Metadata-Version: 2.5
Name: unxts.interop.numpyro
Version: 2.2.0
Summary: NumPyro integration for unxt (canonical: unxts.interop.numpyro)
Project-URL: Bug Tracker, https://github.com/GalacticDynamics/unxt/issues
Project-URL: Homepage, https://github.com/GalacticDynamics/unxt
Author-email: GalacticDynamics <nstarman@users.noreply.github.com>, Nathaniel Starkman <nstarman@users.noreply.github.com>
License-Expression: BSD-3-Clause
License-File: LICENSE
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3 :: Only
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Programming Language :: Python :: 3.14
Classifier: Topic :: Scientific/Engineering
Classifier: Typing :: Typed
Requires-Python: >=3.12
Requires-Dist: numpyro>=0.22.0
Requires-Dist: plum-dispatch>=2.7.0
Requires-Dist: unxt>=2.1.0
Description-Content-Type: text/markdown

# unxts.interop.numpyro

NumPyro integration for [unxt](https://github.com/GalacticDynamics/unxt).

The `unitful` handler lets NumPyro models take `unxt` quantities. It converts them to plain numbers in a unit system you choose before NumPyro sees them. The model gets its sampled values back as quantities, and `attach_units` puts the units on the posterior samples.

## Install

```bash
pip install unxts.interop.numpyro
```

## Usage

```python
import jax
import numpyro
import numpyro.distributions as dist
import unxt as u
from numpyro.infer import MCMC, NUTS
from unxts.interop.numpyro import attach_units, unitful


def model(x=None):
    m = numpyro.sample("m", dist.Normal(u.Q(0.0, "kpc"), u.Q(3.0, "kpc")))
    s = numpyro.sample("s", dist.HalfNormal(u.Q(1.0, "kpc")))
    with numpyro.plate("n", 200):
        numpyro.sample("x", dist.Normal(m, s), obs=x)


x = u.Q(1.0 + 0.2 * jax.random.normal(jax.random.PRNGKey(1), (200,)), "kpc")
M = unitful(model, u.unitsystems.galactic)
mcmc = MCMC(NUTS(M), num_warmup=200, num_samples=200, progress_bar=False)
mcmc.run(jax.random.PRNGKey(0), x)
samples = attach_units(mcmc.get_samples(), M)  # samples["m"] is a Quantity in kpc
```
