92 lines
2.3 KiB
Python
92 lines
2.3 KiB
Python
|
|
|
|
import os
|
|
import time
|
|
import signal
|
|
import sys
|
|
from pathlib import Path
|
|
from redis import Redis
|
|
from rq import Queue
|
|
|
|
# Config via env (matches docker-compose)
|
|
LIB = Path(os.getenv("LIBRARY_ROOT", "/library"))
|
|
TRN = Path(os.getenv("TRANSCRIPT_ROOT", "/transcripts"))
|
|
REDIS_URL = os.getenv("REDIS_URL", "redis://redis:6379/0")
|
|
SCAN_INTERVAL = int(os.getenv("SCAN_INTERVAL", "30")) # seconds
|
|
|
|
# Media types to track
|
|
MEDIA_EXT = {
|
|
".mp3", ".m4a", ".mp4", ".mkv", ".wav", ".flac", ".webm", ".ogg", ".opus"
|
|
}
|
|
|
|
# In-memory seen set to avoid re-enqueueing during a single run
|
|
_seen: set[str] = set()
|
|
|
|
|
|
def already_transcribed(p: Path) -> bool:
|
|
"""Heuristic: if <basename>.json exists in transcripts, consider it done."""
|
|
base_json = TRN / f"{p.stem}.json"
|
|
return base_json.exists()
|
|
|
|
|
|
def iter_media_files(root: Path):
|
|
for path in root.rglob("*"):
|
|
if not path.is_file():
|
|
continue
|
|
if path.suffix.lower() in MEDIA_EXT:
|
|
yield path
|
|
|
|
|
|
def enqueue_new_files():
|
|
q = Queue(connection=Redis.from_url(REDIS_URL))
|
|
|
|
# Ensure target dirs exist
|
|
TRN.mkdir(parents=True, exist_ok=True)
|
|
LIB.mkdir(parents=True, exist_ok=True)
|
|
|
|
new_jobs = 0
|
|
for p in iter_media_files(LIB):
|
|
key = str(p.resolve())
|
|
if key in _seen:
|
|
continue
|
|
if already_transcribed(p):
|
|
_seen.add(key)
|
|
continue
|
|
# Enqueue the worker to process this local file
|
|
q.enqueue("worker.handle_local_file", key)
|
|
_seen.add(key)
|
|
new_jobs += 1
|
|
return new_jobs
|
|
|
|
|
|
_shutdown = False
|
|
|
|
|
|
def _handle_sig(sig, frame):
|
|
global _shutdown
|
|
_shutdown = True
|
|
|
|
|
|
def main():
|
|
signal.signal(signal.SIGINT, _handle_sig)
|
|
signal.signal(signal.SIGTERM, _handle_sig)
|
|
|
|
print(f"[scanner] Watching {LIB} → transcripts in {TRN}; interval={SCAN_INTERVAL}s", flush=True)
|
|
while not _shutdown:
|
|
try:
|
|
jobs = enqueue_new_files()
|
|
if jobs:
|
|
print(f"[scanner] Enqueued {jobs} new file(s)", flush=True)
|
|
except Exception as e:
|
|
print(f"[scanner] Error: {e}", file=sys.stderr, flush=True)
|
|
# Sleep between passes
|
|
for _ in range(SCAN_INTERVAL):
|
|
if _shutdown:
|
|
break
|
|
time.sleep(1)
|
|
|
|
print("[scanner] Shutting down", flush=True)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
main() |