Analyse() should return the score as well

[walles]

Is there an existing issue for this?

• I have searched the existing issues

What problem does this feature solve?

Chroma can classify text by its contents:

lexer := lexers.Analyse("package main\n\nfunc main()\n{\n}\n")

With this API though, it's not possible for me to know if even the best match is bad.

I would like to find that out, so that I can just not highlight if the text contents is uncertain.

What feature do you propose?

One possible suggestion would be to change the API to this...

lexer, certainty := lexers.Analyse("package main\n\nfunc main()\n{\n}\n")

... where certainty is a number on a well defined and documented scale.

Then, if I feel this number is too low, I could choose not to highlight anything.

[alecthomas]

Seems reasonable, how could we do this in a backwards compatible manner?

[walles]

Maybe this?

lexer, certainty := lexers.AnalyseScore("package main\n\nfunc main()\n{\n}\n")

Possibly in combination with deprecating the existing function since it's sort of unpredictable.

[alecthomas]

In what way is it unpredictable?

[walles]

Not sure if "unpredictable" is the right word, but let's say:

1. I get a file
2. lexers.Analyse() says it's a C program, with 1% confidence

This means that even though C is the "best" guess, it's still a bad guess, and it might be better to not highlight at all.

That's why I'd like to have the confidence number as well to be able to make this judgement.