conda-presto 0.6.0

Resolve conda environment.yml files to fully pinned package lists with SHA256 hashes

conda-presto

A fast, dry-run conda solver exposed as both a CLI and an HTTP API. Given package specs or an environment file (environment.yml, pixi.toml, pyproject.toml, requirements.txt, conda-lock, pixi-lock, …), it resolves fully pinned packages for one or more platforms — without downloading or installing anything — and emits the result as native JSON or any conda exporter format (pixi.lock, conda-lock.yml, environment YAML, explicit file, …).

Highlights

• Resolve inline specs or any environment file format
• Full package metadata: sha256, md5, urls, sizes, depends
• Cross-platform solving with automatic virtual package injection
• Multi-platform parallel solves via ProcessPoolExecutor
• Output as JSON or any conda exporter format (--format / ?format=)
• Lockfile-to-lockfile transcode path that skips solving when possible
• Review proposed environments with /preflight, /diff, and /explain
• Content-addressed HTTP result cache with /r/<sha256> lookups and optional file or Redis backing
• HTTP API with interactive docs (Scalar UI), compression, rate limiting
• GitHub Action for CI pipelines (local CLI and hosted API modes)
• Docker images for server and CLI deployment
• Uses conda-rattler-solver for fast SAT solving

Quick start

pixi global install --git https://github.com/jezdez/conda-presto.git
conda presto -c conda-forge -p linux-64 python=3.12 numpy

Run a server

docker run --rm -p 8000:8000 ghcr.io/jezdez/conda-presto:latest
curl http://localhost:8000/health

Pin a versioned image tag for deployments. See the container configuration reference for server tuning and Redis result-cache setup.

Documentation

Full documentation is available at the conda-presto docs site:

• Quick start — install and first resolve
• CLI tutorial — in-depth CLI usage
• HTTP API tutorial — HTTP workflows
• CI pipeline — GitHub Action setup
• Reference — CLI flags, endpoints, formats, env vars
• Architecture — how it works
• Roadmap — shipped foundations and linked future work

Development

git clone https://github.com/jezdez/conda-presto.git
cd conda-presto
pixi install
pixi run lint        # ruff check
pixi run format      # ruff format
pixi run test        # pytest
pixi run bench       # pytest-benchmark
pixi run serve       # uvicorn with --reload
pixi run -e docs docs  # build documentation

License

BSD-3-Clause