final check before we submit to zenodo

[LinguList]

@marctang, this is almost ready now, I added a file .zenodo.json which will organize how this will be rendered on zenodo, but one can later also modify it. I suggest, if you tell me you're ready, we'll make the release and this will then be automatically uploaded to zenodo, so you have your DOI for the code, and can also link that e.g, on your website, etc.

[LinguList]

BTW: you had duplicate sources in your bibtex, I had to remove them.

[LinguList]

Checking this with Python is easy (now you have all the software installed):

>>> from pycldf.sources import Sources
>>> Sources.from_file(yourbibtex)

[marctang]

Blimey, sorry I forgot to check the duplicates in bibtex when I exported then, thanks for pointing it out!

I have a quick question for the references: not all of them were annotated with wals_code in bibtex, which makes (if I read the code correctly) the matching primarily based on wals_code miss some items in values.csv (e.g., mous_alagwa_2016), is it possible to use the match from GSSG_ListOfLanguages.csv ? or does the wals code has to be added in bibtex by default? Thanks!

[LinguList]

Ah, I did not know: do you have another key linking your data to the bibtex? If so, we can just replace it (you can just test this in the code). If there's no key to link sources and langauges, I'd not know how to proceed...

[LinguList]

Okay, sorry, I am stupid: you HAVE the source in yoru data. Of course, then we use the Source link to the sources.bib, and that's all. Do you want to do that yourself, @marctang? If there are problems, I'll help.

[marctang]

No pro, I added the Wals_code in bibtex now. However, a quick question about something that I might be doing wrong: some references are affiliated to more than one wals code (e.g., Kinkade-2001), and these references don't show up in values.csv @@. Maybe I am using a wrong format for the links? for now I use something like this:

@Unpublished{Kinkade-2001, Wals_code = {eya; klp; klm; mll; tli}, Title = {{T}he {A}real {Q}uestion: {N}orthwest {C}oast and {C}alifornia}, Author = {Kinkade, M. Dale}, Year = {2001}, Type = {paper} }

[xrotwang]

addressed here https://github.com/cldf-datasets/tangclassifiers/commit/dd8433a4fd5ea371ee7987fd6e1d27236ecbbddc

[LinguList]

@marctang, please check if this is fixed now.

[LinguList]

perfect, thanks @xrotwang.

[xrotwang]

@marctang the "missing source" issue seems fixed (verified with the csvstat command from csvkit):

$ csvstat -c Source cldf/values.csv 
  7. "Source"

    Type of data:          Text
    Contains null values:  False
    Unique values:         372
    Longest value:         31 characters
    Most common values:    Omar-1983 (16x)
                           Kinkade-2001 (10x)
                           Lynch-1998 (8x)
                           Suarez-1983b (8x)
                           Welmers-1973 (8x)

Row count: 800

[marctang]

Thanks to both for your help! I also did a re-run locally and the code works properly, same for csvstat! Awesome :-)! The output is correct and ready to be released on Zenodo!