sycommon-python-lib 0.2.7a43__tar.gz → 0.2.7a44__tar.gz

{sycommon_python_lib-0.2.7a43 → sycommon_python_lib-0.2.7a44}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sycommon-python-lib
-Version: 0.2.7a43
+Version: 0.2.7a44
 Summary: Add your description here
 Requires-Python: >=3.11
 Description-Content-Type: text/markdown

{sycommon_python_lib-0.2.7a43 → sycommon_python_lib-0.2.7a44}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "sycommon-python-lib"
-version = "0.2.7a43"
+version = "0.2.7a44"
 description = "Add your description here"
 readme = "README.md"
 requires-python = ">=3.11"

{sycommon_python_lib-0.2.7a43 → sycommon_python_lib-0.2.7a44}/src/sycommon/middleware/traceid.py

@@ -1,4 +1,5 @@
 import json
+import os
 import re
 from typing import Dict, Any
 from fastapi import Request, Response
@@ -6,6 +7,11 @@ from sycommon.logging.kafka_log import SYLogger
 from sycommon.tools.merge_headers import merge_headers
 from sycommon.tools.snowflake import Snowflake
 
+# 🔑 OOM 防护：响应体聚合/重写阈值。超过此大小（attachment 下载、大二进制预览、PPTX 图片数组）
+# 的响应跳过整包聚合，避免单请求在内存里同时存原始 bytes + json 解析结果 + 重编码 bytes 三份拷贝。
+# 默认 2MB；可经环境变量 TRACEID_MAX_AGGREGATE_BODY 覆盖。
+_MAX_AGGREGATE_BODY = int(os.environ.get("TRACEID_MAX_AGGREGATE_BODY", 2 * 1024 * 1024))
+
 
 def setup_trace_id_handler(app):
     @app.middleware("http")
@@ -188,38 +194,86 @@ def setup_trace_id_handler(app):
                         pass
 
             # 处理响应体
-            response_body = b""
+            # 🔑 OOM 防护：对大响应体（文件下载/二进制预览/PPTX 图片数组）跳过整包聚合。
+            # 否则上百 MB 的 attachment 会被 response_body += chunk 全量读进内存，再 json.loads/dumps
+            # 产生多份拷贝，单请求即可吃掉数百 MB。大响应只改写 header 不重写 body。
+            content_disposition = response.headers.get(
+                "content-disposition", "").lower()
+            _is_attachment = content_disposition.startswith("attachment")
             try:
-                async for chunk in response.body_iterator:
-                    response_body += chunk
-
-                content_disposition = response.headers.get(
-                    "content-disposition", "").lower()
+                content_length_str = response.headers.get("content-length", "")
+                content_length = int(content_length_str) if content_length_str.isdigit() else None
+            except (TypeError, ValueError):
+                content_length = None
+            # 超过阈值(默认 2MB)或 attachment 下载 → 跳过聚合，直接透传 body（仍注入 traceId header）
+            _skip_aggregate = (
+                _is_attachment
+                or (content_length is not None and content_length > _MAX_AGGREGATE_BODY)
+            )
 
-                # JSON 响应体注入 traceId
-                if "application/json" in response_content_type and not content_disposition.startswith("attachment"):
-                    try:
-                        data = json.loads(response_body)
-                        new_body = response_body
-                        if isinstance(data, dict):
-                            data["traceId"] = trace_id
-                            new_body = json.dumps(
-                                data, ensure_ascii=False).encode()
-
-                        # 重建 Response 以更新 Body 和 Content-Length
-                        response = Response(
-                            content=new_body,
-                            status_code=response.status_code,
-                            headers=dict(response.headers),
-                            media_type=response.media_type
-                        )
-                        response.headers["content-length"] = str(len(new_body))
-                        response.headers["x-traceid-header"] = trace_id
-                        # 恢复 CORS
-                        for k, v in cors_headers.items():
-                            response.headers[k] = v
-                    except json.JSONDecodeError:
-                        # 非 JSON 或解析失败，仅更新长度
+            response_body = b""
+            if _skip_aggregate:
+                # 大响应：不读 body，保留原 StreamingResponse 的 body_iterator 透传。
+                # 仅补 header，不动 body/Content-Length。
+                response.headers["x-traceid-header"] = trace_id
+                for k, v in cors_headers.items():
+                    response.headers[k] = v
+                response_log_body = f"<skipped: large response, content-length={content_length}, attachment={_is_attachment}>"
+            else:
+                try:
+                    async for chunk in response.body_iterator:
+                        response_body += chunk
+                        # 聚合期再兜一道防线上限：流式 body 没有 Content-Length，读到超限就放弃重写
+                        if len(response_body) > _MAX_AGGREGATE_BODY:
+                            _skip_aggregate = True
+                            break
+                except StopAsyncIteration:
+                    pass
+
+                if _skip_aggregate:
+                    # 流式响应超限，恢复为透传已读部分（不重写 JSON，保留原 body）
+                    response.headers["x-traceid-header"] = trace_id
+                    for k, v in cors_headers.items():
+                        response.headers[k] = v
+                    response_log_body = f"<skipped: large response during aggregate, size={len(response_body)}>"
+                else:
+                    # JSON 响应体注入 traceId
+                    if "application/json" in response_content_type and not _is_attachment:
+                        try:
+                            data = json.loads(response_body)
+                            new_body = response_body
+                            if isinstance(data, dict):
+                                data["traceId"] = trace_id
+                                new_body = json.dumps(
+                                    data, ensure_ascii=False).encode()
+
+                            # 重建 Response 以更新 Body 和 Content-Length
+                            response = Response(
+                                content=new_body,
+                                status_code=response.status_code,
+                                headers=dict(response.headers),
+                                media_type=response.media_type
+                            )
+                            response.headers["content-length"] = str(len(new_body))
+                            response.headers["x-traceid-header"] = trace_id
+                            # 恢复 CORS
+                            for k, v in cors_headers.items():
+                                response.headers[k] = v
+                        except json.JSONDecodeError:
+                            # 非 JSON 或解析失败，仅更新长度
+                            response = Response(
+                                content=response_body,
+                                status_code=response.status_code,
+                                headers=dict(response.headers),
+                                media_type=response.media_type
+                            )
+                            response.headers["content-length"] = str(
+                                len(response_body))
+                            response.headers["x-traceid-header"] = trace_id
+                            for k, v in cors_headers.items():
+                                response.headers[k] = v
+                    else:
+                        # 非 JSON 响应
                         response = Response(
                             content=response_body,
                             status_code=response.status_code,
@@ -231,21 +285,6 @@ def setup_trace_id_handler(app):
                         response.headers["x-traceid-header"] = trace_id
                         for k, v in cors_headers.items():
                             response.headers[k] = v
-                else:
-                    # 非 JSON 响应
-                    response = Response(
-                        content=response_body,
-                        status_code=response.status_code,
-                        headers=dict(response.headers),
-                        media_type=response.media_type
-                    )
-                    response.headers["content-length"] = str(
-                        len(response_body))
-                    response.headers["x-traceid-header"] = trace_id
-                    for k, v in cors_headers.items():
-                        response.headers[k] = v
-            except StopAsyncIteration:
-                pass
 
             # 构建响应日志
             response_log_body = response_body.decode('utf-8', errors='ignore')

{sycommon_python_lib-0.2.7a43 → sycommon_python_lib-0.2.7a44}/src/sycommon_python_lib.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sycommon-python-lib
-Version: 0.2.7a43
+Version: 0.2.7a44
 Summary: Add your description here
 Requires-Python: >=3.11
 Description-Content-Type: text/markdown