README.md

metadata

title: Hackers Leaderboard
emoji: 🏆
colorFrom: purple
colorTo: pink
sdk: gradio
sdk_version: 5.50.0
app_file: app.py
pinned: false

Hackers Leaderboard

Tracks engagement from the hf-skills organization for the hackathon leaderboard.

How Points Work

Simple and fair - 1 point per activity:

Activity	Points
💬 Open a discussion	1
📝 Post a comment	1
🔀 Open a PR	1
📦 Own/create a repo	1

Scripts

Collect Points

# Collect org activity only
HF_TOKEN=$HF_TOKEN python collect_points.py

# Also scan trending repos for member PRs/discussions
HF_TOKEN=$HF_TOKEN python collect_points.py --scan-external

# Scan only specific repo types
HF_TOKEN=$HF_TOKEN python collect_points.py --scan-external --repo-type models
HF_TOKEN=$HF_TOKEN python collect_points.py --scan-external --repo-type models datasets

# Push to HF dataset
HF_TOKEN=$HF_TOKEN python collect_points.py --scan-external --push-to-hub

# Custom output
python collect_points.py --output my_leaderboard.json --repo-id my-org/my-dataset

Options

Flag	Description
--scan-external	Scan trending repos across Hub for member activity
--repo-type	Filter external scan to: models, datasets, spaces
--push-to-hub	Push results to HF dataset
--repo-id	Target dataset repo (default: hf-skills/hackers-leaderboard)
--output	Local JSON output path

Run the App

HF_TOKEN=$HF_TOKEN python app.py

API

The collector scans:

• All models, datasets, and spaces in the org
• All discussions and PRs on those repos
• All comments on discussions

Results are saved as JSONL for easy dataset consumption.

Output Format

{
  "username": "user123",
  "total_points": 15,
  "discussions_opened": 3,
  "comments_made": 8,
  "prs_opened": 2,
  "repos_owned": 2
}