Merge pull request #94 from Xpirix/handle_invalid_url

Add timeout handling for URL validation

Author: Xpirix

qgis-app/plugins/tests/test_validator.py

@@ -18,16 +18,19 @@ def setUp(self):
         )
         web_not_exist_plugins = os.path.join(TESTFILE_DIR, "web_not_exist.zip")
         valid_plugins = os.path.join(TESTFILE_DIR, "valid_metadata_link.zip")
+        timeout_plugin_link = os.path.join(TESTFILE_DIR, "timeout_plugin_link.zip_")
         self.valid_metadata_link = open(valid_plugins, "rb")
         self.invalid_metadata_link = open(invalid_plugins, "rb")
         self.web_not_exist = open(web_not_exist_plugins, "rb")
         self.invalid_url_scheme = open(invalid_url_scheme_plugins, "rb")
+        self.timeout_plugin_link = open(timeout_plugin_link, "rb")
 
     def tearDown(self):
         self.valid_metadata_link.close()
         self.invalid_metadata_link.close()
         self.invalid_url_scheme.close()
         self.web_not_exist.close()
+        self.timeout_plugin_link.close()
 
     def test_valid_metadata(self):
         self.assertTrue(
@@ -112,6 +115,28 @@ def test_invalid_metadata_web_does_not_exist(self):
             ),
         )
 
+    def test_timeout_plugin_link(self):
+        """
+        The timeout_plugin_link.zip contains metadata file with
+        invalid scheme.
+        bug tracker : http://www.example.com
+        repo :  http://www.example.com/
+        homepage: http://www.google.com:81/
+        """
+        self.assertRaises(
+            ValidationError,
+            validator,
+            InMemoryUploadedFile(
+                self.timeout_plugin_link,
+                field_name="tempfile",
+                name="testfile.zip",
+                content_type="application/zip",
+                size=39889,
+                charset="utf8",
+            ),
+        )
+
+
     @mock.patch("requests.get", side_effect=requests.exceptions.SSLError())
     def test_check_url_link_ssl_error(self, mock_request):
         urls = [{'url': "http://example.com/", 'forbidden_url': "forbidden_url", 'metadata_attr': "metadata attribute"}]

qgis-app/plugins/tests/testfiles/timeout_plugin_link.zip_

@@ -121,9 +121,9 @@ def error_check_if_exist(url: str)->bool:
             # add the headers parameter to make the request appears like coming
             # from browser, otherwise some websites will return 403
             headers = {
-                "User-Agent": "Mozilla/5.0 (Windows NT 6.1; WOW64) "
+                "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
                 "AppleWebKit/537.36 (KHTML, like Gecko) "
-                "Chrome/56.0.2924.76 Safari/537.36"
+                "Chrome/117.0.0.0 Safari/537.36"
             }
             req = requests.head(url, headers=headers)
         except requests.exceptions.SSLError:
@@ -132,6 +132,31 @@ def error_check_if_exist(url: str)->bool:
             return True
         return req.status_code >= 400
 
+    def error_check_if_timeout(url: str) -> bool:
+        # Check if url exists with a timeout
+        try:
+            # Add headers to make the request appear like it's coming from a browser
+            headers = {
+                "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
+                "AppleWebKit/537.36 (KHTML, like Gecko) "
+                "Chrome/117.0.0.0 Safari/537.36"
+            }
+            req = requests.head(url, headers=headers, timeout=10)  # Set timeout to 10 seconds
+        except requests.exceptions.Timeout:
+            return True
+        except Exception:
+            return True
+        return req.status_code >= 400
+
+    timeout_url_error = [item for item in [url_item['metadata_attr'] for url_item in urls if error_check_if_timeout(url_item['url'])]]
+    if len(timeout_url_error) > 0:
+        timeout_url_error_str = ", ".join(timeout_url_error)
+        raise ValidationError(
+        _(
+            f"Please provide valid url link for the following key(s) in the metadata source: <strong>{timeout_url_error_str}</strong>. "
+            "The website(s) cannot be reached within 10 seconds."
+            )
+    )
     url_error = [item for item in [url_item['metadata_attr'] for url_item in urls if error_check(url_item['url'], url_item['forbidden_url'])]]
     if len(url_error) > 0:
         url_error_str = ", ".join(url_error)