[FIX] Preprocess Text: set highest absolute frequency

[noahnovsak]

Issue

Fixes #669

Description of changes

Changes the behaviour of absolute frequency filter to return no upper cutoff point when the values of both boxes are equal.

Includes

• Code changes
• Tests
• Documentation

[PrimozGodec]

@needNewUsername, was this PR closed by mistake?

[noahnovsak]

Yes, i thought it had been dealt with, but I now remember I was waiting for @ajdapretnar to check if this is the type of solution we are looking for, and make sure it doesn't break anything else.

[codecov-commenter]

Codecov Report

Merging #807 (1393608) into master (4a28f85) will increase coverage by 0.96%.
The diff coverage is 78.26%.

❗ Current head 1393608 differs from pull request most recent head a65f149. Consider uploading reports for the commit a65f149 to get more accurate results

@@            Coverage Diff             @@
##           master     #807      +/-   ##
==========================================
+ Coverage   74.95%   75.92%   +0.96%     
==========================================
  Files          75       76       +1     
  Lines       10189    10644     +455     
  Branches     1626     1788     +162     
==========================================
+ Hits         7637     8081     +444     
- Misses       2303     2308       +5     
- Partials      249      255       +6     

[ajdapretnar]

There are some issues with the spin box, unfortunately.

If I set min=2 and max=max, the results are ok.
Then I increase min to 3. The spin box value of max changes to 3 as well (visually), but the output is really max.
The spin box also allows me to pass max value smaller than min (i.e. min=5, max=3). There are a couple of options how to handle this (raising an error, reversing the values). But I think the nicest one is simply preventing the spin box to go below min value.

What do you think, @PrimozGodec?

[PrimozGodec]

@ajdapretnar I agree. I think that when the value of one spin changes the minimum/maximum values of the other one can be adjusted via setMinimum/setMaximum

[ajdapretnar]

This looks good to me, but I would sleep much easier if there was a test that would check this behavior. 😅
For example passing:

• 5, None (check that it turns to max and that the output is ok)
• 5, 3 (check that it turns to max and that the output is ok)
• raising the final 3 to 6 and checking that it remains a 6 (not max)

[ajdapretnar]

I can tell you that for deerwester you should get:

• 1 word ("of")
• 1 word ("of")
• 1 word ("of")
• then perhaps try also 3-4, where you'd get "system, user, the, trees, graph".

[ajdapretnar]

As I need this quite urgently, I'll open another issue with a request for tests.