTS.md

@@ -205,6 +205,15 @@ MongoDB models and connection logic live in `packages/shared/fastfetchbot_shared
 - On subsequent cache hits, `media_files_packaging` uses stored file_ids directly via `InputMediaPhoto(file_id)` etc., skipping HTTP download entirely
 - The bot has no direct MongoDB access — all database writes go through the async worker via Redis
 
+**Future media asset caching plan** — target architecture for decoupling external media handling from the Telegram bot:
+- Current bot-side video probing with `ffprobe` is only a workaround for Telegram iOS aspect-ratio rendering issues. It should not grow into a full media processing pipeline inside `apps/telegram-bot`
+- Move external media download, video probing (`width`, `height`, `duration`), and any future remux/transcode work into the worker layer (`apps/worker` for synchronous heavy I/O, coordinated by API/async-worker as needed)
+- Introduce a durable `MediaAsset` cache backed by MongoDB metadata plus object storage such as S3. Suggested fields: source URL/hash, media type, MIME type, file size, storage key, width, height, duration, processing status, error details, expiration/lifecycle fields, and `telegram_file_id`
+- Let the bot consume prepared media only: prefer `telegram_file_id` first; otherwise use a ready local/object-storage asset with known dimensions; avoid downloading or probing arbitrary remote media in the bot event loop
+- Keep Telegram `file_id` as the highest-priority cache because it avoids both external downloads and object-storage reads on repeated sends
+- Use object storage primarily for first sends, unstable external sources, sources requiring cookies/referer, and cases where media metadata must be known before Telegram upload
+- Add lifecycle cleanup and deduplication by canonical source URL/content hash before enabling long-term object storage caching
+
 **SQLite/PostgreSQL** (user settings — always enabled for the Telegram bot):
 | Variable | Default | Description |
 |----------|---------|-------------|

apps/async-worker/async_worker/services/file_id_consumer.py

@@ -5,10 +5,12 @@
 
 from async_worker.config import settings
 from fastfetchbot_shared.utils.logger import logger
+from fastfetchbot_shared.utils.number import positive_int
 
 FILEID_QUEUE_KEY = "fileid:updates"
 FILEID_DLQ_KEY = "fileid:updates:dlq"
 _MAX_RETRIES = 3
+VIDEO_METADATA_FIELDS = ("width", "height", "duration")
 
 _redis: aioredis.Redis | None = None
 _consumer_task: asyncio.Task | None = None
@@ -44,7 +46,9 @@ async def _consume_loop() -> None:
                     await r.lpush(FILEID_QUEUE_KEY, raw_payload)
                     logger.info("Requeued in-flight payload before shutdown")
                 except Exception:
-                    logger.warning(f"Failed to requeue payload on shutdown: {raw_payload}")
+                    logger.warning(
+                        f"Failed to requeue payload on shutdown: {raw_payload}"
+                    )
             logger.info("file_id consumer cancelled, shutting down")
             break
         except json.JSONDecodeError as e:
@@ -70,11 +74,15 @@ async def _consume_loop() -> None:
                     try:
                         # Stamp retry count so we can detect repeated failures.
                         payload["_retry_count"] = retry_count + 1
-                        await r.lpush(FILEID_QUEUE_KEY, json.dumps(payload, ensure_ascii=False))
+                        await r.lpush(
+                            FILEID_QUEUE_KEY, json.dumps(payload, ensure_ascii=False)
+                        )
                     except Exception:
                         logger.warning(f"Failed to requeue payload: {raw_payload}")
                 else:
-                    logger.error(f"Max retries ({_MAX_RETRIES}) exceeded, moving to DLQ: {raw_payload}")
+                    logger.error(
+                        f"Max retries ({_MAX_RETRIES}) exceeded, moving to DLQ: {raw_payload}"
+                    )
                     try:
                         await r.lpush(FILEID_DLQ_KEY, raw_payload)
                     except Exception:
@@ -93,9 +101,12 @@ async def _process_file_id_update(payload: dict) -> None:
         logger.warning(f"Invalid file_id update payload: {payload}")
         return
 
