Filtering with pyspelling.filters.context

[gllmflndn]

I am trying to filter out things like this :fontawesome-solid-graduation-cap: from Markdown files with pyspelling.filters.context:

  - pyspelling.filters.context:
      context_visible_first: true
      delimiters:
      - open: '^:'
        content: '[a-z0-9\-]*'
        close: ':$'

but I still get fontawesome flagged as a spelling error. Any solution for this?

[facelessuser]

I don't know what other filters you are using, but assuming you aren't using the Markdown filter and enabling the emoji extension to process the emoji, you could do something like this:

  - pyspelling.filters.context:
      context_visible_first: true
      delimiters:
      - open: ':'
        content: '[\w-]+'
        close: ':'

We don't know if they will be at the start of the chunk being evaluated so we don't want to specify ^. We also don't know if it will be at the end, so we won't use $.

[facelessuser]

I guess you may be processing the files individually. I usually create a task that builds my documents and then applies spellchecking to the rendered docs.

[gllmflndn]

It seems like you are Python Markdown extensions, why not enable things like snippets so you don't have to ignore --8<--?

Thanks for the advice! I added the following and removed the context filter:

  - pyspelling.filters.markdown:
      markdown_extensions:
      - pymdownx.snippets

[gllmflndn]

I guess you may be processing the files individually. I usually create a task that builds my documents and then applies spellchecking to the rendered docs.

I am indeed (well, from what I understand I am doing). So you suggest I first run MkDocs on my Markdown files and then call PySpelling on the generated *.html files.

[facelessuser]

That is what I'm doing. But you don't have to. I let MkDocs build the docs with the appropriate extensions installed, then I use the HTML filter to target what I care about and filter out anything I don't. Then I don't have to build up specific context filters as I just use CSS selectors. I can avoid code blocks, emoji, whatever.

[gllmflndn]

Thanks, that makes sense and matches your CI documentation: https://facelessuser.github.io/pyspelling/#usage-in-ci

I'll try again with this in mind 😄