Updated error message to include server side errors if present. (#183)

neshdev · web-flow · commit c1c8ed5cb1d4 · 2024-12-06T14:27:32.000-05:00
Error message to include server side error messages if present in
payload.

BUG=b/375194356
diff --git a/src/kagglehub/exceptions.py b/src/kagglehub/exceptions.py
@@ -1,9 +1,13 @@
+import logging
 from http import HTTPStatus
 from typing import Any, Optional
 
 import requests
 
 from kagglehub.handle import CompetitionHandle, ResourceHandle
+from kagglehub.logger import EXTRA_CONSOLE_BLOCK
+
+logger = logging.getLogger(__name__)
 
 
 class CredentialError(Exception):
@@ -58,6 +62,13 @@ def kaggle_api_raise_for_status(response: requests.Response, resource_handle: Op
         response.raise_for_status()
     except requests.HTTPError as e:
         message = str(e)
+        server_error_message = ""
+        try:
+            server_error_message = response.json().get("message", "")
+            if server_error_message:
+                server_error_message = f"The server reported the following issues: {server_error_message}\n"
+        except requests.exceptions.JSONDecodeError as ex:
+            logger.info(f"Server payload is not json. See {ex}", extra={**EXTRA_CONSOLE_BLOCK})
         resource_url = resource_handle.to_url() if resource_handle else response.url
         if response.status_code in {HTTPStatus.UNAUTHORIZED, HTTPStatus.FORBIDDEN}:
             if isinstance(resource_handle, CompetitionHandle):
@@ -72,17 +83,19 @@ def kaggle_api_raise_for_status(response: requests.Response, resource_handle: Op
                 message = (
                     f"{response.status_code} Client Error."
                     "\n\n"
-                    f"You don't have permission to access resource at URL: {resource_url}"
-                    "\nPlease make sure you are authenticated if you are trying to access a"
-                    " private resource or a resource requiring consent."
+                    f"You don't have permission to access resource at URL: {resource_url}. "
+                    f"{server_error_message}"
+                    f"Please make sure you are authenticated if you are trying to access a "
+                    f"private resource or a resource requiring consent."
                 )
 
         if response.status_code == HTTPStatus.NOT_FOUND:
             message = (
                 f"{response.status_code} Client Error."
                 "\n\n"
-                f"Resource not found at URL: {resource_url}"
-                "\nPlease make sure you specified the correct resource identifiers."
+                f"Resource not found at URL: {resource_url}\n"
+                f"{server_error_message}"
+                "Please make sure you specified the correct resource identifiers."
             )
 
         # Default handling
diff --git a/tests/server_stubs/kaggle_api_stub.py b/tests/server_stubs/kaggle_api_stub.py
@@ -2,7 +2,7 @@
 import os
 import re
 
-from flask import Flask, Response, jsonify, request
+from flask import Flask, Response, jsonify, make_response, request
 from flask.typing import ResponseReturnValue
 
 from kagglehub.integrity import to_b64_digest
@@ -18,8 +18,26 @@ def head() -> ResponseReturnValue:
 
 @app.errorhandler(404)
 def error(e: Exception):  # noqa: ANN201
-    data = {"code": "404", "error": str(e)}
-    return jsonify(data), 200
+    data = {"code": "404", "error": str(e), "message": "server side error"}
+    return jsonify(data), 404
+
+
+@app.route("/api/v1/content_type_mismatch", methods=["GET"])
+def content_type_and_payload_mismatch() -> ResponseReturnValue:
+    html = """
+    <html>
+    <head>
+        <title>Test</title>
+    </head>
+    <body>
+        <h1>This is HTML content</h1>
+    </body>
+    </html>
+    """
+    resp = make_response(html)
+    resp.headers["Content-Type"] = "application/json"
+    resp.status_code = 404
+    return resp
 
 
 @app.route("/api/v1/no-integrity", methods=["GET"])
diff --git a/tests/test_kaggle_api_client.py b/tests/test_kaggle_api_client.py
@@ -5,7 +5,7 @@
 import kagglehub
 from kagglehub import clients
 from kagglehub.clients import KaggleApiV1Client
-from kagglehub.exceptions import DataCorruptionError
+from kagglehub.exceptions import DataCorruptionError, KaggleApiHTTPError
 from tests.fixtures import BaseTestCase
 
 from .server_stubs import kaggle_api_stub as stub
@@ -68,6 +68,18 @@ def test_download_corrupted_file_fail_integrity_check(self) -> None:
             # Assert the corrupted file has been deleted.
             self.assertFalse(os.path.exists(out_file))
 
+    def test_error_message(self) -> None:
+        api_client = KaggleApiV1Client()
+        with self.assertRaises(KaggleApiHTTPError) as ex:
+            api_client.get("/error")
+        self.assertIn("The server reported the following issues:", str(ex.exception))
+
+    def test_error_message_with_mismatch(self) -> None:
+        api_client = KaggleApiV1Client()
+        with self.assertRaises(KaggleApiHTTPError) as ex:
+            api_client.get("/content_type_mismatch")
+        self.assertNotIn("The server reported the following issues:", str(ex.exception))
+
     @patch.dict("os.environ", {})
     def test_get_user_agent(self) -> None:
         self.assertEqual(clients.get_user_agent(), f"kagglehub/{kagglehub.__version__}")
