clean outputs before commit option

Author: Meriem-BenIsmail

jupyterlab_git/git.py

@@ -1182,6 +1182,56 @@ async def merge(self, branch: str, path: str) -> dict:
             return {"code": code, "command": " ".join(cmd), "message": error}
         return {"code": code, "message": output.strip()}
 
+    async def check_notebooks_with_outputs(self, path):
+        import nbformat, os
+
+        code, stdout, _ = await self.__execute(
+            ["git", "diff", "--cached", "--name-only", "--diff-filter=ACM"], cwd=path
+        )
+        staged_files = stdout.splitlines()
+        notebooks = [f for f in staged_files if f.endswith(".ipynb")]
+
+        dirty_notebooks = []
+
+        for nb_path in notebooks:
+            full_nb_path = os.path.join(path, nb_path)
+            try:
+                nb = nbformat.read(full_nb_path, as_version=nbformat.NO_CONVERT)
+                for cell in nb.get("cells", []):
+                    if cell.get("cell_type") == "code" and cell.get("outputs"):
+                        dirty_notebooks.append(nb_path)
+                        break
+            except Exception:
+                pass
+
+        return {
+            "notebooks_with_outputs": dirty_notebooks,
+            "has_outputs": len(dirty_notebooks) > 0,
+        }
+
+    async def strip_notebook_outputs(self, notebooks: list, repo_path: str):
+        for nb_path in notebooks:
+            full_path = os.path.join(repo_path, nb_path)
+
+            try:
+                # Clear outputs using nbconvert
+                subprocess.run(
+                    [
+                        "jupyter",
+                        "nbconvert",
+                        "--ClearOutputPreprocessor.enabled=True",
+                        "--inplace",
+                        full_path,
+                    ],
+                    check=True,
+                )
+
+                # Re-stage the cleaned notebook
+                subprocess.run(["git", "-C", repo_path, "add", full_path], check=True)
+
+            except Exception as e:
+                print(f"Failed: {nb_path}: {e}")
+
     async def commit(self, commit_msg, amend, path, author=None):
         """
         Execute git commit <filename> command & return the result.

jupyterlab_git/handlers.py

@@ -599,6 +599,44 @@ async def post(self, path: str = ""):
         self.finish(json.dumps(body))
 
 
+class GitStripNotebooksHandler(GitHandler):
+    """Handler to strip outputs from notebooks in a repository."""
+
+    @tornado.web.authenticated
+    async def post(self, path: str = ""):
+        """
+        POST request handler to strip outputs from notebooks.
+        """
+        data = self.get_json_body()
+        notebooks = data.get("notebooks", [])
+
+        try:
+            await self.git.strip_notebook_outputs(notebooks, self.url2localpath(path))
+            self.set_status(200)
+            self.finish(json.dumps({"code": 0, "message": "Notebooks stripped"}))
+        except Exception as e:
+            self.set_status(500)
+            self.finish(
+                json.dumps(
+                    {
+                        "code": 1,
+                        "message": f"Failed to strip notebook outputs: {str(e)}",
+                    }
+                )
+            )
+
+
+class GitCheckNotebooksHandler(GitHandler):
+    """
+    Handler to check staged notebooks for outputs.
+    """
+
+    @tornado.web.authenticated
+    async def get(self, path: str = ""):
+        body = await self.git.check_notebooks_with_outputs(self.url2localpath(path))
+        self.finish(json.dumps(body))
+
+
 class GitUpstreamHandler(GitHandler):
     @tornado.web.authenticated
     async def post(self, path: str = ""):
@@ -1182,6 +1220,8 @@ def setup_handlers(web_app):
         ("/stash_pop", GitStashPopHandler),
         ("/stash_apply", GitStashApplyHandler),
         ("/submodules", GitSubmodulesHandler),
+        ("/check_notebooks", GitCheckNotebooksHandler),
+        ("/strip_notebooks", GitStripNotebooksHandler),
     ]
 
     handlers = [

src/components/GitPanel.tsx

@@ -803,9 +803,47 @@ export class GitPanel extends React.Component<IGitPanelProps, IGitPanelState> {
   ): Promise<void> => {
     const errorMsg = this.props.trans.__('Failed to commit changes.');
     let id: string | null = notificationId ?? null;
+
     try {
       const author = await this._hasIdentity(this.props.model.pathRepository);
 
+      const notebooksWithOutputs =
+        await this.props.model.checkNotebooksForOutputs();
+
+      if (notebooksWithOutputs.length > 0) {
+        const dialog = new Dialog({
+          title: this.props.trans.__('Notebook outputs detected'),
+          body: `You are about to commit ${notebooksWithOutputs.length} notebook(s) with outputs. 
+            Would you like to clean them before committing?`,
+          buttons: [
+            Dialog.cancelButton({
+              label: this.props.trans.__('Commit Anyway')
+            }),
+            Dialog.okButton({
+              label: this.props.trans.__('Clean & Commit')
+            })
+          ]
+        });
+
+        const result = await dialog.launch();
+        dialog.dispose();
+
+        if (result.button.label === this.props.trans.__('Commit Anyway')) {
+        } else if (
+          result.button.label === this.props.trans.__('Clean & Commit')
+        ) {
+          id = Notification.emit(
+            this.props.trans.__('Cleaning notebook outputs…'),
+            'in-progress',
+            { autoClose: false }
+          );
+
+          await this.props.model.stripNotebooksOutputs(notebooksWithOutputs);
+        } else {
+          return;
+        }
+      }
+
       const notificationMsg = this.props.trans.__('Committing changes...');
       if (id !== null) {
         Notification.update({

src/model.ts

@@ -733,6 +733,47 @@ export class GitExtension implements IGitExtension {
     await this.refresh();
   }
 
+  /**
+   * Check staged notebooks for outputs.
+   *
+   * @returns A promise resolving to an array of notebook paths that have outputs
+   *
+   * @throws {Git.NotInRepository} If the current path is not a Git repository
+   * @throws {Git.GitResponseError} If the server response is not ok
+   * @throws {ServerConnection.NetworkError} If the request cannot be made
+   */
+  async checkNotebooksForOutputs(): Promise<string[]> {
+    const path = await this._getPathRepository();
+
+    return this._taskHandler.execute('git:check-notebooks', async () => {
+      const result = await requestAPI<{ notebooks_with_outputs: string[] }>(
+        URLExt.join(path, 'check_notebooks')
+      );
+      return result.notebooks_with_outputs;
+    });
+  }
+
+  /**
+   * Strip outputs from the given staged notebooks.
+   *
+   * @param notebooks - Array of notebook paths to clean
+   *
+   * @returns A promise resolving when the operation completes
+   *
+   * @throws {Git.NotInRepository} If the current path is not a Git repository
+   * @throws {Git.GitResponseError} If the server response is not ok
+   * @throws {ServerConnection.NetworkError} If the request cannot be made
+   */
+  async stripNotebooksOutputs(notebooks: string[]): Promise<void> {
+    const path = await this._getPathRepository();
+
+    await this._taskHandler.execute('git:strip-notebooks', async () => {
+      await requestAPI(URLExt.join(path, 'strip_notebooks'), 'POST', {
+        notebooks
+      });
+    });
+  }
+
   /**
    * Get (or set) Git configuration options.
    *