analysis_template/pixi.toml at main · quadbio/analysis_template · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
[workspace]
name = "analysis-template"
version = "0.1.0"
description = "Template for single-cell/spatial analysis projects (pixi-managed)"
authors = ["Your Name <[email protected]>"]
channels = ["conda-forge"]
platforms = ["osx-arm64", "linux-64"]

[tasks]
lab = "jupyter lab"
test = "pytest"
rename = { cmd = "python scripts/rename_package.py", description = "Rename the placeholder package/project (run once on a fresh clone; see scripts/rename_package.py)" }
install-kernel = "python -m ipykernel install --user --name=analysis-template --display-name='Analysis Template (Pixi)'"
install-hooks = { cmd = "pre-commit install --install-hooks && nbstripout --install && git config diff.ipynb.textconv 'nbstripout -t'", description = "Install pre-commit hooks + the nbstripout notebook-output filter (run once per clone)" }

[dependencies]
python = "3.12.*"
# Pin numpy to avoid conda/pypi conflicts with numba
numpy = ">=2.0,<2.2"

# CellRank accelerated linear algebra (PETSc/SLEPc)
# Pre-built from conda-forge; PyPI wheels require building from source
petsc4py = "*"
slepc4py = "*"

# Shapely from conda to avoid pygeos build issues
shapely = "*"

[pypi-dependencies]
# Compatibility fix: cuML (rapids dependency) incompatible with sklearn 1.8
# https://github.com/rapidsai/cuml/issues/6426
scikit-learn = "<1.8"

# CPU JAX on all platforms by default (MPS handled automatically by torch on
# macOS). The `gpu` environment overrides this with the CUDA build on Linux.
jax = "<0.9.0"

# Core single-cell stack
scanpy = ">=1.10"
squidpy = "*"
scvi-tools = "*"
cellrank = "*"
torch = "*"

# Notebook workflow
jupyterlab = "*"
ipykernel = "*"
ipywidgets = "*"
# Pin pandas < 3.0 until anndata supports Arrow-backed strings
# https://github.com/scverse/anndata/issues/1434
pandas = ">=2.0, <3"
matplotlib = "*"
seaborn = "*"

# Utilities
pre-commit = "*"
nbstripout = "*"
pytest = "*"
session-info2 = "*"

# Local editable package for path helpers
myanalysis = { path = ".", editable = true }
igraph = ">=1.0.0, <2"
celltypist = ">=1.7.1, <2"

# GPU stack — only materialized in the `gpu` environment (Linux + CUDA 12).
# The default environment is CPU-only and is what CI installs; macOS local also
# uses the default. Activate `gpu` explicitly on a Linux/CUDA box (e.g. Euler):
#
#     pixi install -e gpu
#     pixi run -e gpu install-kernel
[feature.gpu.pypi-options]
extra-index-urls = ["https://pypi.nvidia.com"]

[feature.gpu.target.linux-64.pypi-dependencies]
# `rapids-cu12` extra pulls cuml/cupy/cudf from pypi.nvidia.com; without it
# rapids-singlecell installs without the GPU libs and silently runs on CPU.
rapids-singlecell = { version = ">=0.15, <0.16", extras = ["rapids-cu12"] }
jax = { version = "<0.9.0", extras = ["cuda12"] }

[environments]
gpu = ["gpu"]