菜单
Python package and environment management using uv and mamba. Use when installing packages, creating virtual environments, setting up new projects, or managing dependencies. NOT for general Python coding questions.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
MIT Engaging cluster reference. SLURM job scheduling, Coley group nodes, filesystem layout, Python environments, and common commands. Use when working with the Engaging HPC cluster.
Verifies that bibliography references exist and have accurate metadata. Use when the user asks to "check references", "verify a bibliography", "validate citations", "check for hallucinated citations", "check a .bib file", or mentions auditing a BibTeX file for accuracy. Looks up each entry against Crossref and arXiv, then spawns web-search subagents for non-indexed sources, flagging missing DOIs, wrong years, title mismatches, venue mismatches, and broken URLs.
Reference guide for the MolGPU shared GPU cluster. Machine specs, SSH setup, shared filesystem, conda/mamba, multi-GPU training tips, and common commands.
VASP on MIT Engaging. VASP 6.4.2 (GCC/OpenMPI, Rocky 8 compatible), module loading, SLURM submission, POTCAR paths. Use when setting up or running VASP calculations on Engaging.
MassSpecGym benchmark reference — dataset schema, Python API, transforms, evaluation metrics, retrieval pipeline, and common patterns. Use when working with MassSpecGym data, models, or evaluation.
Machine learning workflows for molecular data. Use when building ML models with molecules, splitting datasets, selecting fingerprints for ML, avoiding data leakage, or doing scaffold-based train/test splits. Covers common pitfalls in molecular ML.
| name | pkg-mgmt |
| description | Python package and environment management using uv and mamba. Use when installing packages, creating virtual environments, setting up new projects, or managing dependencies. NOT for general Python coding questions. |
Default to uv for speed/ML; use mamba for heavy C++/Fortran binaries (e.g., OpenMM).
Best for: New projects, PyTorch/JAX, RDKit, CI/CD.
uv is an extremely fast Python package installer and resolver written in Rust.
# macOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh
# Or with Homebrew
brew install uv
# Initialize project
uv init my-project && cd my-project
# (Optional) Pin Python version — uv downloads it automatically if missing
uv python pin 3.11
# Add dependencies (updates pyproject.toml & uv.lock)
uv add rdkit pandas torch
uv add --dev pytest ruff
# Sync environment (guarantees reproducibility)
uv sync
# Run commands in the environment
uv run python train_model.py
uv run pytest
Run a script with dependencies ephemerally (no permanent env created):
uv run --with rdkit --with matplotlib molecular_vis.py
# Create a virtual environment
uv venv
uv venv --python 3.11 # specific version
# Activate
source .venv/bin/activate
# Install packages
uv pip install rdkit scikit-learn pandas
uv pip install -r requirements.txt
uv pip install -e .
Best for: OpenMM, AmberTools, legacy projects, or strict system library requirements.
mamba is a fast, drop-in replacement for conda.
# Install miniforge (includes mamba)
# macOS ARM
curl -L -O "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-MacOSX-arm64.sh"
bash Miniforge3-MacOSX-arm64.sh
# macOS Intel
curl -L -O "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-MacOSX-x86_64.sh"
bash Miniforge3-MacOSX-x86_64.sh
# Linux
curl -L -O "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Linux-x86_64.sh"
bash Miniforge3-Linux-x86_64.sh
Always use environment.yml with conda-forge:
# environment.yml
name: md-sim
channels:
- conda-forge
dependencies:
- python=3.11
- openmm
- ambertools
- rdkit
- numpy
- pandas
- pip # Allow pip for pure python packages if needed
# Create from file
mamba env create -f environment.yml
# Update (use --prune to remove deleted deps)
mamba env update -f environment.yml --prune
# Export environment
mamba env export --no-builds > environment.yml
# Create environment
mamba create -n myenv python=3.11
# Activate/deactivate
mamba activate myenv
mamba deactivate
# Install packages
mamba install rdkit numpy pandas
| Scenario | Tool | Reasoning |
|---|---|---|
| General ML / PyTorch | uv | 100x faster, handles wheels perfectly |
| Cheminformatics (RDKit) | uv | RDKit PyPI wheels are now stable |
| MD Sims (OpenMM/Amber) | mamba | Complex CUDA/C++ bindings are fragile on PyPI |
| Publishing/Sharing | uv | pyproject.toml is the modern standard (PEP 621) |
| Quick Scripts | uv | uv run --with enables single-file reproducibility |
| Legacy Projects | mamba | If it already uses Conda, stick with it |
Need mamba binaries (e.g., OpenMM) but want uv speed for everything else? Create the env with mamba, then use uv pip inside it:
mamba create -n hybrid-env openmm python=3.11 -c conda-forge
mamba activate hybrid-env
uv pip install torch rdkit scikit-learn # Installs into the active Conda env
# uv does this automatically (uv.lock)
uv lock
# For mamba, export without build strings for portability
mamba env export --no-builds > environment.yml
Stop using requirements.txt. Define deps in pyproject.toml:
[project]
name = "my-project"
version = "0.1.0"
dependencies = [
"rdkit",
"numpy>=1.24",
"pandas>=2.0",
]
[project.optional-dependencies]
dev = ["pytest", "ruff"]
Avoid ABI conflicts by strictly prioritizing conda-forge:
conda config --add channels conda-forge
conda config --set channel_priority strict
# Bad
pip install rdkit
# Good
uv venv && source .venv/bin/activate && uv pip install rdkit
# GitHub Actions example
- uses: astral-sh/setup-uv@v4
- run: uv sync
- run: uv run pytest