#!/usr/bin/env python3 import json, time, urllib.request, subprocess from pathlib import Path PROJECT=Path('/data/video-pipeline/Youtube-Video-Maker/Project/01072026-001-Nha-Tay-Son') PROBE='http://192.168.1.104:8018/probe' BLACK='http://192.168.1.104:8018/blackdetect' FINAL=PROJECT/'output/final/final_video.mp4' FINAL_PROBE=PROJECT/'output/final/final_video_probe.json' PKG_REPORT=PROJECT/'output/final/final_package_report.json' QA_REPORT=PROJECT/'output/final/final_qa_report.json' INFO=PROJECT/'output/info.json' THUMB16=PROJECT/'output/thumbnails/thumb_16x9_final.png' THUMB9=PROJECT/'output/thumbnails/thumb_9x16_final.png' MANIFEST=PROJECT/'logs/step19_final_qa_info_manifest.json' def post(url,payload,timeout=21600): req=urllib.request.Request(url,data=json.dumps(payload,ensure_ascii=False).encode(),headers={'Content-Type':'application/json'},method='POST') with urllib.request.urlopen(req,timeout=timeout) as r: return json.loads(r.read().decode()) def load(p, default=None): return json.loads(p.read_text(encoding='utf-8')) if p.exists() else default def dur(pr): return float(pr.get('format',{}).get('duration') or 0) def has(pr,t): return any(s.get('codec_type')==t for s in pr.get('streams',[])) def video_stream(pr): return next((s for s in pr.get('streams',[]) if s.get('codec_type')=='video'), {}) def ffprobe_dimensions(path): out=subprocess.check_output(['ffprobe','-v','error','-select_streams','v:0','-show_entries','stream=width,height,display_aspect_ratio','-of','json',str(path)], text=True, timeout=60) s=(json.loads(out).get('streams') or [{}])[0] return int(s.get('width') or 0), int(s.get('height') or 0), s.get('display_aspect_ratio') def count_files(path, glob): return len(list(Path(path).glob(glob))) if Path(path).exists() else 0 def extract_frames(times): outdir=PROJECT/'output/final/final_qa_frames'; outdir.mkdir(parents=True,exist_ok=True) frames=[] for label,sec in times: out=outdir/f'{label}_{sec:.1f}.jpg' cmd=['ffmpeg','-y','-hide_banner','-loglevel','error','-ss',f'{sec:.3f}','-i',str(FINAL),'-frames:v','1','-q:v','2',str(out)] subprocess.run(cmd,check=True,timeout=120) frames.append({'label':label,'time_sec':sec,'path':str(out),'exists':out.exists()}) return frames def parse_meta(): txt=(PROJECT/'script/source_story.txt').read_text(encoding='utf-8') series='Việt Sử Kể Lại'; title='NGƯỜI CON GÁI HÁI DÂU BƯỚC VÀO LONG MẠCH ĐẠI VIỆT' for line in txt.splitlines(): low=line.lower() if low.startswith('series:'): series=line.split(':',1)[1].strip() if low.startswith('tiêu đề truyện:') or low.startswith('tieu de truyen:'): title=line.split(':',1)[1].strip().title() return series,title def main(): if not FINAL.exists(): raise SystemExit('missing final') pr=post(PROBE,{'input_path':str(FINAL)},120) FINAL_PROBE.write_text(json.dumps(pr,ensure_ascii=False,indent=2),encoding='utf-8') pkg=load(PKG_REPORT,{}) black=post(BLACK,{'input_path':str(FINAL),'log_path':str(PROJECT/'output/final/final_qa_blackdetect.log')},21600) total=dur(pr) intro_dur=dur(pkg.get('input_probes',{}).get('intro',{})) or 16.2 main_dur=dur(pkg.get('input_probes',{}).get('main',{})) or 697.682 outro_start=max(0,total-25) times=[('intro_transition',max(0,intro_dur-1)),('early_main',intro_dur+30),('mid_main',intro_dur+main_dur/2),('late_main',intro_dur+max(0,main_dur-45)),('outro',outro_start+5)] frames=extract_frames(times) master_probe=load(PROJECT/'audio/master_audio_probe.json',{}) main_probe=load(PROJECT/'output/main_story_probe.json',{}) scenes=load(PROJECT/'script/scenes.json',{}).get('scenes',[]) parts=load(PROJECT/'parts/part_ranges.json',{}).get('parts',[]) series,title=parse_meta() vs=video_stream(pr) final_width=int(vs.get('width') or 0); final_height=int(vs.get('height') or 0); final_dar=vs.get('display_aspect_ratio') if not final_width or not final_height: final_width, final_height, final_dar = ffprobe_dimensions(FINAL) validation={ 'final_exists':FINAL.exists() and FINAL.stat().st_size>1000, 'has_audio':has(pr,'audio'),'has_video':has(pr,'video'), 