Stopwords that contain apostrophe are not filtered about retext-keywords

[wooorm]

Thanks, @facundoolano, for reaching out and using retext-keywords.

Thing is, retext-keywords is based on parts of speech, not on stopwords.
The POS tags default to NN, which makes unknown words eligible for inclusion in the results, which is the reason this occurs.

Solving word detection

One fix would be to wait for dariusk/pos-js#21 to be PRd (I believe this is an issue there).

Another would be to create a retext plug-in which expands those elision examples (my preference), or alternatively preprocess the given input and expand them by regexes (easiest):

input.replace(/you[’']ve/, 'you have')/*...*/;
retext.process(input/*, ...*/);

The last method should fix this issue, as then the words will be recognised correctly.

Solving stop-words

Or, this project could start supporting given configurable stop-words. Would you be interested in working on a PR for that? I could help guide you through the project!

[facundoolano]

Thanks for answering. From your explanation it sounds like a plugin that handles the expansions before processing would work better than introducing stopwords. I'll consider working in such plugin in the future, but for now I guess I'll keep on replacing by hand before processing, so I can move forward with my current project.

[wooorm]

It’s quite some work, and I’m not very interested in working on it currently.
If someone wants this: create a PR on dariusk/pos-js!