v1.7.3 前端新增后端选择器 + 暴露 /api/backends + 接通 chrome-cdp 后端
- 新增 GET /api/backends 返回可用后端/默认顺序/说明,供前端动态渲染 - 前端表单新增「捕获后端」下拉框: 自动(默认,优先agent-browser超时自动回退playwright)/agent-browser/playwright/chrome-cdp,不可用的置灰 - 接通 chrome-cdp 后端: 之前函数已实现但 capture_webpage 未接线(传 backend=chrome-cdp 会报 Unknown backend),现支持 screenshot/html/text,连接已开的 Chrome(--remote-debugging-port) - 前端 currentData 携带 backend 字段
This commit is contained in:
@@ -708,6 +708,14 @@ async def capture_with_cdp(
|
|||||||
"html": html
|
"html": html
|
||||||
})
|
})
|
||||||
|
|
||||||
|
elif action == "text":
|
||||||
|
text = await page.evaluate("document.body.innerText")
|
||||||
|
result_data.append({
|
||||||
|
"url": current_url,
|
||||||
|
"title": title,
|
||||||
|
"text": clean_text(text)
|
||||||
|
})
|
||||||
|
|
||||||
if result_data:
|
if result_data:
|
||||||
return {"success": True, "pages": result_data}
|
return {"success": True, "pages": result_data}
|
||||||
else:
|
else:
|
||||||
@@ -983,13 +991,14 @@ def capture_webpage(
|
|||||||
full_page: bool = False,
|
full_page: bool = False,
|
||||||
viewport: dict = None,
|
viewport: dict = None,
|
||||||
wait_time: int = 2000,
|
wait_time: int = 2000,
|
||||||
backend: str = "auto"
|
backend: str = "auto",
|
||||||
|
cdp_port: int = 9222
|
||||||
):
|
):
|
||||||
"""
|
"""
|
||||||
捕获网页
|
捕获网页
|
||||||
|
|
||||||
Args:
|
Args:
|
||||||
backend: "auto", "agent-browser", "playwright"
|
backend: "auto", "agent-browser", "playwright", "chrome-cdp"
|
||||||
"""
|
"""
|
||||||
# 选择后端
|
# 选择后端
|
||||||
if backend == "auto":
|
if backend == "auto":
|
||||||
@@ -1044,6 +1053,33 @@ def capture_webpage(
|
|||||||
except Exception as e:
|
except Exception as e:
|
||||||
return {"success": False, "error": str(e)}
|
return {"success": False, "error": str(e)}
|
||||||
|
|
||||||
|
elif backend == "chrome-cdp":
|
||||||
|
# 连接已打开的 Chrome(需 --remote-debugging-port=<cdp_port>),归一化为单结果
|
||||||
|
if not PLAYWRIGHT_AVAILABLE:
|
||||||
|
return {"success": False, "error": "Playwright not installed (chrome-cdp 依赖 playwright)"}
|
||||||
|
try:
|
||||||
|
loop = asyncio.new_event_loop()
|
||||||
|
asyncio.set_event_loop(loop)
|
||||||
|
r = loop.run_until_complete(
|
||||||
|
capture_with_cdp(url_hint=url, action=action, cdp_port=cdp_port)
|
||||||
|
)
|
||||||
|
loop.close()
|
||||||
|
except Exception as e:
|
||||||
|
return {"success": False, "error": str(e)}
|
||||||
|
if not r.get("success"):
|
||||||
|
return r
|
||||||
|
if not r.get("pages"):
|
||||||
|
return {"success": False, "error": "chrome-cdp 未找到匹配页面(请确认 Chrome 已开 --remote-debugging-port)"}
|
||||||
|
p0 = r["pages"][0]
|
||||||
|
res = {"success": True, "title": p0.get("title", ""), "backend": "chrome-cdp"}
|
||||||
|
if action == "screenshot":
|
||||||
|
res["file_path"] = p0["file_path"]
|
||||||
|
elif action == "html":
|
||||||
|
res["html"] = p0["html"]
|
||||||
|
elif action == "text":
|
||||||
|
res["text"] = p0["text"]
|
||||||
