anbeime/skill / skills/video-frame-extractor/video-frame-extractor
video-frame-extractor
Unverified影片与视频编辑、自媒体创作者在需要拆解优秀视频、提取分镜参考或反推提示词时,使用本技能可自动抽取视频关键帧,调用视觉大模型批量分析画面,一键生成结构化描述与创作提示词,轻松搞定视频内容提取与场景分析。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 7.0 KB |
| references/analysis-guide.md | markdown | 7.9 KB |
| scripts/coze_bot_client.py | script | 10.3 KB |
| scripts/video_frame_extractor.py | script | 6.4 KB |
| scripts/visual_analyzer.py | script | 9.2 KB |
scripts/coze_bot_client.py
View on GitHub#!/usr/bin/env python3
"""
Coze Bot API客户端
调用用户发布的Coze Bot进行视觉分析和内容处理
"""
import os
import sys
import base64
import json
import requests
from typing import List, Dict, Any, Optional
def encode_image_to_base64(image_path: str) -> str:
"""
将图片编码为Base64
参数:
image_path: 图片路径
返回:
Base64编码的字符串
"""
with open(image_path, "rb") as image_file:
return base64.b64encode(image_file.read()).decode('utf-8')
def call_coze_bot(
message: str,
image_path: Optional[str] = None,
bot_id: str = None,
user_id: str = "user_001",
api_key: str = None,
stream: bool = False
) -> Dict[str, Any]:
"""
调用Coze Bot API
参数:
message: 用户消息
image_path: 图片路径(可选)
bot_id: Bot ID
user_id: 用户ID
api_key: API密钥
stream: 是否流式响应
返回:
响应结果字典
"""
# 获取配置
bot_id = bot_id or os.getenv('COZE_BOT_ID')
api_key = api_key or os.getenv('COZE_API_KEY')
if not bot_id:
raise ValueError("缺少Bot ID,请设置环境变量COZE_BOT_ID或通过参数传入")
if not api_key:
raise ValueError("缺少API Key,请设置环境变量COZE_API_KEY或通过参数传入")
# 构建消息
additional_messages = [
{
"content": message,
"content_type": "text",
"role": "user",
"type": "question"
}
]
# 如果有图片,添加图片消息
if image_path:
base64_image = encode_image_to_base64(image_path)
additional_messages.append({
"content": f"data:image/jpeg;base64,{base64_image}",
"content_type": "image",
"role": "user",
"type": "question"
})
# 构建请求
url = "https://api.coze.cn/v3/chat"
headers = {
"Authorization": f"Bearer {api_key}",
"Content-Type": "application/json"
}
payload = {
"bot_id": bot_id,
"user_id": user_id,
"stream": stream,
"additional_messages": additional_messages,
"parameters": {}
}
try:
response = requests.post(url, headers=headers, json=payload, timeout=60)
if response.status_code != 200:
raise Exception(f"API调用失败,状态码: {response.status_code}, 响应: {response.text}")
return {
'success': True,
'status_code': response.status_code,
'response': response.json()
}
except requests.exceptions.RequestException as e:
return {
'success': False,
'error': f"请求失败: {str(e)}"
}
except Exception as e:
return {
'success': False,
'error': str(e)
}
def analyze_image_with_coze_bot(
image_path: str,
prompt: str = "详细描述这张图片的内容,包括场景、人物、构图、光影等",
bot_id: str = None,
user_id: str = "user_001",
api_key: str = None
) -> Dict[str, Any]:
"""
使用Coze Bot分析图片
参数:
image_path: 图片路径
prompt: 分析提示词
bot_id: Bot ID
user_id: 用户ID
api_key: API密钥
返回:
分析结果字典
"""
result = call_coze_bot(
message=prompt,
image_path=image_path,
bot_id=bot_id,
user_id=user_id,
api_key=api_key,
stream=False
)
if result['success']:
# 提取回复内容
response_data = result['response']
# Coze API响应结构
if 'messages' in response_data and len(response_data['messages']) > 0:
# 获取最后一条消息
last_message = response_data['messages'][-1]
if 'content' in last_message:
return {
'success': True,
'description': last_message['content'],
'raw_response': response_data
}
# 检查是否在data字段中
if 'data' in response_data:
return {
'success': True,
'description': str(response_data['data']),
'raw_response': response_data
}
return {
'success': False,
'error': '无法解析API响应',
'raw_response': response_data
}
else:
return result
def batch_analyze_images_with_coze_bot(
image_dir: str,
prompt: str = "详细描述这张图片的内容,包括场景、人物、构图、光影等",
bot_id: str = None,
api_key: str = None,
output_file: str = None
) -> Dict[str, Any]:
"""
批量使用Coze Bot分析图片
参数:
image_dir: 图片目录
prompt: 分析提示词
bot_id: Bot ID
api_key: API密钥
output_file: 输出JSON文件路径(可选)
