diff --git a/app/modules/themoviedb/tmdbv3api/tmdb.py b/app/modules/themoviedb/tmdbv3api/tmdb.py index acca40c1..ae7f7c11 100644 --- a/app/modules/themoviedb/tmdbv3api/tmdb.py +++ b/app/modules/themoviedb/tmdbv3api/tmdb.py @@ -130,11 +130,16 @@ class TMDb(object): @classmethod def _snapshot_response(cls, response): + """ + 生成可缓存的响应快照，并在入缓存前拦截明显异常的TMDB响应结构。 + """ + json_data = cls._decode_response_json(response) + cls._validate_json_response(json_data) # Redis 不能稳定序列化 requests/httpx 响应对象，缓存里只保留当前流程会用到的数据。 return { cls._RESPONSE_SNAPSHOT_MARKER: True, "headers": dict(response.headers.items()), - "json": response.json(), + "json": json_data, } @classmethod @@ -148,7 +153,78 @@ class TMDb(object): if isinstance(response, dict) and response.get(cls._RESPONSE_SNAPSHOT_MARKER): # 调用方会补充 media_type 等字段，缓存快照必须隔离这些原地修改。 return deepcopy(response.get("json")) - return response.json() + return cls._decode_response_json(response) + + @classmethod + def _decode_response_json(cls, response): + """ + 解析TMDB响应JSON，并把空响应或代理错误页统一转换为TMDB异常。 + """ + try: + return response.json() + except ValueError as err: + raise TMDbException(cls._build_invalid_json_message(response)) from err + + @staticmethod + def _build_invalid_json_message(response): + """ + 生成非JSON响应的诊断信息，避免日志只保留JSONDecodeError文本。 + """ + status_code = getattr(response, "status_code", None) + headers = getattr(response, "headers", {}) or {} + try: + content_type = headers.get("content-type") or headers.get("Content-Type") + except AttributeError: + content_type = None + + try: + response_text = getattr(response, "text", "") or "" + except Exception as err: # pragma: no cover - 防御异常响应对象 + response_text = f"<读取响应内容失败：{err!r}>" + if not isinstance(response_text, str): + response_text = repr(response_text) + response_text = response_text.strip() + if len(response_text) > 200: + response_text = f"{response_text[:200]}..." + + message_parts = ["TheMovieDb 返回数据不是有效JSON"] + if status_code is not None: + message_parts.append(f"HTTP状态码：{status_code}") + if content_type: + message_parts.append(f"Content-Type：{content_type}") + if response_text: + message_parts.append(f"响应内容：{response_text!r}") + else: + message_parts.append("响应内容为空") + return "，".join(message_parts) + + @staticmethod + def _validate_json_response(json_data): + """ + 校验TMDB响应JSON顶层结构，避免代理错误页等标量值继续按字典解析。 + """ + if isinstance(json_data, (dict, list)): + return + + payload_preview = repr(json_data) + if len(payload_preview) > 200: + payload_preview = f"{payload_preview[:200]}..." + raise TMDbException( + "TheMovieDb 返回数据格式异常：期望JSON对象或数组，" + f"实际为{type(json_data).__name__}，内容：{payload_preview}" + ) + + @staticmethod + def _get_json_key(json_data, key): + """ + 从TMDB对象响应中读取指定字段，避免异常顶层结构触发AttributeError。 + """ + if not isinstance(json_data, dict): + raise TMDbException( + "TheMovieDb 返回数据格式异常：" + f"期望JSON对象包含字段 {key!r}，实际为{type(json_data).__name__}" + ) + return json_data.get(key) def cache_clear(self): return self.request.cache_clear() @@ -190,6 +266,12 @@ class TMDb(object): return 0 def _process_json_response(self, json_data, is_async=False): + """ + 从TMDB对象响应中记录分页信息；数组响应没有分页字段，直接跳过。 + """ + if not isinstance(json_data, dict): + return + if "page" in json_data: self._page = json_data["page"] @@ -201,6 +283,12 @@ class TMDb(object): @staticmethod def _handle_errors(json_data): + """ + 将TMDB标准错误字段转换为统一异常，非对象响应由结构校验提前处理。 + """ + if not isinstance(json_data, dict): + return + if "errors" in json_data: raise TMDbException(json_data["errors"]) @@ -228,11 +316,12 @@ class TMDb(object): return self._request_obj(action, params, False, method, data, json, key) json_data = self._get_response_json(req) + self._validate_json_response(json_data) self._process_json_response(json_data, is_async=False) self._handle_errors(json_data) if key: - return json_data.get(key) + return self._get_json_key(json_data, key) return json_data async def _async_request_obj(self, action, params="", call_cached=True, @@ -256,11 +345,12 @@ class TMDb(object): return await self._async_request_obj(action, params, False, method, data, json, key) json_data = self._get_response_json(req) + self._validate_json_response(json_data) self._process_json_response(json_data, is_async=True) self._handle_errors(json_data) if key: - return json_data.get(key) + return self._get_json_key(json_data, key) return json_data def close(self): diff --git a/tests/test_tmdb_response_cache.py b/tests/test_tmdb_response_cache.py index 98c32e2f..af33c31e 100644 --- a/tests/test_tmdb_response_cache.py +++ b/tests/test_tmdb_response_cache.py @@ -124,12 +124,15 @@ def _load_tmdb_class(): TMDb = _load_tmdb_class() +TMDbException = sys.modules["app.modules.themoviedb.tmdbv3api.exceptions"].TMDbException class _FakeResponse: - def __init__(self, payload: dict, headers: dict): + def __init__(self, payload, headers: dict, status_code: int = 200, text: str = ""): self._payload = payload self.headers = headers + self.status_code = status_code + self.text = text self._lock = RLock() def json(self): @@ -152,6 +155,56 @@ class TmdbResponseCacheTest(TestCase): self.assertEqual(result["headers"]["X-RateLimit-Remaining"], "39") pickle.dumps(result) + def test_request_rejects_scalar_json_response(self): + """ + 标量JSON响应不应进入TMDB响应缓存，避免后续按对象解析崩溃。 + """ + tmdb = TMDb() + response = _FakeResponse(payload="upstream error", headers={}) + tmdb._req.get_res = lambda *args, **kwargs: response + + with self.assertRaisesRegex(TMDbException, "返回数据格式异常"): + TMDb.request.__wrapped__(tmdb, "GET", "https://example.com", None, None) + + def test_request_rejects_invalid_json_response(self): + """ + 非JSON响应应转换为TMDbException，调用方可按连接异常统一处理。 + """ + class _InvalidJsonResponse: + headers = {"Content-Type": "text/html"} + status_code = 502 + text = "bad gateway" + + def json(self): + """ + 模拟上游返回无法解析为JSON的响应体。 + """ + raise ValueError("invalid json") + + tmdb = TMDb() + tmdb._req.get_res = lambda *args, **kwargs: _InvalidJsonResponse() + + with self.assertRaisesRegex(TMDbException, "不是有效JSON.*HTTP状态码：502.*bad gateway"): + TMDb.request.__wrapped__(tmdb, "GET", "https://example.com", None, None) + + def test_get_response_json_rejects_invalid_live_response(self): + """ + 未缓存的实时响应解析失败时也应输出统一诊断信息。 + """ + class _InvalidJsonResponse: + headers = {} + status_code = 200 + text = "" + + def json(self): + """ + 模拟HTTP 200但响应体为空的情况。 + """ + raise ValueError("empty") + + with self.assertRaisesRegex(TMDbException, "不是有效JSON.*响应内容为空"): + TMDb._get_response_json(_InvalidJsonResponse()) + def test_async_request_returns_pickleable_snapshot(self): tmdb = TMDb() response = _FakeResponse( @@ -229,3 +282,40 @@ class TmdbResponseCacheTest(TestCase): self.assertEqual(second_results[0]["media_type"], "movie") self.assertIsNot(first_results, second_results) self.assertIsNot(first_results[0], second_results[0]) + + def test_request_obj_rejects_scalar_snapshot_before_key_lookup(self): + """ + 旧缓存中的标量快照不应在读取results字段时触发AttributeError。 + """ + tmdb = TMDb() + snapshot = { + TMDb._RESPONSE_SNAPSHOT_MARKER: True, + "headers": {"x-ratelimit-remaining": "39", "x-ratelimit-reset": "1234567890"}, + "json": "upstream error", + } + tmdb.request = lambda *args, **kwargs: snapshot + + with self.assertRaisesRegex(TMDbException, "返回数据格式异常"): + tmdb._request_obj("/search/movie", key="results") + + def test_async_request_obj_rejects_scalar_snapshot_before_key_lookup(self): + """ + 异步对象请求读取旧标量快照时也应走统一TMDB异常路径。 + """ + tmdb = TMDb() + snapshot = { + TMDb._RESPONSE_SNAPSHOT_MARKER: True, + "headers": {"x-ratelimit-remaining": "39", "x-ratelimit-reset": "1234567890"}, + "json": "upstream error", + } + + async def _fake_async_request(*args, **kwargs): + """ + 模拟异步请求命中已缓存的异常快照。 + """ + return snapshot + + tmdb.async_request = _fake_async_request + + with self.assertRaisesRegex(TMDbException, "返回数据格式异常"): + asyncio.run(tmdb._async_request_obj("/search/movie", key="results"))