|
return res
|
||||||
|
|
||||||
else:
|
else:
|
||||||
return {"success": False, "error": f"Unknown backend: {backend}"}
|
return {"success": False, "error": f"Unknown backend: {backend}"}
|
||||||
|
|
||||||
@@ -1054,6 +1090,37 @@ def index():
|
|||||||
return render_template('index.html')
|
return render_template('index.html')
|
||||||
|
|
||||||
|
|
||||||
|
@app.route('/api/backends')
|
||||||
|
def api_backends():
|
||||||
|
"""返回可用后端及默认选择顺序,供前端渲染选择器"""
|
||||||
|
available = []
|
||||||
|
if AGENT_BROWSER_AVAILABLE:
|
||||||
|
available.append("agent-browser")
|
||||||
|
if PLAYWRIGHT_AVAILABLE:
|
||||||
|
available.append("playwright")
|
||||||
|
# chrome-cdp 依赖 playwright 的 connect_over_cdp
|
||||||
|
available.append("chrome-cdp")
|
||||||
|
unavailable = []
|
||||||
|
if not AGENT_BROWSER_AVAILABLE:
|
||||||
|
unavailable.append("agent-browser")
|
||||||
|
if not PLAYWRIGHT_AVAILABLE:
|
||||||
|
unavailable.append("playwright")
|
||||||
|
unavailable.append("chrome-cdp")
|
||||||
|
return jsonify({
|
||||||
|
"success": True,
|
||||||
|
"default": "auto",
|
||||||
|
"available": available,
|
||||||
|
"unavailable": unavailable,
|
||||||
|
"priority": available + unavailable,
|
||||||
|
"labels": {
|
||||||
|
"auto": "自动(推荐:优先 agent-browser,打开超时自动回退 Playwright)",
|
||||||
|
"agent-browser": "agent-browser(Rust 版,反爬强、快,默认首选)",
|
||||||
|
"playwright": "Playwright(Python 版,domcontentloaded 更宽容,慢页面更稳)",
|
||||||
|
"chrome-cdp": "Chrome CDP(连接已打开的 Chrome,需 --remote-debugging-port 启动)"
|
||||||
|
}
|
||||||
|
})
|
||||||
|
|
||||||
|
|
||||||
@app.route('/api')
|
@app.route('/api')
|
||||||
def api_info():
|
def api_info():
|
||||||
"""API信息"""
|
"""API信息"""
|
||||||
@@ -1066,6 +1133,7 @@ def api_info():
|
|||||||
},
|
},
|
||||||
"endpoints": {
|
"endpoints": {
|
||||||
"/api/capture": "POST - Capture webpage (screenshot/html/text/smart) + 自动入库历史(记录调用者/方式/原始HTML)",
|
"/api/capture": "POST - Capture webpage (screenshot/html/text/smart) + 自动入库历史(记录调用者/方式/原始HTML)",
|
||||||
|
"/api/backends": "GET - 可用后端列表与默认顺序(供前端渲染后端选择器)",
|
||||||
"/api/history": "GET - 历史记录分页列表 (?page&page_size&action&search)",
|
"/api/history": "GET - 历史记录分页列表 (?page&page_size&action&search)",
|
||||||
"/api/history/<id>": "GET - 历史记录详情 / DELETE - 删除记录",
|
"/api/history/<id>": "GET - 历史记录详情 / DELETE - 删除记录",
|
||||||
"/api/history/<id>/file": "GET - 读取历史截图文件",
|
"/api/history/<id>/file": "GET - 读取历史截图文件",
|
||||||
@@ -1320,7 +1388,8 @@ def capture():
|
|||||||
full_page=full_page,
|
full_page=full_page,
|
||||||
viewport=viewport,
|
viewport=viewport,
|
||||||
wait_time=wait_time,
|
wait_time=wait_time,
|
||||||
backend=backend
|
backend=backend,
|
||||||
|
cdp_port=cdp_port
|
||||||
)
|
)
|
||||||
|
|
||||||
backend_used = result.get("backend", backend)
|
backend_used = result.get("backend", backend)
|
||||||
|
|||||||
+35
-1
@@ -624,6 +624,16 @@
|
|||||||
</div>
|
</div>
|
||||||
</div>
|
</div>
|
||||||
|
