Bring your own token for commits (#10)

Dockerfile

@@ -1,6 +1,6 @@
-FROM python:3.11-slim as base
+FROM python:3.11-slim AS base
 
-RUN apt-get update && apt-get install -y git
+RUN apt-get update && apt-get install -y git && apt clean
 
 COPY requirements.txt /
 RUN pip install --no-cache-dir -r requirements.txt

README.md

@@ -73,12 +73,14 @@ jobs:
         run: python /app/alttexter-ghclient.py
 ```
 
-You'll need to set the following repo secrets:
+You'll need to [add the following repo secrets](https://docs.github.com/en/codespaces/managing-codespaces-for-your-organization/managing-development-environment-secrets-for-your-repository-or-organization#adding-secrets-for-a-repository):
 
 * `ALTTEXTER_ENDPOINT`: Endpoint URL of the running `alttexter` (eg. `https://alttexter-prod.westeurope.cloudapp.azure.com:9100/alttexter`)
 * `ALTTEXTER_TOKEN`: Single token for service.
 * `ALTTEXTER_RATEMINUTE`: Maximum number of calls to `alttexter` service per minute. Consider [OpenAI rate limits](https://platform.openai.com/docs/guides/rate-limits) for your tier. I'd reccomend no more than 3.
 
+Optionally you can [add repo environment variables](https://docs.github.com/en/actions/learn-github-actions/variables#creating-configuration-variables-for-a-repository) `ALTTEXTER_GITHUB_TOKEN_OVERRIDE` (text of a repo secret name, for example `CR_TOKEN`—if using pass the secret in `alttexter.yml`), `ALTTEXTER_GITHUB_USERNAME` & `ALTTEXTER_GITHUB_EMAIL` to override the default GitHub token, username and email used for commits.
+
 Once that's done you can open a PR with files that need alt text and/or title attributes.
 
 ## Features

alttexter-ghclient.py

@@ -216,7 +216,7 @@ async def get_image_metadata(self, session, markdown_content, file_path, encoded
 
         return response_structure
 
-    def update_image_metadata(self, markdown_content, file_path, image_metadata, base_dir, repo_root, is_ipynb):
+    def update_image_metadata(self, markdown_content, image_metadata, base_dir, repo_root, is_ipynb):
         """
         Updates markdown content with new alt-text and title attributes.
 
@@ -234,7 +234,7 @@ def update_image_metadata(self, markdown_content, file_path, image_metadata, bas
 
         def replacement(match):
             groups = match.groups()
-            image_alt, image_path = groups[:2]
+            _, image_path = groups[:2]
             title_double, title_single = groups[2:4] if len(groups) == 4 else ("", "")
             image_title = title_double or title_single or ""
 
@@ -311,7 +311,7 @@ async def process_file(session, file, alttexter_endpoint, github_handler, metada
 
         image_metadata = response["data"].get('images', [])
         updated_content, images_not_updated = metadata_updater.update_image_metadata(
-            markdown_content, file_path, image_metadata, base_dir, repo_root, is_ipynb
+            markdown_content, image_metadata, base_dir, repo_root, is_ipynb
         )
 
         if updated_content != markdown_content:

ghutils.py

@@ -24,7 +24,19 @@ def __init__(self, repo_name, pr_number):
             repo_name (str): Repository name.
             pr_number (int): Pull request number.
         """
-        self.github_obj = Github(os.getenv('GITHUB_TOKEN'))
+        token_var_name = os.getenv('ALTTEXTER_GITHUB_TOKEN_OVERRIDE')
+        if token_var_name:
+            github_token = os.getenv(token_var_name)
+            if github_token:
+                logging.info(f"Using custom token provided by the environment variable: {token_var_name}")
+            else:
+                logging.error(f"The environment variable {token_var_name} does not exist or is not set. Falling back to GITHUB_TOKEN.")
+                github_token = os.getenv('GITHUB_TOKEN')
+        else:
+            logging.debug("No custom token override provided; using GITHUB_TOKEN.")
+            github_token = os.getenv('GITHUB_TOKEN')
+
+        self.github_obj = Github(github_token)
         self.repo = self.github_obj.get_repo(repo_name)
         self.pr = self.repo.get_pull(pr_number)
 
@@ -72,12 +84,25 @@ def commit_and_push(self, updated_files, commit_message):
         file_paths_str = "[" + ", ".join(updated_files) + "]"
         logging.info(f"{file_paths_str} Initiating commit and push process")
 
+        git_username = os.getenv('ALTTEXTER_GITHUB_USERNAME') or 'github-actions'
+        git_email = os.getenv('ALTTEXTER_GITHUB_EMAIL') or '[email protected]'
+
+        if os.getenv('ALTTEXTER_GITHUB_USERNAME'):
+            logging.info(f"Using custom Git username: {git_username}")
+        else:
+            logging.info("Using default Git username: 'github-actions'")
+
+        if os.getenv('ALTTEXTER_GITHUB_EMAIL'):
+            logging.info(f"Using custom Git email: {git_email}")
+        else:
+            logging.info("Using default Git email: '[email protected]'")
+
         try:
             # Configure Git to allow operations in the current directory
             current_directory = os.getcwd()
             subprocess.run(['git', 'config', '--global', '--add', 'safe.directory', current_directory], check=True)
-            subprocess.run(['git', 'config', 'user.name', 'github-actions'], check=True)
-            subprocess.run(['git', 'config', 'user.email', '[email protected]'], check=True)
+            subprocess.run(['git', 'config', '--global', 'user.name', git_username], check=True)
+            subprocess.run(['git', 'config', '--global', 'user.email', git_email], check=True)
 
             if subprocess.run(['git', 'status', '--porcelain'], capture_output=True, text=True).stdout:
                 subprocess.run(['git', 'add'] + updated_files, check=True)