fix: github adapter only applies to URLs formatted as expected (#1708)

- fixes part of #1607 In the absence of other hints (e.g. a Controller object associated to another plugin), all URLs from `github.com` were being processed by the `GithubPlugin`, to create a Github Data Portal adapter. With this PR, `GithubPlugin` only handles the URL if it has the expected format "https://github.com/user_or_org/repo", and "https://github.com/user_or_org", otherwise it is passed other to be treated as a plain URL. I also made some refactorings along the way, to have more readable and less nested code. As tests github portal are currently skipped, I checked manually that I have not broken this functionality.
frictionlessdata · Nov 22, 2024 · d97ae46 · d97ae46
1 parent 72b74a8
commit d97ae46
Showing 1 changed file with 37 additions and 12 deletions.
diff --git a/frictionless/portals/github/plugin.py b/frictionless/portals/github/plugin.py
@@ -24,21 +24,46 @@ def create_adapter(
         basepath: Optional[str] = None,
         packagify: bool = False,
     ):
-        if isinstance(source, str):
-            parsed = urlparse(source)
-            if not control or isinstance(control, GithubControl):
-                if parsed.netloc == "github.com":
-                    control = control or GithubControl()
-                    splited_url = parsed.path.split("/")[1:]
-                    if len(splited_url) == 1:
-                        control.user = splited_url[0]
-                        return GithubAdapter(control)
-                    if len(splited_url) == 2:
-                        control.user, control.repo = splited_url
-                    return GithubAdapter(control)
+        """Checks if the source is meant to access a Github Data portal, and returns the adapter if applicable
+
+        The source is expected to be in one of the formats :
+
+        - https://github.com/user_or_org
+        - https://github.com/user_or_org/repo
+
+        Alternatively, user and/or repo information can be provided in the
+        GithubControl, with an empty source.
+
+        """
+        if control and not isinstance(control, GithubControl):
+            # Explicit control for other plugin
+            return
+
         if source is None and isinstance(control, GithubControl):
+            # Source informations are inside the control
             return GithubAdapter(control=control)
 
+        if not isinstance(source, str):
+            return
+
+        parsed_url = urlparse(source)
+        splited_url = parsed_url.path.split("/")[1:]
+
+        has_expected_format = (
+            parsed_url.netloc == "github.com"
+            and len(splited_url) >= 1
+            and len(splited_url) <= 2
+        )
+
+        if has_expected_format:
+            control = control or GithubControl()
+            control.user = splited_url[0]
+
+            if len(splited_url) == 2:
+                control.repo = splited_url[1]
+
+            return GithubAdapter(control)
+
     def select_control_class(self, type: Optional[str] = None):
         if type == "github":
             return GithubControl