[dy] Fix git integration bugs (mage-ai#2199)

* [dy] Fix bugs

* [dy] Fix lint

docs/developing-in-the-cloud/setting-up-git.mdx

@@ -15,11 +15,9 @@ sidebarTitle: "Git"
 
 ## 1. Open terminal
 
-1. Open a pipeline
-2. In the right panel of the screen (aka Sidekick), click on the tab named
-   **Terminal**.
+1. In the left navigation menu, click the Terminal icon.
 3. Click the terminal (large dark background area), and you should see a
-   blinking cursor. You can start typing once that cursor appears.
+   yellow cursor. You can start typing once that cursor appears.
 
 ---
 

mage_ai/api/resources/GitBranchResource.py

@@ -1,3 +1,4 @@
+from mage_ai.api.errors import ApiError
 from mage_ai.api.resources.GenericResource import GenericResource
 from mage_ai.data_preparation.git import Git
 from mage_ai.data_preparation.preferences import get_preferences
@@ -37,6 +38,12 @@ async def update(self, payload, **kwargs):
             self.model = dict(name=git_manager.current_branch, status=status)
         elif action_type == 'commit':
             message = payload.get('message')
+            if not message:
+                error = ApiError.RESOURCE_ERROR
+                error.update({
+                    'message': 'Message is empty, please add a message for your commit.',
+                })
+                raise ApiError(error)
             git_manager.commit(message)
         elif action_type == 'push':
             await git_manager.check_connection()

mage_ai/api/resources/SyncResource.py

@@ -1,7 +1,7 @@
 from mage_ai.api.resources.GenericResource import GenericResource
-from mage_ai.data_preparation.git import Git
 from mage_ai.data_preparation.preferences import get_preferences
 from mage_ai.data_preparation.sync import GitConfig
+from mage_ai.data_preparation.sync.git_sync import GitSync
 
 
 class SyncResource(GenericResource):
@@ -22,7 +22,7 @@ def create(self, payload, user, **kwargs):
             dict(sync_config=sync_config.to_dict())
         )
 
-        Git(sync_config)
+        GitSync(sync_config)
 
         return self(get_preferences().sync_config, user, **kwargs)
 
@@ -31,11 +31,10 @@ def member(self, pk, user, **kwargs):
         return self(get_preferences().sync_config, user, **kwargs)
 
     def update(self, payload, **kwargs):
-        git_manager = Git.get_manager()
-        config = git_manager.git_config
+        config = GitConfig.load(config=self.model)
+        sync = GitSync(config)
         action_type = payload.get('action_type')
         if action_type == 'sync_data':
-            git_manager.check_connection()
-            git_manager.reset(config.branch)
+            sync.sync_data()
 
         return self

mage_ai/data_preparation/git/__init__.py

@@ -4,6 +4,8 @@
 import git
 import subprocess
 
+REMOTE_NAME = 'mage-repo'
+
 
 class Git:
     def __init__(self, git_config: GitConfig):
@@ -18,12 +20,14 @@ def __init__(self, git_config: GitConfig):
             self.commit('Initial commit')
 
         try:
-            self.repo.create_remote('origin', self.remote_repo_link)
+            self.repo.create_remote(REMOTE_NAME, self.remote_repo_link)
         except git.exc.GitCommandError:
             # if the remote already exists
-            self.repo.remotes.origin.set_url(self.remote_repo_link)
+            pass
 
-        self.origin = self.repo.remotes.origin
+        self.origin = self.repo.remotes[REMOTE_NAME]
+        if self.remote_repo_link not in self.origin.urls:
+            self.origin.set_url(self.remote_repo_link)
 
     @classmethod
     def get_manager(self):
@@ -56,7 +60,9 @@ async def check_connection(self):
         if return_code is None:
             proc.kill()
             raise Exception(
-                "Connecting to remote timed out, make sure your SSH key is set up properly.")
+                "Connecting to remote timed out, make sure your SSH key is set up properly"
+                " and your repository host is added as a known host. More information here:"
+                " https://docs.mage.ai/developing-in-the-cloud/setting-up-git#5-add-github-com-to-known-hosts")  # noqa: E501
 
     def all_branches(self):
         return [head.name for head in self.repo.heads]
@@ -65,10 +71,14 @@ def reset(self, branch: str = None):
         self.origin.fetch()
         if branch is None:
             branch = self.current_branch
-        self.repo.git.reset('--hard', f'origin/{branch}')
+        self.repo.git.reset('--hard', f'{self.origin.name}/{branch}')
 
     def push(self):
-        self.repo.git.push('--set-upstream', self.origin.name, self.current_branch)
+        self.repo.git.push(
+            '--set-upstream',
+            self.origin.name,
+            self.current_branch
+        )
 
     def pull(self):
         self.origin.pull(self.current_branch)

mage_ai/data_preparation/sync/git_sync.py

@@ -1,37 +1,20 @@
+from mage_ai.data_preparation.git import Git
 from mage_ai.data_preparation.sync import GitConfig
 from mage_ai.data_preparation.sync.base_sync import BaseSync
 from mage_ai.shared.logger import VerboseFunctionExec
-import git
-import subprocess
+import asyncio
 
 
 class GitSync(BaseSync):
     def __init__(self, sync_config: GitConfig):
-        self.remote_repo_link = sync_config.remote_repo_link
-        self.repo_path = sync_config.repo_path
         self.branch = sync_config.branch
-        try:
-            self.repo = git.Repo(self.repo_path)
-        except git.exc.InvalidGitRepositoryError:
-            self.repo = git.Repo.init(self.repo_path)
-
-        try:
-            self.repo.create_remote('origin', self.remote_repo_link)
-        except git.exc.GitCommandError:
-            # if the remote already exists
-            self.repo.remotes.origin.set_url(self.remote_repo_link)
-
-        self.origin = self.repo.remotes.origin
+        self.remote_repo_link = sync_config.remote_repo_link
+        self.git_manager = Git(sync_config)
 
     def sync_data(self):
         with VerboseFunctionExec(
             f'Syncing data with remote repo {self.remote_repo_link}',
             verbose=True,
         ):
-            # use subprocess because the gitpython command wasn't timing out correctly
-            subprocess.run(
-                ['git', 'fetch', self.origin.name],
-                cwd=self.repo_path,
-                timeout=30,
-            )
-            self.repo.git.reset('--hard', f'{self.origin.name}/{self.branch}')
+            asyncio.run(self.git_manager.check_connection())
+            self.git_manager.reset(self.branch)