Fixing coverage in TextFeaturizer

[freddyaboulton]

Pull Request Description

With the removal of text_columns in #1652, the _get_feature_provenance method of TextFeaturizer and LSA was no longer being covered because the test dataset doesn't have any text features and the text features were specified with text_columns.

---

After creating the pull request: in order to pass the release_notes_updated check you will need to update the "Future Release" section of docs/source/release_notes.rst to include this pull request by adding :pr:123.

[codecov]

Codecov Report

Merging #1842 (4916f3a) into main (0fe8bdd) will increase coverage by 0.1%.
The diff coverage is 100.0%.

@@           Coverage Diff            @@
##            main    #1842     +/-   ##
========================================
+ Coverage   99.9%   100.0%   +0.1%     
========================================
  Files        255      255             
  Lines      20655    20658      +3     
========================================
+ Hits       20633    20650     +17     
+ Misses        22        8     -14     

Impacted Files	Coverage Δ
...understanding_tests/test_permutation_importance.py	100.0% <100.0%> (ø)
...lines/components/transformers/preprocessing/lsa.py	100.0% <0.0%> (+2.5%)	⬆️
...ents/transformers/preprocessing/text_featurizer.py	100.0% <0.0%> (+17.9%)	⬆️

---

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 0fe8bdd...4916f3a. Read the comment docs.

[chukarsten]

Cool, took me a second to understand what was going on, but cool.

[chukarsten]

Just for my edification, is the reason this test covers those lines because the change in logical types forces a derivation of the provenance?

[freddyaboulton]

Yea the provenance is only computed if the text featurizer creates features. If you run it on a dataset without text features, the provenance will always be an empty dict. Since the text featurizer will (now) only compute features on columns of logical type NaturalLanguage, we have to create one for the component to be able to do its thing.