Replace some uses of TensorType.broadcastable with TensorType.shape

[brandonwillard]

This PR contains important and independent TensorType.shape refactoring from #1280. Aside from superficial TensorType.broadcastable-to-TensorType.shape refactoring, these changes contain improvements to our static Type-level shape inference capabilities.

[codecov]

Codecov Report

Merging #1297 (4a8ea3d) into main (eace7f6) will decrease coverage by 0.00%.
The diff coverage is 85.19%.

❗ Current head 4a8ea3d differs from pull request most recent head 75274ef. Consider uploading reports for the commit 75274ef to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1297      +/-   ##
==========================================
- Coverage   74.12%   74.11%   -0.01%     
==========================================
  Files         174      174              
  Lines       48654    48673      +19     
  Branches    10353    10379      +26     
==========================================
+ Hits        36064    36076      +12     
- Misses      10300    10305       +5     
- Partials     2290     2292       +2     

Impacted Files	Coverage Δ
aesara/compile/debugmode.py	60.45% <0.00%> (ø)
aesara/link/c/params_type.py	80.41% <ø> (ø)
aesara/tensor/fourier.py	60.81% <0.00%> (ø)
aesara/tensor/nnet/corr.py	16.81% <0.00%> (ø)
aesara/tensor/nnet/corr3d.py	17.26% <0.00%> (ø)
aesara/tensor/rewriting/basic.py	92.12% <ø> (ø)
aesara/tensor/signal/pool.py	28.30% <50.00%> (ø)
aesara/sparse/sandbox/sp.py	73.48% <66.66%> (ø)
aesara/tensor/elemwise.py	88.59% <75.00%> (-0.15%)	⬇️
aesara/tensor/nnet/abstract_conv.py	18.80% <81.81%> (+0.18%)	⬆️
... and 27 more