Spatialxgene

v0.2.1 — Browser-based spatial transcriptomics viewer and preprocessing (cellxgene-style). Launch the viewer to explore spatial plots (zoom/pan, color by gene or annotation, optional tissue image), or run infer / preprocess on raw h5ad.

Requirements

• Python ≥3.9
• If you install from PyPI (pip install spatialxgene): no other requirements; the frontend is included.
• If you install from source (pip install -e .): Node.js is needed to build the frontend once (see docs/development.md).

Install

# From PyPI (recommended; frontend included)
pip install spatialxgene

# From source (e.g. for development)
git clone <repo> && cd spatialxgene
pip install -e .
# Then build frontend once: see docs/development.md

Commands

For development (frontend watch): spatialxgene dev — see docs/development.md.

Quick start

1. Launch the viewer

   spatialxgene launch --data-dir /path/to/preprocessed/h5ad/dir
   

   Opens http://127.0.0.1:8000. Use the sidebar to pick a dataset, color by annotation or gene, choose a tissue background image (if present), and zoom/pan the plot.

2. Infer spatial (when raw h5ad has no coordinates)

   spatialxgene infer /path/to/raw.h5ad
   # or a directory: spatialxgene infer /path/to/dir
   

3. Preprocess expression (QC, reduce, cluster)

   spatialxgene preprocess /path/to/raw.h5ad
   # Writes *_preprocessed.h5ad (or input/preprocessed/ for a directory)
   

Data format (viewer)

Preprocessed h5ad for launch should have:

• Spatial: obsm['spatial'] or obs['x'] and obs['y'].
• Expression: X, var index = gene names.
• Annotations (optional): e.g. obs['leiden'] for coloring.
• Tissue images (optional): Visium-style uns['spatial'][sample]['images'] (e.g. hires, lowres) and scalefactors.

More options: spatialxgene launch --help, spatialxgene infer --help, spatialxgene preprocess --help.

Documentation

• CHANGELOG.md — Release history.
• docs/ — Development, project layout, design, and plans: development.md, master_plan.md, implementation_plan.md, design_guidelines.md, tasks.md.