LB 2.0 rc01

[FredericBlum]

@LinguList @chrzyki @johenglisch @xrotwang I have made some final changes to the representation of collections (individual columns for LexICore, ClicsCore, etc). I also went through workflow.md to update all scripts accordingly and to make sure that everythings runs smoothly. It would be great if you could check this PR.

Following the merge, I propose we publish an alpha release that can be used for the initial submission of the ddescriptive article of the updated release.

[LinguList]

@FredericBlum, can it be that you do not code for 0 in Selexion? I looked at the first three lines only, and we find this:

aaleykusunda-KusundaGM,Gyani Maiya,Eurasia,28.0,82.26,kusu1250,,aaleykusunda,224,224,224,LexiCore,1,1,0,0,,,,Kusunda
aaleykusunda-KusundaK,Kamala,Eurasia,28.0,82.26,kusu1250,,aaleykusunda,227,227,227,LexiCore Selexion,1,1,0,0,1,,,Kusunda

More specifically:

LexiCore,1,1,0,0,
LexiCore Selexion,1,1,0,0,1

For consistency reasons, we'd expect a 0 in the first row, that has only LexiCore, right?

[FredericBlum]

Found the problem and am fixing it. The commit will probably have to wait until tomorrow morning.

[chrzyki]

Thanks for preparing this. Still working through the changes and testing everything. It seems as if wordlist-metadata.json doesn't specify the new Incollections column in its metadata.

[chrzyki]

Given the outliers for ConsonantQualitySize the color coding on the map doesn't make for the best example I think. Is there a different, sensible, feature or maybe filter for outliers? Does make it a bit better.

[FredericBlum]

How did you filter for outliers there? Not sure how this is done with CLDFviz. I like your version, and since the main point is not about the feature but about the continuous variable plotting, I would be completely fine with that version.

[chrzyki]

How did you filter for outliers there? Not sure how this is done with CLDFviz. I like your version, and since the main point is not about the feature but about the continuous variable plotting, I would be completely fine with that version.

After inspecting the data, I removed some glaring outliers with:

cldfbench cldfviz.map cldf/phonology-metadata.json --parameters ConsonantQualitySize --language-filters '{"Name":"^(?!Nyagrong Minyag|Chuanqiandian, Northeast Yunnan|rGyalrong|Bagvalal$).*$"}' --colormaps plasma --pacific-centered

[FredericBlum]

Cool, thanks! I added a new plot accordingly with CVQualityRatio, seemed like a good fit