BUG: TypeError: Variables do not support boolean operations with compute_test_value = "raise" for observed random variable

[vandalt]

Describe the issue:

Hi!

When running the tests for dfm/celerite2, with pytensor.config.compute_test_value = "raise", for PyMC v5 and v4, I encountered the error below. I'm adding a simple example to reproduce.

The error seems to come from the use of np.broadcast_to() on rv_var.shape here.
I believe replacing it rv_var.type.shape fixes the issue.

Should I submit a PR with the change? If I write a test with the example below, where is the best place to put it?

Thank you!

Reproduceable code example:

import numpy as np
import pytensor
import pytensor.tensor as pt

import pymc as pm

pytensor.config.compute_test_value = "raise"

rng = np.random.default_rng(123)
data = rng.standard_normal(100)
with pm.Model():
    # NOTE: pt -> np yields same error
    obs = pm.Normal("obs", mu=pt.zeros_like(data), sigma=pt.ones_like(data), observed=data)

Error message:

```python Traceback (most recent call last): File "/home/vandal/repos/astro/pymc/repro_error.py", line 13, in obs = pm.Normal("obs", mu=pt.zeros_like(data), sigma=pt.ones_like(data), observed=data) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/vandal/repos/astro/pymc/pymc/distributions/distribution.py", line 316, in __new__ rv_out = model.register_rv( ^^^^^^^^^^^^^^^^^^ File "/home/vandal/repos/astro/pymc/pymc/model/core.py", line 1315, in register_rv rv_var = self.make_obs_var(rv_var, observed, dims, transform, total_size) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/vandal/repos/astro/pymc/pymc/model/core.py", line 1356, in make_obs_var rv_var.tag.test_value = np.broadcast_to(test_value, rv_var.shape) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/vandal/repos/astro/exoplanet-dev/celerite2-pymc5/venv/lib/python3.11/site-packages/numpy/lib/stride_tricks.py", line 413, in broadcast_to return _broadcast_to(array, shape, subok=subok, readonly=True) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/vandal/repos/astro/exoplanet-dev/celerite2-pymc5/venv/lib/python3.11/site-packages/numpy/lib/stride_tricks.py", line 345, in _broadcast_to if any(size < 0 for size in shape): ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/home/vandal/repos/astro/exoplanet-dev/celerite2-pymc5/venv/lib/python3.11/site-packages/pytensor/tensor/variable.py", line 72, in __bool__ raise TypeError("Variables do not support boolean operations.") TypeError: Variables do not support boolean operations. ```

PyMC version information:

Output of `pymc.__version__`: 5.9.1+1.g58217cc8.dirty (I played with the code to debug the issue, but it's basically the main branch)

Context for the issue:

As mentioned above, the error affects dfm/celerite2 when running the test suite. Here is the output of that:

```python python/test/pymc/test_pymc_celerite2.py:115: _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ python/celerite2/pymc/celerite2.py:96: in marginal return CeleriteNormal( ../../pymc/pymc/distributions/distribution.py:316: in __new__ rv_out = model.register_rv( ../../pymc/pymc/model/core.py:1315: in register_rv rv_var = self.make_obs_var(rv_var, observed, dims, transform, total_size) ../../pymc/pymc/model/core.py:1356: in make_obs_var rv_var.tag.test_value = np.broadcast_to(test_value, rv_var.shape) venv/lib/python3.11/site-packages/numpy/lib/stride_tricks.py:413: in broadcast_to return _broadcast_to(array, shape, subok=subok, readonly=True) venv/lib/python3.11/site-packages/numpy/lib/stride_tricks.py:345: in _broadcast_to if any(size < 0 for size in shape): _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

self = Lt.0

def __bool__(self):
    # This is meant to prohibit stuff like a < b < c, which is internally
    # implemented as (a < b) and (b < c). The trouble with this is the
    # side-effect that checking for a non-NULL a by typing "if a: ..."
    # uses the same __nonzero__ method.  We want these both to work, but
    # it seems impossible.  Currently, all vars evaluate to nonzero except
    # the return values of comparison operators, which raise this
    # exception.  If you can think of a better solution, go for it!
    #
    # __bool__ is Python 3.x data model. __nonzero__ is Python 2.x.
    if self._is_nonzero:
        return True
    else:

      raise TypeError("Variables do not support boolean operations.")

E TypeError: Variables do not support boolean operations.

venv/lib/python3.11/site-packages/pytensor/tensor/variable.py:72: TypeError

</details>

[welcome]

🎉 Welcome to PyMC! 🎉 We're really excited to have your input into the project! 💖

If you haven't done so already, please make sure you check out our Contributing Guidelines and Code of Conduct.