Fix GitDagBundle to support https (include 46073/46179) (apache#46226)

astronomer · Jan 30, 2025 · 0637366 · 0637366
1 parent b4b6de3
commit 0637366
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 4 deletions.
diff --git a/airflow/config_templates/config.yml b/airflow/config_templates/config.yml
@@ -2615,7 +2615,6 @@ dag_processor:
                 "classpath": "airflow.dag_processing.bundles.git.GitDagBundle",
                 "kwargs": {
                     "subdir": "dags",
-                    "repo_url": "[email protected]:example.com/my-dags.git",
                     "tracking_ref": "main",
                     "refresh_interval": 0
             }

diff --git a/airflow/dag_processing/bundles/git.py b/airflow/dag_processing/bundles/git.py
@@ -140,9 +140,11 @@ def initialize(self) -> None:
             raise AirflowException(f"Connection {self.git_conn_id} doesn't have a host url")
         if isinstance(self.repo_url, os.PathLike):
             self._initialize()
-        elif not self.repo_url.startswith("git@") or not self.repo_url.endswith(".git"):
+        elif not (
+            self.repo_url.startswith("git@") or self.repo_url.startswith("https")
+        ) or not self.repo_url.endswith(".git"):
             raise AirflowException(
-                f"Invalid git URL: {self.repo_url}. URL must start with git@ and end with .git"
+                f"Invalid git URL: {self.repo_url}. URL must start with git@ or https and end with .git"
             )
         else:
             self._initialize()
@@ -236,6 +238,8 @@ def view_url(self, version: str | None = None) -> str | None:
             return None
         if url.startswith("git@"):
             url = self._convert_git_ssh_url_to_https(url)
+        if url.endswith(".git"):
+            url = url[:-4]
         parsed_url = urlparse(url)
         host = parsed_url.hostname
         if not host:

diff --git a/tests/dag_processing/test_dag_bundles.py b/tests/dag_processing/test_dag_bundles.py
@@ -399,7 +399,8 @@ def test_repo_url_validation_for_ssh(self, mock_hook, repo_url, session):
             tracking_ref=GIT_DEFAULT_BRANCH,
         )
         with pytest.raises(
-            AirflowException, match=f"Invalid git URL: {repo_url}. URL must start with git@ and end with .git"
+            AirflowException,
+            match=f"Invalid git URL: {repo_url}. URL must start with git@ or https and end with .git",
         ):
             bundle.initialize()
 
@@ -413,6 +414,10 @@ def test_repo_url_validation_for_ssh(self, mock_hook, repo_url, session):
                 "[email protected]:apache/airflow.git",
                 "https://myorg.github.com/apache/airflow/tree/0f0f0f",
             ),
+            (
+                "https://myorg.github.com/apache/airflow.git",
+                "https://myorg.github.com/apache/airflow/tree/0f0f0f",
+            ),
         ],
     )
     @mock.patch("airflow.dag_processing.bundles.git.Repo")