david/personal_development
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

syncing latest revisions - dk123

Browse Source
This commit is contained in:
david
2026-01-12 15:15:37 -05:00
parent 56494208bd
commit e471d3e1f8
11 changed files with 100 additions and 89 deletions
Download Patch File Download Diff File Expand all files Collapse all files
video_transcription/ai_transcriber_v2/extractor.py
+15 -19
View File
@@ -2,6 +2,7 @@ import os
import subprocess
import sys
import json
import tracker
from utils import verify_file_not_empty
def run_ffmpeg(args):
@@ -15,10 +16,10 @@ def run_ffmpeg(args):
subprocess.run(cmd, check=True)
return True
except subprocess.CalledProcessError as e:
print(f"FFmpeg Error: {e}")
tracker.logger.error(f"FFmpeg Error: {e}")
return False
except FileNotFoundError:
print("Error: 'ffmpeg' command not found. Please ensure it is installed on your host system.")
tracker.logger.error("Error: 'ffmpeg' command not found. Please ensure it is installed on your host system.")
return False
def get_video_duration(file_path):
@@ -54,10 +55,10 @@ def extract_audio(video_path, output_path=None):
output_path = f"{base_name}.wav"
if verify_file_not_empty(output_path):
print(f"Audio file already exists: {output_path}")
tracker.logger.info(f"Audio file already exists: {output_path}")
return output_path
print(f"Extracting audio from {video_path}...")
tracker.logger.info(f"Extracting audio from {video_path}...")
args = [
"-i", video_path,
@@ -72,7 +73,7 @@ def extract_audio(video_path, output_path=None):
if run_ffmpeg(args):
if not verify_file_not_empty(output_path):
raise Exception("FFmpeg succeeded but output is empty.")
print(f"Audio extracted to: {output_path}")
tracker.logger.info(f"Audio extracted to: {output_path}")
return output_path
else:
sys.exit(1)
@@ -83,7 +84,7 @@ def embed_subtitles(video_path, srt_path, output_path=None):
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.")
tracker.logger.error("Error: Video or SRT file not found for embedding.")
return False
# Capture original stats
@@ -91,14 +92,14 @@ def embed_subtitles(video_path, srt_path, output_path=None):
original_duration = get_video_duration(video_path)
if original_size == 0:
print("Error: Source video is 0 bytes.")
tracker.logger.error("Error: Source video is 0 bytes.")
return False
if output_path is None:
base, ext = os.path.splitext(video_path)
output_path = f"{base}.subbed{ext}"
print(f"Embedding subtitles into: {output_path}...")
tracker.logger.info(f"Embedding subtitles into: {output_path}...")
sub_codec = "mov_text" if video_path.lower().endswith(".mp4") else "srt"
@@ -124,7 +125,7 @@ def embed_subtitles(video_path, srt_path, output_path=None):
if run_ffmpeg(args):
# --- Safety Checks ---
if not os.path.exists(output_path):
print("Error: Output file was not created.")
tracker.logger.error("Error: Output file was not created.")
return False
new_size = os.path.getsize(output_path)
@@ -132,29 +133,24 @@ def embed_subtitles(video_path, srt_path, output_path=None):
# 1. Zero Byte Check
if new_size == 0:
print("❌ CRITICAL: Output file is 0 bytes. Deleting corrupted output.")
tracker.logger.error("❌ 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")
tracker.logger.error(f"❌ CRITICAL: Output file is significantly smaller than source!")
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.")
tracker.logger.error(f"❌ CRITICAL: Duration mismatch detected!")
os.remove(output_path)
raise Exception("Embedding failed: Duration mismatch > 1s.")
print(f"Subtitles embedded successfully: {output_path}")
tracker.logger.info(f"Subtitles embedded successfully: {output_path}")
return True
else:
print("Error: Embedding failed.")
tracker.logger.error("Error: Embedding failed.")
return False