Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
61 lines (42 loc) · 3.21 KB

CONTRIBUTING.md

File metadata and controls

61 lines (42 loc) · 3.21 KB
Raw

Contributing to AIStudio Server

Thanks for your interest in contributing. This document covers how to get started, what areas are open for extension, and what to verify before opening a pull request.

Getting Started

  1. Fork the repo and create a branch from master.
  2. Set up your local environment — see the README for prerequisites (PostgreSQL, RabbitMQ, Python deps).
  3. Run the test suite to confirm everything passes before making changes.
  4. Open a pull request against master with a clear description of what and why.

For significant changes (new workload types, schema changes, new API endpoints), open an issue first to discuss the approach before writing code.

Areas Open for Extension

Area Examples
New workload types Add support for training jobs, embedding benchmarks, diffusion model inference
GPU vendor support AMD ROCm nodes, Intel Gaudi — the manifest builder and SSH executor are the right entry points
Model catalog Add new HuggingFace models to catalog.json with correct license and gating metadata
Benchmark metrics Extend benchmark.py cost engine or add new metric types to the result schema
Dashboard New charts or filters in the FastAPI dashboard (dashboard/backend/)
Storage backends S3/GCS dataset and result storage alongside the existing NFS support
CI/CD Additional GitHub Actions workflows for linting, security scanning, or test coverage

Best Practices

  • One concern per PR. Keep changes focused — a new workload type should not also refactor the SSH executor.
  • Tests required. New API endpoints, services, and state machine transitions need tests in AIStudio/tests/.
  • No secrets in code. Tokens, passwords, and keys belong in ~/.aistudio/env on GPU nodes or in Helm values secrets — never hardcoded or logged.
  • License compliance. Any new dependency must have a permissive licence (Apache-2.0, MIT, BSD). Update THIRD-PARTY-NOTICES.md if you add a package with notice obligations. Do not add packages with GPL, AGPL, or contested provenance without prior discussion.
  • Workload isolation. Workload containers must remain self-contained — no direct database access, all communication via RabbitMQ events and the REST API.

Before Raising a PR

Run these checks locally and confirm they all pass:

# 1. Full test suite (requires PostgreSQL — see README for setup)
cd AIStudio
pytest tests/ -v --log-cli-level=INFO

# 2. License compliance check
make check-licenses

# 3. Model access check (verifies catalog.json entries are reachable)
python3 scripts/check_model_access.py

# 4. Node environment check (if you changed manifest_builder or SSH execution)
make check-node-env NODE=<your-gpu-node>

CI runs the test suite and compliance checks automatically on every PR. PRs with failing tests will not be merged.

Reporting Bugs

Open a GitHub issue with: steps to reproduce, expected vs actual behaviour, relevant log output, and your deployment environment (Kubernetes version, GPU type, Python version).

Security Issues

Do not open a public issue for security vulnerabilities. Email muraharirao@corespan.ai directly.