hackerrank-fetcher

Small helper that fetches public HackerRank profile data. Built to plug into Student-Analytics (which today only stores HackerRank_Username + HackerRank_URL and shows "coming soon" in the profile tab). Lazy by design: fetch one user when their profile opens, never a big batch at once. No login, no secrets.

Unofficial endpoints used (proven, from tashifkhan/hackerrank-stats-api):

Kind	URL
PROFILE	/rest/contests/master/hackers/{u}/profile → model
SCORES	/rest/hackers/{u}/scores_elo → list
BADGES	/rest/hackers/{u}/badges → models[]
CONTESTS	/rest/hackers/{u}/contest_participation?offset=0&limit=50
RATINGS	/rest/hackers/{u}/rating_histories_elo
SUBMISSIONS	/rest/hackers/{u}/submission_histories
RECENT	/rest/hackers/{u}/recent_challenges?limit=100&response_version=v2 (page by page, max 25 pages)

Browser headers, 20s timeout, answers remembered for 1 hour.

Files

src/hackerrank_client/client.py   # talks to HackerRank (7 fetches)
src/hackerrank_client/service.py  # fetches everything at once, cleans it up
src/hackerrank_client/schemas.py  # our clean data format
src/hackerrank_client/cache.py    # remembers answers for 1 hour (no Redis)
api.py                            # website: /{username}/profile | /badges | /contests | /heatmap
tests/                            # tests with fake answers (no internet needed)

Clean format

Only these keys are ever returned — HackerRank's messy keys (badge_type, track_id, hacker_rank, ch_slug, …) never leak out:

{
  "username": "shashank21j",
  "display_name": "Shashank Sharma",
  "badges": [{"track": "Problem Solving", "stars": 6, "solved": 202}],
  "practice_score": 8803,
  "total_solved": 202,
  "contests": [{"name": "Weekly Challenges - Week 4", "slug": "w4"}]
}

practice_score adds up points from every topic. total_solved uses the Problem Solving badge (else adds everything up).

Run

pip install -r requirements.txt
python -m uvicorn api:app --reload --port 8000

Try it

curl.exe http://localhost:8000/health
curl.exe http://localhost:8000/shashank21j/profile
curl.exe http://localhost:8000/shashank21j/badges
curl.exe http://localhost:8000/shashank21j/contests   # new, thin data
curl.exe http://localhost:8000/shashank21j/heatmap    # new, thin data

Errors are simple JSON: unknown user → 404 {"error":"user_not_found"}, HackerRank broken → 502 {"error":"upstream_error"}. Badges/activity/recent give empty defaults for quiet accounts instead of errors.

Use it in Python

import asyncio
from hackerrank_client import HackerRankAPI, get_full_profile

async def main():
    api = HackerRankAPI()
    try:
        print((await get_full_profile("shashank21j", api)).to_dict())
    finally:
        await api.close()

asyncio.run(main())

Tests

python -m pytest tests/ -q

Fake HackerRank answers — no internet. Covers unknown users, empty history, broken answers, and the 25-page guard on recent challenges.

Limits (read before trusting the numbers)

• Unofficial pages — HackerRank can change or block them anytime; we show empty sections / upstream_error instead of retrying forever.
• No worldwide rank — HackerRank only gives per-topic ranks, so there is no single global rank number to show.
• Empty history is normal — new/quiet accounts return no activity and no badges; that shows as "no data", not an error.
• Totals are estimates — points/solved are added up by us, not official totals.
• /contests + /heatmap are new — upstream data is thin (name/slug, date→count). /profile + /badges are the solid ones.
• No Redis/Docker — memory cache only (1 hour).