|
||||||
|
<div class="form-group">
|
||||||
|
<label>捕获后端</label>
|
||||||
|
<select id="backend" style="max-width:100%;">
|
||||||
|
<option value="auto">加载中…</option>
|
||||||
|
</select>
|
||||||
|
<small id="backendHint" style="color: #666; display: block; margin-top: 4px;">
|
||||||
|
自动 = 优先 agent-browser,打开页面超时自动回退 Playwright(推荐)
|
||||||
|
</small>
|
||||||
|
</div>
|
||||||
|
|
||||||
<div class="form-group">
|
<div class="form-group">
|
||||||
<label>等待设置(重要!)</label>
|
<label>等待设置(重要!)</label>
|
||||||
<div class="row">
|
<div class="row">
|
||||||
@@ -844,12 +854,35 @@ DELETE /api/history/<id> // 删除历史</code></pre>
|
|||||||
let smartConfig = null;
|
let smartConfig = null;
|
||||||
let lastSmartHistoryId = null;
|
let lastSmartHistoryId = null;
|
||||||
|
|
||||||
// 页面加载时读取历史 + 智能配置
|
// 页面加载时读取历史 + 智能配置 + 可用后端
|
||||||
window.addEventListener('load', () => {
|
window.addEventListener('load', () => {
|
||||||
loadHistory(1);
|
loadHistory(1);
|
||||||
loadSmartConfig();
|
loadSmartConfig();
|
||||||
|
loadBackends();
|
||||||
});
|
});
|
||||||
|
|
||||||
|
async function loadBackends() {
|
||||||
|
const sel = document.getElementById('backend');
|
||||||
|
const hint = document.getElementById('backendHint');
|
||||||
|
try {
|
||||||
|
const res = await fetch('/api/backends');
|
||||||
|
const data = await res.json();
|
||||||
|
if (!data.success) throw new Error(data.error || '加载失败');
|
||||||
|
const labels = data.labels || {};
|
||||||
|
const opts = ['auto'].concat(data.available || []);
|
||||||
|
sel.innerHTML = opts.map(v =>
|
||||||
|
`<option value="${v}">${escapeHtml(v)} — ${escapeHtml(labels[v] || '')}</option>`
|
||||||
|
).join('') + (data.unavailable || []).map(v =>
|
||||||
|
`<option value="${v}" disabled>${escapeHtml(v)} — ${escapeHtml(labels[v] || '')}(不可用)</option>`
|
||||||
|
).join('');
|
||||||
|
sel.value = 'auto';
|
||||||
|
hint.textContent = labels['auto'] || hint.textContent;
|
||||||
|
} catch (err) {
|
||||||
|
sel.innerHTML = '<option value="auto">自动</option>';
|
||||||
|
hint.textContent = '后端列表加载失败:' + err.message;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
// 操作类型切换:智能模式显示提示条
|
// 操作类型切换:智能模式显示提示条
|
||||||
document.querySelectorAll('input[name="action"]').forEach(r => {
|
document.querySelectorAll('input[name="action"]').forEach(r => {
|
||||||
r.addEventListener('change', () => {
|
r.addEventListener('change', () => {
|
||||||
@@ -957,6 +990,7 @@ DELETE /api/history/<id> // 删除历史</code></pre>
|
|||||||
scroll_delay: scrollDelay,
|
scroll_delay: scrollDelay,
|
||||||
full_page: fullPage,
|
full_page: fullPage,
|
||||||
wait_time: waitTime,
|
wait_time: waitTime,
|
||||||
|
backend: document.getElementById('backend').value || 'auto',
|
||||||
viewport: {
|
viewport: {
|
||||||
width: viewportWidth,
|
width: viewportWidth,
|
||||||
height: viewportHeight
|
height: viewportHeight
|
||||||
|
|||||||
Reference in New Issue
Block a user