e0a2b323d9
- obsidian-kb: Add OCR processors (easyocr, tesseract) - obsidian-kb: Fix subtitle detection with test files - obsidian-kb: Add cookie verification script - Add find-skills skill for discovering capabilities - Add video processing scripts - Update clawhub lock file
150 lines
5.6 KiB
Python
150 lines
5.6 KiB
Python
#!/usr/bin/env python3
|
|
# -*- coding: utf-8 -*-
|
|
"""
|
|
测试动态字幕加载 - 播放视频并监听字幕请求
|
|
"""
|
|
import asyncio
|
|
import json
|
|
from playwright.async_api import async_playwright
|
|
|
|
async def test_dynamic_subtitle(bvid):
|
|
"""测试动态字幕加载"""
|
|
url = f"https://www.bilibili.com/video/{bvid}"
|
|
|
|
async with async_playwright() as playwright:
|
|
browser = await playwright.chromium.launch(headless=True)
|
|
context = await browser.new_context()
|
|
page = await context.new_page()
|
|
|
|
print(f"正在测试视频: {url}")
|
|
print("=" * 80)
|
|
|
|
# 监听所有网络请求,查找字幕相关的请求
|
|
subtitle_requests = []
|
|
|
|
def handle_request(request):
|
|
url = request.url
|
|
# 记录可能包含字幕的请求
|
|
if 'subtitle' in url.lower() or 'cc' in url.lower():
|
|
subtitle_requests.append({
|
|
'url': url,
|
|
'method': request.method,
|
|
'resource_type': request.resource_type
|
|
})
|
|
print(f"📡 发现字幕请求: {url[:100]}...")
|
|
|
|
page.on('request', handle_request)
|
|
|
|
print("\n1. 访问页面并等待加载...")
|
|
await page.goto(url, wait_until='networkidle', timeout=60000)
|
|
await asyncio.sleep(3)
|
|
|
|
# 模拟播放视频(点击播放按钮)
|
|
print("\n2. 尝试播放视频...")
|
|
try:
|
|
# 尝试点击播放按钮
|
|
play_button = await page.query_selector('.bpx-player-ctrl-play')
|
|
if play_button:
|
|
await play_button.click()
|
|
await asyncio.sleep(5) # 等待视频播放
|
|
print(" ✅ 已点击播放按钮")
|
|
else:
|
|
print(" ⚠️ 未找到播放按钮")
|
|
await asyncio.sleep(5) # 无论如何等待一下
|
|
except Exception as e:
|
|
print(f" ⚠️ 播放异常: {e}")
|
|
await asyncio.sleep(5)
|
|
|
|
# 检查是否有字幕按钮
|
|
print("\n3. 检查字幕按钮...")
|
|
try:
|
|
subtitle_button = await page.query_selector('.bpx-player-ctrl-subtitle')
|
|
if subtitle_button:
|
|
print(" ✅ 找到字幕按钮")
|
|
is_enabled = await subtitle_button.is_enabled()
|
|
print(f" 字幕按钮状态: {'可用' if is_enabled else '不可用'}")
|
|
|
|
# 尝试点击字幕按钮
|
|
if is_enabled:
|
|
await subtitle_button.click()
|
|
await asyncio.sleep(2)
|
|
print(" ✅ 已点击字幕按钮")
|
|
else:
|
|
print(" ❌ 未找到字幕按钮")
|
|
except Exception as e:
|
|
print(f" ⚠️ 检查字幕按钮异常: {e}")
|
|
|
|
# 等待更多网络请求
|
|
print("\n4. 等待字幕请求...")
|
|
await asyncio.sleep(3)
|
|
|
|
# 显示捕获的字幕请求
|
|
print(f"\n5. 捕获的字幕请求 (共 {len(subtitle_requests)} 个):")
|
|
if subtitle_requests:
|
|
for i, req in enumerate(subtitle_requests):
|
|
print(f"\n [{i+1}] {req['method']} {req['resource_type']}")
|
|
print(f" {req['url']}")
|
|
else:
|
|
print(" ❌ 未捕获到字幕请求")
|
|
|
|
# 尝试从页面中提取字幕文本
|
|
print("\n6. 尝试从页面提取字幕文本...")
|
|
try:
|
|
# B站字幕通常在 .bpx-player-subtitle-item 元素中
|
|
subtitle_items = await page.query_selector_all('.bpx-player-subtitle-item')
|
|
|
|
if subtitle_items:
|
|
print(f" ✅ 找到 {len(subtitle_items)} 个字幕元素")
|
|
# 获取前几个字幕
|
|
for i, item in enumerate(subtitle_items[:5]):
|
|
text = await item.inner_text()
|
|
print(f" [{i+1}] {text}")
|
|
else:
|
|
print(" ❌ 未找到字幕元素")
|
|
except Exception as e:
|
|
print(f" ⚠️ 提取字幕文本异常: {e}")
|
|
|
|
# 尝试获取所有文本内容(包含动态加载的字幕)
|
|
print("\n7. 检查页面中是否有字幕相关的数据...")
|
|
try:
|
|
# 尝试从 window 对象中获取字幕数据
|
|
script_result = await page.evaluate('''
|
|
() => {
|
|
const results = {};
|
|
|
|
// 检查不同的位置
|
|
if (window.__playinfo__?.data?.subtitle) {
|
|
results.playinfo_subtitle = window.__playinfo__.data.subtitle;
|
|
}
|
|
|
|
if (window.__subtitle__ || window.__subtitles__) {
|
|
results.subtitle_var = window.__subtitle__ || window.__subtitles__;
|
|
}
|
|
|
|
// 检查播放器实例
|
|
const player = document.querySelector('.bpx-player-container')?.__player__;
|
|
if (player?.subtitlePlugin) {
|
|
results.player_subtitle = Object.keys(player.subtitlePlugin);
|
|
}
|
|
|
|
return results;
|
|
}
|
|
''')
|
|
|
|
if script_result:
|
|
print(f" 找到的字幕数据: {json.dumps(script_result, indent=8, ensure_ascii=False)}")
|
|
else:
|
|
print(" 未找到字幕相关的数据")
|
|
except Exception as e:
|
|
print(f" ⚠️ 检查页面数据异常: {e}")
|
|
|
|
await browser.close()
|
|
|
|
print("\n" + "=" * 80)
|
|
print("测试完成")
|
|
print("=" * 80)
|
|
|
|
if __name__ == '__main__':
|
|
bvid = "BV1KtFzzqEUG"
|
|
asyncio.run(test_dynamic_subtitle(bvid))
|