Installation

Shapash requires Python 3.11 or newer. Install it with pip:

pip install shapash

Optional extras

Some features need extra dependencies, available as extras:

Extra	Use
report	Generate the Shapash Report.
nlp	Explain text classifiers (preview). Installs PyTorch, Transformers, Captum and related libraries.
lime	Use LIME instead of SHAP to compute contributions.
xgboost, lightgbm, catboost	Install a version of these libraries that Shapash supports.

For example:

pip install "shapash[report]"
pip install "shapash[report,lime]"

Jupyter

Shapash plots are Plotly figures. They display in JupyterLab, Jupyter Notebook 7 and VS Code without any extra configuration. To check that Plotly works in your environment:

import plotly.graph_objects as go

fig = go.Figure()
fig.add_trace(go.Scatter(y=[2, 1, 4, 3]))
fig.add_trace(go.Bar(y=[1, 4, 3, 2]))
fig.update_layout(title="Hello Figure")
fig.show()

If the figure doesn’t display, see Plotly’s troubleshooting guide.

Compatibility issues

If you get an error from one of the model libraries (XGBoost, LightGBM, CatBoost), install the matching extra to get a version that Shapash supports, for example:

pip install "shapash[xgboost]"