返回:
批量分析结果字典
"""
import time
from pathlib import Path
results = {
'success': True,
'total_images': 0,
'analyzed_count': 0,
'failed_count': 0,
'analysis': [],
'errors': []
}
try:
# 收集图片文件
image_files = []
for file in os.listdir(image_dir):
if file.lower().endswith(('.jpg', '.jpeg', '.png')):
image_files.append(os.path.join(image_dir, file))
image_files.sort()
results['total_images'] = len(image_files)
if len(image_files) == 0:
raise ValueError("没有找到图片文件")
print(f"找到 {len(image_files)} 张图片")
# 批量分析
for i, image_file in enumerate(image_files):
print(f"\n分析图片 {i+1}/{len(image_files)}: {os.path.basename(image_file)}")
# 调用Coze Bot分析
analysis_result = analyze_image_with_coze_bot(
image_path=image_file,
prompt=prompt,
bot_id=bot_id,
api_key=api_key
)
# 构建结果
frame_data = {
'frame_file': os.path.basename(image_file),
'frame_path': image_file,
}
if analysis_result['success']:
frame_data['description'] = analysis_result['description']
frame_data['analysis_status'] = 'success'
results['analyzed_count'] += 1
print(f" ✅ 分析成功")
else:
frame_data['error'] = analysis_result.get('error', '未知错误')
frame_data['analysis_status'] = 'failed'
results['failed_count'] += 1
print(f" ❌ 分析失败: {frame_data['error']}")
results['analysis'].append(frame_data)
# 控制请求速率,避免频繁调用
if i < len(image_files) - 1:
print(f" 等待2秒...")
time.sleep(2)
# 保存结果
if output_file:
os.makedirs(os.path.dirname(output_file), exist_ok=True)
with open(output_file, 'w', encoding='utf-8') as f:
json.dump(results, f, indent=2, ensure_ascii=False)
print(f"\n✅ 结果已保存到: {output_file}")
print(f"\n批量分析完成:")
print(f" 总图片数: {results['total_images']}")
print(f" 分析成功: {results['analyzed_count']}")
print(f" 分析失败: {results['failed_count']}")
except Exception as e:
results['success'] = False
results['errors'].append(str(e))
print(f"❌ 批量分析失败: {str(e)}")
return results
def main():
import argparse
parser = argparse.ArgumentParser(description='Coze Bot API客户端')
parser.add_argument('--message', type=str, help='用户消息')
parser.add_argument('--image', type=str, help='图片路径')
parser.add_argument('--image_dir', type=str, help='图片目录(批量分析)')
parser.add_argument('--prompt', type=str,
default='详细描述这张图片的内容,包括场景、人物、构图、光影等,适合AI视频创作参考。',
help='分析提示词')
parser.add_argument('--bot_id', type=str, help='Bot ID(默认从环境变量COZE_BOT_ID读取)')
parser.add_argument('--api_key', type=str, help='API Key(默认从环境变量COZE_API_KEY读取)')
parser.add_argument('--output', type=str, help='输出JSON文件路径(批量分析时使用)')
args = parser.parse_args()
# 获取配置
bot_id = args.bot_id or os.getenv('COZE_BOT_ID')
api_key = args.api_key or os.getenv('COZE_API_KEY')
# 示例配置(如果用户未配置)
if not bot_id:
bot_id = "7572557757883383858"
print("使用默认Bot ID(可通过环境变量COZE_BOT_ID自定义)")
if not api_key:
api_key = "cztei_qHZQ0A5OSJjsmfZWmVb8bqu2BTbtB240YGbDYLhZpsIr8jER4aL4Aevyii8rnKfNs"
print("使用默认API Key(可通过环境变量COZE_API_KEY自定义)")
# 执行操作
if args.image_dir:
# 批量分析
results = batch_analyze_images_with_coze_bot(
image_dir=args.image_dir,
prompt=args.prompt,
bot_id=bot_id,
api_key=api_key,
output_file=args.output
)
sys.exit(0 if results['success'] else 1)
elif args.image:
# 单张图片分析
result = analyze_image_with_coze_bot(
image_path=args.image,
prompt=args.message or args.prompt,
bot_id=bot_id,
api_key=api_key
)
if result['success']:
print(f"\n分析结果:\n{result['description']}")
sys.exit(0)
else:
print(f"\n分析失败: {result.get('error', '未知错误')}")
sys.exit(1)
elif args.message:
# 纯文本对话
result = call_coze_bot(
message=args.message,
bot_id=bot_id,
api_key=api_key
)
if result['success']:
print(f"\nBot回复:\n{json.dumps(result['response'], indent=2, ensure_ascii=False)}")
sys.exit(0)
else:
print(f"\n调用失败: {result.get('error', '未知错误')}")
sys.exit(1)
else:
parser.print_help()
sys.exit(1)
if __name__ == '__main__':
main()
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 197k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |