Update Github Profile Readme Dynamically using a Github Workflow

TL;DR: Update your Github Profile Readme dynamically with recent blog posts. There’s an example (profile and code).

Github Profile Readme

Github introduced the option to a add profile readme, which is a special repository that is used to display a readme on your profile page. I wanted to post links to my recent blog posts there, so I wrote a simple python script to check for the latest X posts, and update the readme with links to them. Then I used a scheduled workflow to run the script every day so my new posts are always displayed.

The Python Script

The python script is pretty simple. It takes in the url of the xml feed for the blog, so on mine for example, it is here. Then it parses out the title and url of the latest X posts and replases a placeholder in the readme with the links. You can see the full script in the repo here, but here’s the main part:

def main():
    blog_feed_url = os.getenv('BLOG_FEED_URL', None)
    number_of_posts = os.getenv('NUMBER_OF_POSTS', 10)
    readme_name = os.getenv('README_TEMPLATE_NAME', 'README_template.md')
    if not blog_feed_url:
        print('Please set the BLOG_FEED_URL environment variable')
        return
    print('Scraping blog feed from: {}'.format(blog_feed_url))
    print('Number of posts: {}'.format(number_of_posts))
    url_title_map = xml_to_url(
        get_blog_feed_xml(blog_feed_url), number_of_posts)
    add_to_readme(url_title_map, readme_name)

The BLOG_FEED_URL needs to be set in the Github repository settings, I set it as a variable (under ‘reponame/settings/variables/actions’ on github). The other variables (NUMBER_OF_POSTS and README_TEMPLATE_NAME) are optional, but can be set in the same way. The README_TEMPLATE_NAME is the name of the readme that you want to update, it should contain the placeholder text BLOG_POST_LIST to replace with the list of blog posts.

The Github Workflow

Now that we have the script to update the readme, we need to run it on a regular schedule, and then commit the updated reamdme to the repo automatically. Here’s the Workflow that I used to do that:

name: "Get latest blogs"
on:
  schedule:
  - cron: "0 0 * * *"
jobs:
  scrape_blog:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
      - name: Set up python 3.10
        uses: actions/setup-python@v4
        with:
          python-version: "3.10"
          cache: "pip"
      - name: Install deps
        run: pip install -r requirements.txt
      - name: Run blog scraper
        env:
          BLOG_FEED_URL: ${{ vars.BLOG_FEED_URL }}
        run: python scrape_blogs.py
      - name: Update commit author
        run: |
          git config user.name 'github-actions[bot]'
          git config user.email 'github-actions[bot]@users.noreply.github.com'          
      - name: Add and commit readme
        run: |
          git add README.md
          git commit -m "Automatic readme update (blog scraper)" || exit 0
          git push origin main          

We can see the Run blog scraper step passes in the environment variables from the repo variables. That step runs the python script above, which gets the latest blogs and copies them into the readme. Then we just need to commit the result and push it to the repo. The || exit 0 on the commit step is so that if there is nothing to commit (no new blog posts), the workflow doesn’t fail.

That’s all

That’s all there is to it. Now you can have a dynamic readme on your Github Profile, you may need to make some changes to the script to fit your specific blog url, but this is the idea.