added diag script

video_transcription/ai_transcriber_v2/diagnostic_check.py

@@ -0,0 +1,95 @@
+#!/usr/bin/env python3
+import sys
+import shutil
+import subprocess
+import torch
+import os
+
+def check_flatpak():
+    return os.path.exists("/.flatpak-info")
+
+def run_cmd(cmd_list):
+    in_flatpak = check_flatpak()
+    if in_flatpak:
+        full_cmd = ["flatpak-spawn", "--host"] + cmd_list
+    else:
+        full_cmd = cmd_list
+    
+    try:
+        result = subprocess.run(full_cmd, capture_output=True, text=True)
+        return result.returncode == 0, result.stdout.strip(), result.stderr.strip()
+    except FileNotFoundError:
+        return False, "", "Command not found"
+    except Exception as e:
+        return False, "", str(e)
+
+def main():
+    print("========================================")
+    print("   AI Transcriber Diagnostic Tool")
+    print("========================================")
+    
+    in_flatpak = check_flatpak()
+    print(f"Environment: {'Flatpak Sandbox' if in_flatpak else 'Native Host'}")
+    print("-" * 40)
+
+    # 1. GPU Check (PyTorch)
+    print("\n[1] GPU Availability (Internal PyTorch)")
+    if torch.cuda.is_available():
+        print(f"✅ GPU Detected: {torch.cuda.get_device_name(0)}")
+        print(f"   VRAM: {torch.cuda.get_device_properties(0).total_memory / 1024**3:.2f} GB")
+        print(f"   CUDA Version: {torch.version.cuda}")
+    else:
+        print("❌ GPU NOT Detected by PyTorch.")
+        print("   Whisper will run on CPU (Slow).")
+
+    # 2. Nvidia Driver Check (System)
+    print("\n[2] Nvidia Driver Check (System)")
+    ok, out, err = run_cmd(["nvidia-smi", "--query-gpu=name,driver_version", "--format=csv,noheader"])
+    if ok:
+        print(f"✅ Driver Active: {out}")
+    else:
+        print("❌ 'nvidia-smi' failed. Drivers might be missing or inaccessible.")
+        if err: print(f"   Error: {err}")
+
+    # 3. FFmpeg Check
+    print("\n[3] FFmpeg Check")
+    ok, out, err = run_cmd(["ffmpeg", "-version"])
+    if ok:
+        version_line = out.split('\n')[0]
+        print(f"✅ FFmpeg Ready: {version_line}")
+    else:
+        print("❌ FFmpeg NOT found.")
+        print("   On Bazzite, install it with: 'brew install ffmpeg'")
+
+    # 4. FFprobe Check
+    print("\n[4] FFprobe Check (Required for duration checks)")
+    ok, out, err = run_cmd(["ffprobe", "-version"])
+    if ok:
+        version_line = out.split('\n')[0]
+        print(f"✅ FFprobe Ready: {version_line}")
+    else:
+        print("❌ FFprobe NOT found.")
+
+    # 5. Ollama Check
+    print("\n[5] Local LLM (Ollama)")
+    ok, out, err = run_cmd(["ollama", "--version"])
+    if ok:
+        print(f"✅ Ollama Installed: {out}")
+        
+        # Check server
+        import socket
+        sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
+        sock.settimeout(1)
+        result = sock.connect_ex(('127.0.0.1', 11434))
+        if result == 0:
+            print("✅ Ollama Server is RUNNING.")
+        else:
+            print("⚠️  Ollama Server is NOT running (Scripts will try to auto-start it).")
+    else:
+        print("❌ Ollama NOT found.")
+        print("   Run: ./ai_transcriber_v2/install_local_llm.sh")
+
+    print("\n========================================")
+
+if __name__ == "__main__":
+    main()

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):
+        # --- Safety Checks ---
-             raise Exception("FFmpeg succeeded but output video is empty.")
+        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

video_transcription/ai_transcriber_v2/main.py

@@ -195,10 +195,9 @@ def process_file(file_path, args, source_lang=None, loaded_model=None, service_s
             should_embed = False
 
         if should_embed:
-            embed_subtitles(file_path, final_srt_path)
+            success_embed = embed_subtitles(file_path, final_srt_path)
 
-            if args.delete_source:
+            if success_embed and args.delete_source:
-                if args.embed:
                 base, ext = os.path.splitext(file_path)
                 expected_output = f"{base}.subbed{ext}"
                 
@@ -210,8 +209,6 @@ def process_file(file_path, args, source_lang=None, loaded_model=None, service_s
                         tracker.logger.error(f"Error deleting source: {e}")
                 else:
                     tracker.logger.error(f"SAFETY ABORT: Output '{expected_output}' not found.")
-                else:
-                    tracker.logger.warning("SAFETY ABORT: Enable --embed to delete source.")
         tracker.update_step(file_path, "step_embed", "done")
 
         # 5. Cleanup