Inspiration
I chose this problem statement because I feel that it is the most impactful solution as everyone can use it for their own project regardless of industry. GitHub actions published by others have helped me in some of my previous workflows and I have been wanting to look into creating my own one day.
What it does
This action checks markdown (.md) files for images and flags out missing inline alt-text. Optional usage with the Microsoft Azure Cognitive Services integration to get suggestions for the missing alt-text. Can be run as part of a PR pre-merge check or periodically. For both public and private repositories. More detailed information can be found on the GitHub page!
How I built it
The solution uses NodeJS runtime. I used the GitHub REST API for to authenticate and interact with the checked out repository's content and do some preprocessing mainly using string manipulation. For the suggestions portion, I used the Microsoft Azure Cognitive Services (Computer Vision Image Analysis) REST API to return the image recognised text. I then make use of the GitHub REST API again to create a check run that generates a summary report and annotation on line level.
Challenges I ran into
It was tedious to do testing as I had to use a separate test repository to run the actual solutions repository. I also had to think of various configurations that might interest people to use the action in different ways for their project.
Accomplishments that I'm proud of
A full end to end working solution!
What I learned
I once saw this quote that says "I am better than I was yesterday but not as good as I will be tomorrow". And that was how I encouraged myself to continue building the solution one day at a time, one line of code at a time.
What's next for Markdown Image Alt-text Checker
The repository will be maintained on GitHub and bug fixes are definitely in the pipeline. I am also always open to community contributions to the project as any other GitHub project!
Log in or sign up for Devpost to join the conversation.