Remove special-cases for scalars in contractions

[eric-wieser]

It's possible this will be marginally slower, but it makes everything much easier to read.

[utensil]

So, should it actually be removed?

[eric-wieser]

I need to investigate this some more. Some of the failures are just addition reordering, which can be easily fixed.

The other half look like actual bugs. Leaving this as a draft for now, I probably need a piecewise approach to work out which bit is broken.

Edit: fixed all the issues in the earlier PRs, this now works just fine

[codecov]

Codecov Report

Merging #99 into master will decrease coverage by 0.31%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master      #99      +/-   ##
==========================================
- Coverage   67.54%   67.23%   -0.32%     
==========================================
  Files           8        8              
  Lines        4794     4785       -9     
==========================================
- Hits         3238     3217      -21     
- Misses       1556     1568      +12

Impacted Files	Coverage Δ
galgebra/ga.py	73.77% <100%> (-1.32%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 159cc2a...3e7c389. Read the comment docs.

[eric-wieser]

Some timings:

In [2]: from galgebra.ga import Ga

In [3]: g, *a = Ga.build('e*1|2|3|4')

In [4]: A, B = g.mv('A', 'mv'), g.mv('B', 'mv')

Before:

In [5]: %timeit A < B
102 ms ± 2.34 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)

After:

In [7]: %timeit A < B
105 ms ± 1.91 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)

No significant change in the general case. Probably faster for a pure scalar contraction, but that doesn't seem like a case worth optimizing for.

[utensil]

We can always have them back if one day these special treatments are really needed.