matricize based on just the codomain length

[mtfishman]

Change matricize(randn(2, 2, 2, 2), Val(2), Val(2)) to matricize(randn(2, 2, 2, 2), Val(2)), since the number of dimensions in the domain can be determined automatically.

This also removes the dependency of TensorProducts.jl and move TensorProducts.tensor_product to this package (as tensor_product_axis), and introduces new interface functions trivial_axis, tensor_product_axis, and matricize_axes which dispatch on the FusionStyle and are used to determine that axes of the matricization.

Closes #69.

[codecov]

Codecov Report

❌ Patch coverage is 89.09091% with 12 lines in your changes missing coverage. Please review.
✅ Project coverage is 90.06%. Comparing base (7dc4184) to head (a0f686f).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/matricize.jl	78.26%	10 Missing ⚠️
src/blockarrays.jl	94.28%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #101      +/-   ##
==========================================
+ Coverage   89.34%   90.06%   +0.72%     
==========================================
  Files          16       17       +1     
  Lines         563      604      +41     
==========================================
+ Hits          503      544      +41     
  Misses         60       60              

Flag	Coverage Δ
docs	0.00% <0.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.