优化浏览器context复用

2026-06-09 16:04:44 +08:00
parent bc0f7161c0
commit 3b8bdf09f7
4 changed files with 399 additions and 93 deletions
@@ -0,0 +1,124 @@
+"""保活上下文（cf_clearance 复用）逻辑的单元测试（不启动真实 Chromium）。"""
+from __future__ import annotations
+
+from contextlib import asynccontextmanager
+from types import SimpleNamespace
+
+import pytest
+
+from app.core.config import Settings
+from app.services.cloudflare_session import CloudflareSessionManager
+from app.services.session_store import SessionStore
+
+_DETAIL_URL = "https://www.suruga-ya.jp/product/detail/{}"
+
+
+class _FakePage:
+    """最小页面替身：goto 状态由所属上下文是否被标记失败决定。"""
+
+    def __init__(self, pool: "_FakeBrowserPool", context: "_FakeContext"):
+        self._pool = pool
+        self._context = context
+
+    async def goto(self, url: str, **_kw):
+        status = 403 if self._context.idx in self._pool.fail_idxs else 200
+        return SimpleNamespace(status=status)
+
+    async def title(self) -> str:
+        return "ok"
+
+    async def content(self) -> str:
+        return self._pool.content_html
+
+    async def screenshot(self, **_kw):  # pragma: no cover - 非 200 分支才用到
+        return None
+
+    async def wait_for_timeout(self, _ms):  # pragma: no cover
+        return None
+
+
+class _FakeContext:
+    def __init__(self, idx: int):
+        self.idx = idx
+        self.closed = False
+
+    async def cookies(self):
+        return [{"name": "cf_clearance", "value": f"cv-{self.idx}"}]
+
+    async def close(self):
+        self.closed = True
+
+
+class _FakeBrowserPool:
+    """记录上下文创建次数，并允许把指定上下文标记为「复用必失败」。"""
+
+    def __init__(self):
+        self.content_html = "<html>ok</html>"
+        self.create_calls = 0
+        self.contexts: list[_FakeContext] = []
+        self.fail_idxs: set[int] = set()
+
+    async def create_persistent_context(self, storage_state=None) -> _FakeContext:
+        self.create_calls += 1
+        ctx = _FakeContext(self.create_calls)
+        self.contexts.append(ctx)
+        return ctx
+
+    @asynccontextmanager
+    async def page_in_context(self, context: _FakeContext):
+        yield _FakePage(self, context)
+
+    async def close_context(self, context: _FakeContext):
+        await context.close()
+
+
+def _make_manager(pool: _FakeBrowserPool) -> CloudflareSessionManager:
+    settings = Settings(session_ttl_seconds=900, cloudflare_reuse_wait_timeout_seconds=8.0)
+    return CloudflareSessionManager(settings, pool, SessionStore(settings))
+
+
+@pytest.mark.asyncio
+async def test_warm_context_is_reused_without_rebuild():
+    pool = _FakeBrowserPool()
+    mgr = _make_manager(pool)
+
+    html1, sess1 = await mgr.fetch_html(_DETAIL_URL.format(1))
+    assert html1 == pool.content_html
+    assert pool.create_calls == 1  # 首次构建
+
+    # 第二次请求（不同商品）应复用同一保活上下文，不再新建
+    html2, sess2 = await mgr.fetch_html(_DETAIL_URL.format(2))
+    assert html2 == pool.content_html
+    assert pool.create_calls == 1
+    assert sess2.session_id == sess1.session_id
+
+
+@pytest.mark.asyncio
+async def test_warm_context_rebuilds_and_closes_on_failure():
+    pool = _FakeBrowserPool()
+    mgr = _make_manager(pool)
+
+    _, sess1 = await mgr.fetch_html(_DETAIL_URL.format(1))
+    assert pool.create_calls == 1
+
+    # 让首个上下文在后续复用导航时返回 403（模拟 clearance 失效）
+    pool.fail_idxs.add(1)
+
+    html2, sess2 = await mgr.fetch_html(_DETAIL_URL.format(2))
+    assert html2 == pool.content_html
+    assert pool.create_calls == 2                 # 复用失败后重建
+    assert pool.contexts[0].closed is True        # 旧上下文已退役关闭
+    assert sess2.session_id != sess1.session_id   # 换了新会话
+
+
+@pytest.mark.asyncio
+async def test_close_releases_all_warm_contexts():
+    pool = _FakeBrowserPool()
+    mgr = _make_manager(pool)
+
+    await mgr.fetch_html(_DETAIL_URL.format(1))
+    assert pool.contexts[0].closed is False
+
+    await mgr.close()
+    assert pool.contexts[0].closed is True
+    assert mgr._warm_contexts == {}