Metadata-Version: 2.1
Name: bayes_traj
Version: 1.0.0
Summary: bayes_traj
Home-page: https://github.com/acil-bwh/bayes_traj
Author: James Ross
Author-email: jross@bwh.harvard.edu
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: Topic :: Software Development :: Build Tools
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Description-Content-Type: text/markdown
License-File: LICENSE.txt
Requires-Dist: provenance-tools>=0.0.5
Requires-Dist: pandas<2.2.2
Requires-Dist: numpy>=1.26.4
Requires-Dist: matplotlib>=3.3.1
Requires-Dist: scipy>=1.5.2
Requires-Dist: argparse>=1.1
Requires-Dist: statsmodels>=0.11.1
Requires-Dist: torch>=2.0.1
Requires-Dist: pyro-ppl>=1.8.5
Requires-Dist: pytest>=7.0.0
Requires-Dist: numexpr>=2.10.0
Requires-Dist: bottleneck>=1.3.8

# bayes_traj

Algorithms and tools for Bayesian trajectory modeling

```sh
pip install provenance_tools
pip install bayes_traj
```

## Overview of code

* `mult_dp_regression.py` is the current non-pyro model
  "multivariate Dirichlet process regression"
* `generate_prior.py` creates a prior file.
* `viz_data_prior_draws.py` prior predictive checks.
* `bayes_traj_main.py --out_model PICKLE_FILE` saves a trajectory model
  configured by other tools.
  This inputs a prior file.
* `viz_model_trajs.py` is posterior predictive check,
  visualizes the learned clusters.
* `sumarize_traj_model.py` prints additional info.
* `assign_trajectory.py` reads in the model + data.csv and outputs a
  data.csv file with one appended column that is the assigned trajectory.

Other files are legacy.

## Running tests

```sh
pytest
```
or when debugging
```sh
pytest -vsx --pdb
```
