search

calc-project / grouping-sounds

Datand Code Accompanying the Study "Grouping Sounds into Evolving Units for the Purpose of Historical Language Comparison"
MIT License
0 stars 0 forks source link

Minimal example #3

Closed FredericBlum closed 3 weeks ago

FredericBlum commented 10 months ago

@LinguList Here comes the minimal example for reproducing the error.

LinguList commented 10 months ago

Did you see my remark in the issue? I think I have proposed a fix, right?

FredericBlum commented 10 months ago

I saw it but only got to test it now. Seems to work fine. We might change the condition so that this only is triggered when the length is exactly 1. This would then still throw an error when trying to give an empty sequence. Alternatively, we could throw a warning and proceed. Should I set such a thing up?

LinguList commented 10 months ago

Throwing errors for empty sequences is useful, in my experience. It is a symptom of empty rows in tables. Setting it up would be nice. Maybe, we should think of making a mini-package with tests. Did you test it on different data?

LinguList commented 10 months ago

The test is useful for grouped sounds, since it points to cases where grouping is too adhoc and not systematic. I had errors in the Karen data we use as example, and had to correct them.

FredericBlum commented 10 months ago

Do we still use the functions that we defined here, or do we switch to "linse"?