'final_width':final_width,'final_height':final_height,'final_display_aspect_ratio':final_dar, 'resolution_16x9_1080p': final_width==1920 and final_height==1080, 'thumbnail_16x9_exists':THUMB16.exists(),'thumbnail_9x16_exists':THUMB9.exists(), 'black_event_count':black.get('event_count',0),'blackdetect_clear':black.get('event_count',0)==0, 'qa_frames_extracted':all(f['exists'] for f in frames), 'package_completed':pkg.get('status') in ('completed','needs_review'), 'bgm_volume':pkg.get('validation',{}).get('bgm_volume'), 'decision':'approved_output_candidate_after_automated_qa; human spot-check recommended for subjective BGM/sync' } passed=all([validation['final_exists'],validation['has_audio'],validation['has_video'],validation['resolution_16x9_1080p'],validation['thumbnail_16x9_exists'],validation['thumbnail_9x16_exists'],validation['blackdetect_clear'],validation['qa_frames_extracted'],validation['package_completed']]) qa={'project_id':PROJECT.name,'step':19,'passed':passed,'status':'passed' if passed else 'failed','final_video_path':str(FINAL),'final_duration_sec':total,'qa_points':times,'qa_frames':frames,'blackdetect':black,'validation':validation,'notes':['Automated QA checks streams, duration, black frames, frames at intro/early/mid/late/outro. User should still listen/spot-check BGM and sync before publishing.']} QA_REPORT.write_text(json.dumps(qa,ensure_ascii=False,indent=2),encoding='utf-8') yt_tags=['#vetmucthoigian','#thanhcattuhan','#genghiskhan','#mongol','#lichsuthegioi','#lichsu','#worldhistory'] info={ 'project_id':PROJECT.name,'series':series,'title':title,'approved_output':passed,'qa_final_status':qa['status'], 'paths':{ 'source_script':str(PROJECT/'script/source_story.txt'),'cleaned_story':str(PROJECT/'script/cleaned_story.txt'),'dense_plan':str(PROJECT/'script/scenes.json'),'character_bible':str(PROJECT/'script/character_bible.json'), 'scene_images_dir':str(PROJECT/'images/scenes'),'thumbnails':{'16x9':str(THUMB16),'9x16':str(THUMB9)},'voice_segments_dir':str(PROJECT/'audio/voice_segments'),'scene_tts_dir':str(PROJECT/'audio/scenes'),'master_narration':str(PROJECT/'audio/master_narration.wav'), 'subtitles':str(PROJECT/'subtitles/master_subtitles_clamped.srt'),'part_ranges':str(PROJECT/'parts/part_ranges.json'),'rendered_parts_dir':str(PROJECT/'render_parts'),'main_story':str(PROJECT/'output/main_story.mp4'), 'intro':str(PROJECT/'output/intro/intro.mp4'),'final_output':str(FINAL),'final_qa_report':str(QA_REPORT),'final_package_report':str(PKG_REPORT) }, 'durations':{'master_audio_sec':master_probe.get('master_duration_sec') or dur(load(PROJECT/'audio/master_audio_probe.json',{})),'main_story_sec':main_probe.get('main_duration_sec') or dur(main_probe),'final_sec':total}, 'counts':{'scenes':len(scenes),'scene_images':count_files(PROJECT/'images/scenes','scene_*.png'),'parts':len(parts)}, 'social_metadata':{ 'title_youtube_1': title, 'title_youtube_2': f'{title} | {series}', 'title_youtube_3': f'{title} - Vết Mực Thời Gian', 'description_youtube': f'Cùng Vết Mực Thời Gian khám phá {title}. Một câu chuyện thuộc series {series}.', 'hashtag_youtube': '#vetmucthoigian #lichsu #lichsuvietnam #lichsuthegioi #worldhistory', 'title_facebook': title, 'description_facebook': f'Câu chuyện {title}, thuộc series {series}.', 'hashtag_facebook': '#vetmucthoigian #lichsu #lichsuvietnam #lichsuthegioi #worldhistory', 'title_tiktok': title, 'description_tiktok': f'{title} | Vết Mực Thời Gian.', 'hashtag_tiktok': '#vetmucthoigian #lichsu #lichsuvietnam #lichsuthegioi #worldhistory' } } INFO.write_text(json.dumps(info,ensure_ascii=False,indent=2),encoding='utf-8') MANIFEST.write_text(json.dumps({'project_id':PROJECT.name,'step':19,'status':qa['status'],'info_path':str(INFO),'qa_report_path':str(QA_REPORT),'final_output':str(FINAL)},ensure_ascii=False,indent=2),encoding='utf-8') print(time.strftime('%Y-%m-%d %H:%M:%S ')+'DONE '+json.dumps({'status':qa['status'],'final_duration':total,'info':str(INFO)},ensure_ascii=False),flush=True) if not passed: raise SystemExit('failed_validation') if __name__=='__main__': main()