-    doc = await Metadata.find(
-        Metadata.url == metadata_url
-    ).sort("-version").limit(1).first_or_none()
+    doc = (
+        await Metadata.find(Metadata.url == metadata_url)
+        .sort("-version")
+        .limit(1)
+        .first_or_none()
+    )
 
     if doc is None or not doc.media_files:
         logger.warning(f"No metadata found for file_id update: {metadata_url}")
@@ -104,9 +115,20 @@ async def _process_file_id_update(payload: dict) -> None:
     matched = 0
     for update in updates:
         for mf in doc.media_files:
-            if mf.url == update["url"] and mf.telegram_file_id is None:
-                mf.telegram_file_id = update["telegram_file_id"]
-                matched += 1
+            if mf.url == update["url"]:
+                changed = False
+                if mf.telegram_file_id is None:
+                    mf.telegram_file_id = update["telegram_file_id"]
+                    changed = True
+                if update.get("media_type") == "video":
+                    for field in VIDEO_METADATA_FIELDS:
+                        value = positive_int(update.get(field))
+                        if value is not None and getattr(mf, field, None) != value:
+                            setattr(mf, field, value)
+                            changed = True
+                if changed:
+                    matched += 1
+                break
 
     if matched > 0:
         await doc.save()

apps/telegram-bot/Dockerfile

@@ -25,6 +25,7 @@ COPY --from=ghcr.io/astral-sh/uv:0.10.4 /uv /usr/local/bin/uv
 RUN apt-get update \
     && apt-get install --no-install-recommends -y \
         libmagic1 \
+        ffmpeg \
         build-essential
 
 # copy workspace files for dependency resolution
@@ -42,7 +43,8 @@ FROM python-base AS production
 ENV PYTHONPATH=/app/apps/telegram-bot:$PYTHONPATH
 RUN apt-get update \
     && apt-get install --no-install-recommends -y \
-        libmagic1
+        libmagic1 \
+        ffmpeg
 COPY --from=builder-base $PYSETUP_PATH $PYSETUP_PATH
 COPY packages/ /app/packages/
 COPY apps/telegram-bot/ /app/apps/telegram-bot/

apps/telegram-bot/core/services/file_id_capture.py

@@ -6,8 +6,10 @@
 
 from core.config import settings
 from fastfetchbot_shared.utils.logger import logger
+from fastfetchbot_shared.utils.number import positive_int
 
 FILEID_QUEUE_KEY = "fileid:updates"
+VIDEO_METADATA_FIELDS = ("width", "height", "duration")
 
 _redis: aioredis.Redis | None = None
 
@@ -34,6 +36,18 @@ def extract_file_id(message: Message, media_type: str) -> Optional[str]:
     return None
 
 
+def _video_metadata_for_update(info: dict, message: Message) -> dict[str, int]:
+    metadata = {}
+    video = getattr(message, "video", None)
+    for field in VIDEO_METADATA_FIELDS:
+        value = positive_int(info.get(field))
+        if value is None and video:
+            value = positive_int(getattr(video, field, None))
+        if value is not None:
+            metadata[field] = value
+    return metadata
+
+
 async def capture_and_push_file_ids(
     uncached_info: list[dict],
     sent_messages: tuple[Message, ...],
@@ -57,21 +71,27 @@ async def capture_and_push_file_ids(
             break
         file_id = extract_file_id(sent_messages[i], info["media_type"])
         if file_id:
-            file_id_updates.append({
+            update = {
                 "url": info["url"],
                 "media_type": info["media_type"],
                 "telegram_file_id": file_id,
-            })
+            }
+            if info["media_type"] == "video":
+                update.update(_video_metadata_for_update(info, sent_messages[i]))
+            file_id_updates.append(update)
 
     if not file_id_updates:
         return
 
     try:
         r = await _get_redis()
-        payload = json.dumps({
-            "metadata_url": metadata_url,
-            "file_id_updates": file_id_updates,
-        }, ensure_ascii=False)
+        payload = json.dumps(
+            {
+                "metadata_url": metadata_url,
+                "file_id_updates": file_id_updates,
+            },
+            ensure_ascii=False,
+        )
         await r.lpush(FILEID_QUEUE_KEY, payload)
         logger.info(f"Pushed {len(file_id_updates)} file_id updates for {metadata_url}")
     except Exception: