Installation¶

PyMetaAnalysis supports Python 3.10 through 3.13. The numerical core depends on NumPy, pandas, and SciPy; Matplotlib is optional.

Install from PyPI¶

Install the core package with:

python -m pip install PyMetaAnalysis

The distribution name is PyMetaAnalysis, while Python code imports the shorter module name:

import meta_analyze as ma

Optional plotting support¶

Install the plot extra for forest and funnel plots:

python -m pip install "PyMetaAnalysis[plot]"

Numerical analyses do not import Matplotlib. Calling a plotting method without the extra raises an ImportError containing the installation command.

Install a source checkout¶

Clone the repository and install it in editable mode:

git clone https://github.com/ZhaoboDing/PyMetaAnalysis.git
cd PyMetaAnalysis
python -m pip install -e ".[test,dev,docs,plot]"

The extras have separate purposes:

Extra	Contents
plot	Matplotlib plotting backend
test	pytest, Hypothesis, coverage, and plotting test dependencies
dev	Ruff, Mypy, type stubs, and build tooling
docs	MkDocs
notebook	JupyterLab, notebook execution, and Matplotlib

To run the committed example notebook locally:

python -m pip install -e ".[notebook]"
python tools/execute_notebooks.py

Verify the installation¶

python -c "import meta_analyze as ma; print(ma.__version__)"

Then run a small common-effect analysis:

import meta_analyze as ma

result = ma.meta_analysis(
    effect=[0.1, 0.3, 0.2],
    variance=[0.02, 0.03, 0.01],
    model="common",
)

assert result.k == 3
print(result.summary())

Upgrade or uninstall¶

python -m pip install --upgrade PyMetaAnalysis
python -m pip uninstall PyMetaAnalysis

PyMetaAnalysis is early-stage software. Pin the version used by a consequential analysis and record it with the analysis environment.