Preprocess text widget should remove stop-words

[PrimozGodec]

Text version

0.7.3dev

Orange version

3.24.0dev

Expected behavior

• Preprocess text widget should remove stop-words from the text for the Slovene language.
• Stopwords languages should be ordered in alphabetical order.

Actual behavior

It does not happen when stopwords from NLTK are used - the problem is spaces beside the stopwords.

Steps to reproduce the behavior

Read some data with Import documents and connect with Preprocess Text widget. Observe the output of Preprocess Text widget.

Additional info (worksheets, data, screenshots, ...)

[ajdapretnar]

@PrimozGodec So in essence NLTK has the wrong stopwords? 😱

[PrimozGodec]

They have additional space after each stopword. It at least holds for the Slovene language. The fix for this is in #456

[PrimozGodec]

I opened the issue on NLTK nltk/nltk_data#139, anyway I think #456 can be a quick fix. I think it is an OK solution since it is linear.