The big cooccurrence refactor

[cjweir]

The big refactoring is basically complete. The only draw back is that your kernel functions (if you are doing multiple window sizes) need to all be the same. This is a numba list of functions issue. It will take some work to add this back in I think... it can be another PR.

[codecov-commenter]

Codecov Report

Merging #99 (c2f43a7) into master (3bb7902) will decrease coverage by 0.05%. The diff coverage is 89.45%.

@@            Coverage Diff             @@
##           master      #99      +/-   ##
==========================================
- Coverage   90.55%   90.50%   -0.06%     
==========================================
  Files          28       32       +4     
  Lines        4491     4644     +153     
==========================================
+ Hits         4067     4203     +136     
- Misses        424      441      +17     

Impacted Files	Coverage Δ
vectorizers/edge_list_vectorizer.py	77.27% <ø> (-0.34%)	:arrow_down:
vectorizers/skip_gram_vectorizer.py	90.99% <ø> (-0.09%)	:arrow_down:
vectorizers/preprocessing.py	80.28% <58.18%> (-9.26%)	:arrow_down:
vectorizers/_window_kernels.py	75.74% <69.56%> (-8.20%)	:arrow_down:
vectorizers/utils.py	77.10% <71.42%> (+11.11%)	:arrow_up:
vectorizers/ngram_vectorizer.py	90.90% <84.84%> (+4.35%)	:arrow_up:
vectorizers/base_cooccurrence_vectorizer.py	88.19% <88.19%> (ø)
vectorizers/ngram_token_cooccurence_vectorizer.py	92.78% <92.78%> (ø)
vectorizers/coo_utils.py	93.22% <93.33%> (+1.00%)	:arrow_up:
vectorizers/__init__.py	100.00% <100.00%> (ø)
... and 10 more

:mega: Codecov can now indicate which changes are the most critical in Pull Requests. Learn more