refactor: Replace subprocess with yt-dlp for subtitle download

summarize_yt/cli.py

@@ -3,9 +3,9 @@ import argparse
 import os
 import re
 import signal
-import subprocess
 import sys
 from datetime import datetime
+from yt_dlp import YoutubeDL
 from pathlib import Path
 from typing import Optional, Tuple
 
@@ -56,33 +56,27 @@ def ensure_transcript_dir() -> Path:
 def download_subtitles(url: str, quiet: bool = False) -> Tuple[bool, Optional[Path]]:
     """Download subtitles from YouTube using yt-dlp."""
     try:
-        cmd = ["yt-dlp", "--skip-download", "--write-auto-sub", "--sub-lang", "en"]
+        ydl_opts = {
-        if quiet:
+            'skip_download': True,
-            cmd.append("--quiet")
+            'writeautomaticsub': True,
-        cmd.append(url)
+            'subtitleslangs': ['en'],
+            'quiet': quiet,
+            'no_warnings': quiet,
+        }
 
-        print(f"Debug: Running command: {cmd}")  # Debug line
+        if not quiet:
+            print(f"Debug: Downloading subtitles for {url}")
 
-        process = subprocess.Popen(
+        with YoutubeDL(ydl_opts) as ydl:
-            cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True  # This might help with encoding issues
+            ydl.download([url])
-        )
-
-        try:
-            stdout, stderr = process.communicate(timeout=30)
-            print(f"Debug: stdout: {stdout}")  # Debug line
-            print(f"Debug: stderr: {stderr}")  # Debug line
-            if process.returncode != 0:
-                print(f"Error downloading subtitles: {stderr}", file=sys.stderr)
-                return False, None
-        except subprocess.TimeoutExpired:
-            process.kill()
-            print("Download timed out after 30 seconds", file=sys.stderr)
-            return False, None
 
         # Find the downloaded VTT file
         current_dir = Path(".")
         vtt_files = list(current_dir.glob("*.en.vtt"))
-        print(f"Debug: Found VTT files: {vtt_files}")  # Debug line
+        
+        if not quiet:
+            print(f"Debug: Found VTT files: {vtt_files}")
+            
         if not vtt_files:
             print("No VTT file found after download", file=sys.stderr)
             return False, None
@@ -91,10 +85,10 @@ def download_subtitles(url: str, quiet: bool = False) -> Tuple[bool, Optional[Pa
 
     except Exception as e:
         print(f"Error during subtitle download: {str(e)}", file=sys.stderr)
-        print(f"Debug: Full exception info: {type(e).__name__}: {str(e)}")  # Debug line
+        if not quiet:
+            print(f"Debug: Full exception info: {type(e).__name__}: {str(e)}")
             import traceback
-
+            traceback.print_exc()
-        traceback.print_exc()  # This will print the full traceback
         return False, None