Ricardo Vieira

Deprecate `test_value` machinery

11

### Description This adds a lot of complexity for little user benefit (most don't know about this functionality in the first place)

### Description https://pytensor.readthedocs.io/en/latest/tutorial/adding.html I think we should leave types discussion for later and emphasize the laziness until compilation instead. Could also mention the general `tensor` class

Implement equivalent to `numpy.nan_to_num`

2

### Description A couple of switches should do the job, no need to implement a new Op https://numpy.org/doc/stable/reference/generated/numpy.nan_to_num.html

Consider replacing flake8/pylint by ruff in pre-commit

2

### Description Besides being way way faster, it would allow us to get rid of `setup.cfg` which AFAICT exists only because flake8 does not support `pyproject.toml`: https://github.com/PyCQA/flake8/issues/234. See #295 https://github.com/pymc-devs/pytensor/blob/main/setup.cfg

Use environment.yml (or something like that) in CI

1

### Description The CI dependencies are completely dissociated from the conda pytensor-dev environment specified by `environment.yml`. This led to need a separate commit in #448 that should have gone into...

Implement helper `@as_jax_op` to wrap JAX functions in PyTensor

3

### Description This blogpost walks through the logic for 3 different examples: https://www.pymc-labs.com/blog-posts/jax-functions-in-pymc-3-quick-examples/ and shows the logic is always the same: 1. Wrap jitted forward pass in Op 2. Wrap...

### Description This Composite `Op` computes both `Max` and `Argmax` and is returned by default when you call `at.max(...)`. This makes the graphs unnecessarily more complex from the get-go (see...

Marginalize over latent variables

35

I wonder whether it would be possible to rewrite the logp graphs to marginalize over finite discrete variables, indicated by the user (not necessarily all that are in the graph)....

Implement Truncated RVs

18

Closes #96

Transforms fail in FAST_COMPILE

1

```python import aesara import aesara.tensor as at from aeppl.transforms import TransformValuesRewrite, LogTransform x_rv = at.random.exponential() opt = TransformValuesRewrite({x_rv: LogTransform()}) logp, (x_vv,) = aeppl.joint_logprob(x_rv, extra_rewrites=opt) logp_fn = aesara.function([x_vv], logp, mode="FAST_COMPILE") aesara.dprint(logp_fn)...