fix: django content length extraction bug (#160)

Author: daniel-sanche

google/cloud/logging_v2/handlers/_helpers.py

@@ -24,6 +24,7 @@
 
 from google.cloud.logging_v2.handlers.middleware.request import _get_django_request
 
+_DJANGO_CONTENT_LENGTH = "CONTENT_LENGTH"
 _DJANGO_TRACE_HEADER = "HTTP_X_CLOUD_TRACE_CONTEXT"
 _DJANGO_USERAGENT_HEADER = "HTTP_USER_AGENT"
 _DJANGO_REMOTE_ADDR_HEADER = "REMOTE_ADDR"
@@ -93,11 +94,18 @@ def get_request_data_from_django():
 
     if request is None:
         return None, None
+
+    # convert content_length to int if it exists
+    content_length = None
+    try:
+        content_length = int(request.META.get(_DJANGO_CONTENT_LENGTH))
+    except (ValueError, TypeError):
+        content_length = None
     # build http_request
     http_request = {
         "requestMethod": request.method,
         "requestUrl": request.build_absolute_uri(),
-        "requestSize": len(request.body),
+        "requestSize": content_length,
         "userAgent": request.META.get(_DJANGO_USERAGENT_HEADER),
         "remoteIp": request.META.get(_DJANGO_REMOTE_ADDR_HEADER),
         "referer": request.META.get(_DJANGO_REFERER_HEADER),

tests/unit/handlers/test__helpers.py

@@ -172,6 +172,9 @@ def test_http_request_populated(self):
             HTTP_USER_AGENT=expected_agent,
             HTTP_REFERER=expected_referrer,
         )
+        # ensure test passes even after request has been read
+        # context: https://github.com/googleapis/python-logging/issues/159
+        django_request.read()
 
         middleware = request.RequestMiddleware(None)
         middleware.process_request(django_request)