Visidata: Resizing columns and wide characters

[polm]

Adjusting column width with _ doesn't seem to work as intended with wide/full width characters as used in East Asian languages.

Consider this sample csv data:

壱弐参四五六七八九拾,ok,test-test
this is a test,test two, test three

On the second row everything looks fine, but on the first row _ is weird for all columns, not just the first one.

You should be able to compute the display width of a string with something like this:

import unicodedata

def get_display_width(text):
    width = 0
    for cc in text:
        if unicodedata.east_asian_width(cc) in 'FWA':
            width += 2
        else:
            width += 1
    return width

[anjakefala]

Hi @polm!

Can you clarify what you meant by it is weird for all columns?

I have been able to reproduce using benchmarks.csv that _ underestimates what the max width is for cells with special characters.

If: that is what you meant, yeah that is definitely an issue, nice catch. =) Thanks for letting us know. We will fix it up soon.

Else: Would you mind sharing a screenshot? :blush:

[polm]

If: that is what you meant, yeah that is definitely an issue, nice catch. =) Thanks for letting us know. We will fix it up soon.

That's the issue, yes. I though resizing was also weird for columns to the right of the one with wide characters (thus "all columns") but I can't reproduce it so I must have been confused.

Could I submit a PR to fix this? I see you have code like what I suggested above already but are just not using it in the column resizing part.

[anjakefala]

@polm Yes!

Please do. :blush: We use the develop branch for development. Please submit the PR against that branch. =)

[anjakefala]

@polm fixed the issue in https://github.com/saulpw/visidata/pull/284. Closing.