Automate PR Merging & Improve Contributor Card Management

[shahiakhilesh1304]

This issue suggests two enhancements:

1. Automate PR Merging:

   • Introduce a GitHub Action that automates PR approvals and merging.
   • The action should validate PR changes, ensure compliance with guidelines, and merge automatically if all checks pass.
   • This will reduce manual approval overhead for repository owners.

2. Improve Contributor Card Management:

   • Currently, when contributors add their cards, the previous entry gets overwritten.
   • Instead of modifying contributors.js directly, contributors should add a new constant.
   • Modify the .sh script to ensure contributor data is appended, not replaced.
   • Show all contributor cards on the index page or create a separate card.html file listing all contributors.

3. Possible Solution:

   • Implement a GitHub Action that automates PR validation & merging.
   • Modify contributors.js to prevent overwriting existing entries.
   • Update the .sh script to handle multiple contributors properly.

• Benefits:
  ✔️ Speeds up PR handling and reduces manual intervention.
  ✔️ Ensures every contributor's card remains visible.
  ✔️ Improves repository maintenance and contributor experience.

[Roshanjossey]

Hi @shahiakhilesh1304, thank you for creating this issue. I do like to have a workflow for merge automation.

One additional constrain I'd like for this is to have checks for offensive or NSFW content.

What you wrote under Improve Contributor Card Management: is not correct.
generate cards script doesn't replace anything. It looks for html files inside contributors directory and puts them all in contributors.js. You could try this on your local to verify.

Instead of modifying contributors.js directly, contributors should add a new constant.

I didn't understand this.

[shahiakhilesh1304]

Hi @Roshanjossey,
I can help you automate the PR workflow using the Hugging Face AI API to detect NSFW content. Additionally, we can use regex for validation both methods are effective. Let me know if you’d like assistance with this.

Regarding Improve Contributor Card Management, I wasn't familiar with that part of the code, but it looks good.

What I was trying to suggest is that there should be a dedicated page displaying all contributor cards in one place.

[Roshanjossey]

Hugging Face AI API to detect NSFW content

Could you share a link to this so I can check their usage limits, terms etc.

we can use regex for validation

Not sure about this solution. Can you give me examples?

What I was trying to suggest is that there should be a dedicated page displaying all contributor cards in one place.

Like https://roshanjossey.github.io/code-contributions ?

[xarical]

Could you share a link to this so I can check their usage limits, terms etc.

Hugging Face inference API: https://huggingface.co/docs/inference-providers/index. It's limited to $0.10 of usage a month, so it depends on what model you want to use.

---

Alternatively:

1. Added NSFW Content Detection #108 proposes downloading and running a fine-tuned BERT text-classification model (unitary/toxic-bert) in the GitHub Action, which would be free, albeit a tad slow (2-3 minute runtime, mostly because it has to install transformers and torch). Works pretty well for English text, and the model could be switched out for a multilingual one such as unitary/multilingual-toxic-xlm-roberta for decent multilingual detetection.

2. An LLM such as Gemma 2 or LLama 3.1 through Groq might provide the best multilingual detection of offensive content, speed, and rate limits (30 requests per minute and 14,400 requests per day on the free plan, with token usage being negligible for this use case, though some safeguards may be needed to prevent abuse which might make it less ideal).

I have the beginnings of a PR based on #108 using Groq that I can finish and clean up a bit if you're interested @Roshanjossey.