""" backfill_media_by_bot_id.py ---------------------------- Manually fetches the audio and video for a specific meeting from Recall AI and uploads them to S3, updating the database exactly as the normal pipeline does. Use this when the server crashed mid-processing and the recording row in the database has null audio_download_url / video_download_url. Usage: # By bot ID python scripts/backfill_media_by_bot_id.py --bot-id "2d9750c1-3777-4cf8-a439-7ca57077a908" # By meeting ID python scripts/backfill_media_by_bot_id.py --meeting-id 123 # Force re-upload even if S3 URLs already exist python scripts/backfill_media_by_bot_id.py --bot-id "..." --force Docker: docker exec -it iih_ai-notemaker-app python /app/scripts/backfill_media_by_bot_id.py --bot-id "YOUR_BOT_ID" """ import argparse import logging import sys from pathlib import Path import requests PROJECT_ROOT = Path(__file__).resolve().parents[1] if str(PROJECT_ROOT) not in sys.path: sys.path.insert(0, str(PROJECT_ROOT)) from app import crud from app.database import SessionLocal from app.services import recall_service logging.basicConfig( level=logging.INFO, format="%(asctime)s | %(levelname)s | %(message)s", ) logger = logging.getLogger(__name__) def backfill_media(meeting_id: int = None, bot_id: str = None, force: bool = False) -> int: db = SessionLocal() try: # ── Step 1: Resolve the meeting ───────────────────────────────────── if bot_id: meeting = db.query(crud.Meeting).filter(crud.Meeting.recall_bot_id == bot_id).first() if not meeting: logger.error("No meeting found in the database for bot ID: %s", bot_id) return 1 logger.info("Resolved bot ID %s → meeting ID %s ('%s')", bot_id, meeting.id, meeting.title) meeting_id = meeting.id else: meeting = crud.get_meeting_by_id(db, meeting_id) if not meeting: logger.error("No meeting found in the database for meeting ID: %s", meeting_id) return 1 logger.info("Found meeting ID %s ('%s')", meeting.id, meeting.title) if not meeting.recall_bot_id: logger.error("Meeting %s has no recall_bot_id in the database. Cannot fetch media.", meeting.id) return 1 # ── Step 2: Check current recording state ─────────────────────────── recording = crud.get_recording_by_meeting(db, meeting_id) if recording: has_audio = bool(recording.audio_download_url) has_video = bool(recording.video_download_url) if has_audio and has_video and not force: logger.info( "Meeting %s already has both audio and video in the database. " "Use --force to re-upload anyway. Skipping.", meeting_id, ) logger.info(" Audio: %s", recording.audio_download_url) logger.info(" Video: %s", recording.video_download_url) return 0 if has_audio and not force: logger.info("Meeting %s already has audio. Only video will be re-fetched.", meeting_id) if has_video and not force: logger.info("Meeting %s already has video. Only audio will be re-fetched.", meeting_id) else: logger.info("No recording row exists yet for meeting %s. A new one will be created.", meeting_id) # ── Step 3: Fetch the latest bot data from Recall AI ──────────────── logger.info("Fetching bot data from Recall AI for bot ID: %s ...", meeting.recall_bot_id) try: bot_data = recall_service.get_bot(meeting.recall_bot_id) except requests.RequestException as exc: logger.error("Failed to fetch bot data from Recall AI: %s", exc) return 1 recording_data = recall_service.latest_recording(bot_data) if not recording_data: logger.error( "Recall AI returned no recording data for bot ID %s. " "The recording may not be ready yet or may have expired.", meeting.recall_bot_id, ) return 1 recall_recording_id = recording_data.get("id") recording_status = (recording_data.get("status") or {}).get("code", "unknown") logger.info( "Recall recording found. Recording ID: %s. Status: %s.", recall_recording_id, recording_status, ) if recording_status != "done": logger.warning( "Recording status is '%s', not 'done'. Media URLs may not be available yet. " "Proceeding anyway...", recording_status, ) # ── Step 4: Upload to S3 and update the database ──────────────────── logger.info("Uploading media to S3 and updating the database for meeting %s ...", meeting_id) try: updated_recording = crud.upsert_recording_from_recall( db, meeting_id, recording_data, mark_links_refreshed=True, ) except Exception as exc: logger.error("Database update failed: %s", exc) return 1 if not updated_recording: logger.error("upsert_recording_from_recall returned None. Something went wrong.") return 1 # ── Step 5: Report results ─────────────────────────────────────────── logger.info("✅ Success! Media has been saved for meeting %s.", meeting_id) logger.info(" Recall Recording ID : %s", updated_recording.recall_recording_id) logger.info(" Recording Status : %s", updated_recording.status) logger.info(" Audio URL : %s", updated_recording.audio_download_url or "NOT AVAILABLE") logger.info(" Video URL : %s", updated_recording.video_download_url or "NOT AVAILABLE") logger.info(" Refreshed At : %s", updated_recording.download_urls_refreshed_at) if not updated_recording.audio_download_url: logger.warning("Audio URL is still null. Recall AI may not have the audio ready yet.") if not updated_recording.video_download_url: logger.warning("Video URL is still null. Recall AI may not have the video ready yet.") return 0 except Exception as exc: logger.exception("Unexpected error: %s", exc) return 1 finally: db.close() def main() -> int: parser = argparse.ArgumentParser( description="Backfill audio/video media for a specific meeting from Recall AI to S3." ) group = parser.add_mutually_exclusive_group(required=True) group.add_argument("--bot-id", type=str, help="The Recall AI bot ID") group.add_argument("--meeting-id", type=int, help="The meeting ID in the database") parser.add_argument( "--force", action="store_true", help="Re-upload even if audio/video URLs already exist in the database.", ) args = parser.parse_args() return backfill_media(meeting_id=args.meeting_id, bot_id=args.bot_id, force=args.force) if __name__ == "__main__": raise SystemExit(main())