Installation

M2M-PostAViz is tested with Python version 3.12, 3.13 and 3.14.

At the command line:

pip install m2m-postaviz

To install the latest development version from source:

git clone https://gitlab.inria.fr/postaviz/m2m-postaviz.git
cd m2m-postaviz
pip install -r requirements.txt
pip install .

Dependencies

M2M-PostAViz dependencies (installed automatically with pip):

  • pandas

  • padmet

  • scipy

  • skbio

  • plotly

  • scikit-bio

  • shiny

  • shinywidgets

  • faicons

  • matplotlib

  • polars

  • pyarrow

  • seaborn

  • statsmodels

To install dependencies manually:

pip install -r requirements.txt

Troubleshooting

Polars compatibility (polars)

If installation succeeds but running m2m_postaviz fails because polars is unavailable or incompatible on your machine, install the long-term-support CPU wheel manually:

pip install polars-lts-cpu

macOS ARM binary mismatch (biom-format / scikit-bio)

On Apple Silicon (arm64), you may occasionally hit an error like:

  • ImportError: ... biom/_filter...so ... incompatible architecture (have 'x86_64', need 'arm64')

This means a transitive binary extension (from biom-format, required by scikit-bio) was installed for the wrong architecture.

In most cases, reinstalling biom-format with arm64 build flags is enough:

pip uninstall -y biom-format
ARCHFLAGS="-arch arm64" pip install --no-binary=biom-format biom-format==2.1.17

Then verify with:

m2m_postaviz --test

Citation

If you use the application for research, do not forget to cite the works associated to those dependencies.

Quickstart

To test the application with example data:

m2m_postaviz --test

It takes a few seconds to launch because data needs to be uncompressed and processed in a temporary directory. Shiny will launch automatically afterward.

A dataset with test data is available in this repository in postaviz_test_data and can be used to test the main functionalities of the tool.

Once on the homepage you’re free to explore the test data.

Note

Metacyc database information related to the ontology of metabolites and pathways is not included in test option.

Warning

We assume that you arrive at this step having installed the tool first (see above), for instance in a Python virtual environment, or conda (mamba) environment.