added diag script

video_transcription/ai_transcriber_v2/extractor.py

@@ -1,6 +1,7 @@
 import os
 import subprocess
 import sys
+import json
 from utils import verify_file_not_empty
 
 def run_ffmpeg(args):
@@ -20,6 +21,27 @@ def run_ffmpeg(args):
         print("Error: 'ffmpeg' command not found. Please ensure it is installed on your host system.")
         return False
 
+def get_video_duration(file_path):
+    """
+    Gets video duration in seconds using ffprobe.
+    """
+    in_flatpak = os.path.exists("/.flatpak-info")
+    cmd_base = ["flatpak-spawn", "--host"] if in_flatpak else []
+    
+    cmd = cmd_base + [
+        "ffprobe", 
+        "-v", "error", 
+        "-show_entries", "format=duration", 
+        "-of", "default=noprint_wrappers=1:nokey=1", 
+        file_path
+    ]
+    
+    try:
+        result = subprocess.run(cmd, capture_output=True, text=True, check=True)
+        return float(result.stdout.strip())
+    except Exception:
+        return 0.0
+
 def extract_audio(video_path, output_path=None):
     """
     Extracts audio from a video file using ffmpeg.
@@ -58,10 +80,19 @@ def extract_audio(video_path, output_path=None):
 def embed_subtitles(video_path, srt_path, output_path=None):
     """
     Embeds subtitles into the video file (Soft Subs) and sets them as primary.
+    Includes strict safety checks (Size & Duration) to prevent replacing videos with corrupted files.
     """
     if not os.path.exists(video_path) or not os.path.exists(srt_path):
         print("Error: Video or SRT file not found for embedding.")
-        return
+        return False
+
+    # Capture original stats
+    original_size = os.path.getsize(video_path)
+    original_duration = get_video_duration(video_path)
+    
+    if original_size == 0:
+        print("Error: Source video is 0 bytes.")
+        return False
 
     if output_path is None:
         base, ext = os.path.splitext(video_path)
@@ -91,8 +122,39 @@ def embed_subtitles(video_path, srt_path, output_path=None):
     ]
 
     if run_ffmpeg(args):
-        if not verify_file_not_empty(output_path):
-             raise Exception("FFmpeg succeeded but output video is empty.")
+        # --- Safety Checks ---
+        if not os.path.exists(output_path):
+             print("Error: Output file was not created.")
+             return False
+             
+        new_size = os.path.getsize(output_path)
+        new_duration = get_video_duration(output_path)
+        
+        # 1. Zero Byte Check
+        if new_size == 0:
+             print("❌ CRITICAL: Output file is 0 bytes. Deleting corrupted output.")
+             os.remove(output_path)
+             raise Exception("Embedding failed: Output is empty.")
+
+        # 2. Significant Size Drop Check
+        if new_size < (original_size * 0.8):
+             print(f"❌ CRITICAL: Output file is significantly smaller than source!")
+             print(f"   Original: {original_size/1024/1024:.2f} MB")
+             print(f"   New:      {new_size/1024/1024:.2f} MB")
+             os.remove(output_path)
+             raise Exception("Embedding failed: Suspicious file size reduction.")
+
+        # 3. Duration Mismatch Check (New)
+        if abs(original_duration - new_duration) > 1.0:
+             print(f"❌ CRITICAL: Duration mismatch detected!")
+             print(f"   Original: {original_duration:.2f}s")
+             print(f"   New:      {new_duration:.2f}s")
+             print("   This indicates truncated video stream. Aborting.")
+             os.remove(output_path)
+             raise Exception("Embedding failed: Duration mismatch > 1s.")
+
         print(f"Subtitles embedded successfully: {output_path}")
+        return True
     else:
         print("Error: Embedding failed.")
+        return False