MDV Tools

Multi-Dimensional Viewer (MDV) is a web-based application for analysing and visualising large, multi-dimensional data (single-cell, spatial, and general tabular data). mdvtools is the Python package for creating, manipulating, and viewing MDV projects — including a lightweight local server and the ability to export a project as a static website.

• Homepage / source: https://github.com/Taylor-CCB-Group/MDV
• Issues: https://github.com/Taylor-CCB-Group/MDV/issues

Installation

pip install mdvtools

This installs the slim core: everything you need to build, manipulate, view, and serve MDV projects, including spatial-data support.

For the full application — database-backed project management, the chat/LLM assistant, and authentication — install the app extra:

pip install "mdvtools[app]"

If you call a full-app feature on a slim install, you'll get a clear message telling you to pip install "mdvtools[app]" rather than a cryptic import error.

Testing a pre-release from TestPyPI:

pip install --index-url https://test.pypi.org/simple/ \
            --extra-index-url https://pypi.org/simple/ mdvtools

Quickstart

Create and serve an MDV project from the pbmc3k example dataset. folder is where the project is stored — any location you can write to; it's created if it doesn't exist.

import scanpy
from mdvtools.conversions import convert_scanpy_to_mdv
from mdvtools.serverlite import serve_project

folder = "/path/to/mytestproject"
data = scanpy.datasets.pbmc3k_processed()

project = convert_scanpy_to_mdv(folder, data)
serve_project(project)            # serves at http://localhost:5050

Open http://localhost:5050 in your browser to explore the project. The project files live in folder and can be re-served later or exported as a static site.

There's also a command-line interface:

mdvtools --help

License

GPL-3.0-only. See the repository for full details.