1a7567231a
- 视频分析: MiMo 2.5 大模型直接分析 - 场景检测: AI 视觉模型识别分镜头 - 语音识别: 每个场景单独 ASR - 语音克隆: MiMo TTS 语音克隆 - 下载功能: 单文件/批量 ZIP 下载 - 系统设置: API 密钥配置 - 文件夹管理: 创建/移动/筛选 - 扁平化 UI: 简约按钮样式
54 lines
2.2 KiB
Python
54 lines
2.2 KiB
Python
import json
|
||
import urllib.request
|
||
|
||
resp = urllib.request.urlopen("http://localhost:5001/api/footage?per_page=60")
|
||
data = json.loads(resp.read().decode())
|
||
items = data.get("items", [])
|
||
|
||
# Find the target video
|
||
for item in items:
|
||
if "763456798450254550912" in item.get("filename", ""):
|
||
screener = item.get("screener", {})
|
||
scenes = screener.get("scenes", [])
|
||
transcription = screener.get("transcription", "")
|
||
|
||
print(f"Video: {item['filename']}")
|
||
print(f"Duration: {item.get('duration', 'N/A')}s")
|
||
print(f"Total scenes: {len(scenes)}")
|
||
print()
|
||
|
||
# Check scene timing
|
||
print("=== Scene Timing Analysis ===")
|
||
total_scene_duration = 0
|
||
for i, scene in enumerate(scenes):
|
||
start = scene.get("start_time", 0)
|
||
end = scene.get("end_time", 0)
|
||
duration = scene.get("duration", 0)
|
||
total_scene_duration += duration
|
||
|
||
# Check if duration matches end - start
|
||
calculated_duration = end - start
|
||
match = abs(duration - calculated_duration) < 0.1
|
||
|
||
print(f"Scene {scene.get('shot_index')}: {start}s - {end}s ({duration}s) {'✅' if match else '❌ MISMATCH'}")
|
||
|
||
print(f"\nTotal scene duration: {total_scene_duration}s")
|
||
print(f"Video duration: {item.get('duration', 'N/A')}s")
|
||
|
||
# Check speech alignment
|
||
print("\n=== Speech Alignment ===")
|
||
all_dialogue = " ".join([s.get("dialogue", "") for s in scenes if s.get("dialogue")])
|
||
print(f"Total dialogue length: {len(all_dialogue)} chars")
|
||
print(f"Transcription length: {len(transcription)} chars")
|
||
|
||
# Check if dialogue matches transcription
|
||
if transcription:
|
||
# Simple check: see if dialogue words appear in transcription
|
||
dialogue_words = set(all_dialogue.replace("，", "").replace("。", "").replace("！", ""))
|
||
trans_words = set(transcription.replace("，", "").replace("。", "").replace("！", ""))
|
||
|
||
overlap = dialogue_words & trans_words
|
||
print(f"Dialogue words overlap with transcription: {len(overlap)} chars")
|
||
|
||
break
|