fix: decode raw gzip tmdb responses

app/modules/themoviedb/tmdbv3api/tmdb.py

@@ -1,6 +1,8 @@
 # -*- coding: utf-8 -*-
 
 import asyncio
+import gzip
+import json as jsonlib
 import logging
 import time
 from copy import deepcopy
@@ -19,6 +21,7 @@ logger = logging.getLogger(__name__)
 
 class TMDb(object):
     _RESPONSE_SNAPSHOT_MARKER = "__mp_tmdb_response_snapshot__"
+    _JSON_DECODE_FAILED = object()
 
     def __init__(self, session=None, language=None):
         self._api_key = settings.TMDB_API_KEY
@@ -164,9 +167,62 @@ class TMDb(object):
             return response.json()
         except (ValueError, UnicodeDecodeError) as err:
             # httpx.Response.json() 在响应体是压缩字节或错误编码时会直接抛 UnicodeDecodeError，
-            # 这里统一收敛成 TMDbException，避免上层把脏响应当作未捕获异常。
+            # 先尝试兼容未被客户端解压的 gzip JSON，仍失败时再收敛成 TMDbException。
+            json_data = cls._decode_compressed_response_json(response)
+            if json_data is not cls._JSON_DECODE_FAILED:
+                return json_data
             raise TMDbException(cls._build_invalid_json_message(response)) from err
 
+    @classmethod
+    def _decode_compressed_response_json(cls, response):
+        """
+        尝试解析未被HTTP客户端自动解压的压缩JSON响应。
+        """
+        response_content = getattr(response, "content", b"") or b""
+        if isinstance(response_content, str):
+            response_content = response_content.encode("utf-8")
+        if not isinstance(response_content, (bytes, bytearray)):
+            return cls._JSON_DECODE_FAILED
+
+        content_bytes = bytes(response_content)
+        content_encoding = cls._get_header_value(
+            getattr(response, "headers", {}) or {},
+            "Content-Encoding",
+        ) or ""
+        encodings = {
+            encoding.strip().lower()
+            for encoding in str(content_encoding).split(",")
+            if encoding.strip()
+        }
+        if "gzip" not in encodings and not content_bytes.startswith(b"\x1f\x8b"):
+            return cls._JSON_DECODE_FAILED
+
+        try:
+            return jsonlib.loads(gzip.decompress(content_bytes))
+        except (OSError, EOFError, ValueError, UnicodeDecodeError):
+            return cls._JSON_DECODE_FAILED
+
+    @staticmethod
+    def _get_header_value(headers, name):
+        """
+        从不同响应头对象中按大小写兼容读取指定响应头。
+        """
+        try:
+            value = headers.get(name)
+        except AttributeError:
+            return None
+        if value is not None:
+            return value
+
+        lower_name = name.lower()
+        try:
+            for header_name, header_value in headers.items():
+                if str(header_name).lower() == lower_name:
+                    return header_value
+        except AttributeError:
+            return None
+        return None
+
     @staticmethod
     def _build_invalid_json_message(response):
         """
@@ -174,10 +230,7 @@ class TMDb(object):
         """
         status_code = getattr(response, "status_code", None)
         headers = getattr(response, "headers", {}) or {}
-        try:
+        content_type = TMDb._get_header_value(headers, "Content-Type")
-            content_type = headers.get("content-type") or headers.get("Content-Type")
-        except AttributeError:
-            content_type = None
 
         try:
             response_text = getattr(response, "text", "") or ""
@@ -194,7 +247,7 @@ class TMDb(object):
             message_parts.append(f"HTTP状态码：{status_code}")
         if content_type:
             message_parts.append(f"Content-Type：{content_type}")
-        content_encoding = headers.get("content-encoding") or headers.get("Content-Encoding")
+        content_encoding = TMDb._get_header_value(headers, "Content-Encoding")
         if content_encoding:
             message_parts.append(f"Content-Encoding：{content_encoding}")
         if response_text:

tests/test_tmdb_response_cache.py

@@ -1,5 +1,7 @@
 import asyncio
+import gzip
 import importlib.util
+import json
 import pickle
 import sys
 from contextlib import asynccontextmanager, contextmanager
@@ -144,13 +146,14 @@ class _UnicodeDecodeErrorResponse:
     模拟 httpx.Response.json() 直接抛 UnicodeDecodeError 的异常响应。
     """
 
-    def __init__(self):
+    def __init__(self, content: bytes = b"\x8b"):
         """
         初始化一个带有压缩响应特征的伪响应对象。
         """
         self.headers = {"Content-Type": "application/json", "Content-Encoding": "gzip"}
         self.status_code = 200
         self.text = ""
+        self.content = content
 
     def json(self):
         """
@@ -159,6 +162,18 @@ class _UnicodeDecodeErrorResponse:
         raise UnicodeDecodeError("utf-8", b"\x8b", 1, 2, "invalid start byte")
 
 
+class _GzipJsonResponse(_UnicodeDecodeErrorResponse):
+    """
+    模拟响应对象带着尚未解压的 gzip JSON 字节。
+    """
+
+    def __init__(self, payload):
+        """
+        将JSON载荷压缩成 gzip 字节，复现代理返回原始压缩体的情况。
+        """
+        super().__init__(gzip.compress(json.dumps(payload).encode("utf-8")))
+
+
 class TmdbResponseCacheTest(TestCase):
     def test_request_returns_pickleable_snapshot(self):
         tmdb = TMDb()
@@ -217,6 +232,20 @@ class TmdbResponseCacheTest(TestCase):
         with self.assertRaisesRegex(TMDbException, "不是有效JSON.*Content-Encoding：gzip"):
             TMDb.request.__wrapped__(tmdb, "GET", "https://example.com", None, None)
 
+    def test_request_decodes_raw_gzip_json_response(self):
+        """
+        客户端未自动解压 gzip JSON 时，应手动解压后正常进入响应快照。
+        """
+        tmdb = TMDb()
+        tmdb._req.get_res = lambda *args, **kwargs: _GzipJsonResponse(
+            {"page": 1, "results": [{"id": 100}]}
+        )
+
+        result = TMDb.request.__wrapped__(tmdb, "GET", "https://example.com", None, None)
+
+        self.assertTrue(result[TMDb._RESPONSE_SNAPSHOT_MARKER])
+        self.assertEqual(result["json"]["results"], [{"id": 100}])
+
     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。