Updated to new pylexibank about logos

[LinguList]

Nice. For Evobib, the new keys are important, as they allow for a consistent storing of Software across versions, with varying authors. The most recent version is, similare to latestdoi n zenodo, always the name of the software without versioning. I could switch to underscores, but bibtex keys in latex allow this treatment. One might tweak evobib in such a way that it returns another bibtex-key on demand.

[LinguList]

The question about evobipping these sources is also a general question. It has the advantage that updates will get included immediately.

See Ian's (Joo Ian) paper for which Kristina added the conceptlist yesterday: the paper is available ahead of print. This means, in half a year, we have to update it. If IDs stay consistent, this means that the updates are simplified.

The alternative is: just place the evobib in sources.bib, and never call the download function, and manually adjust it, as we also do in many datasets.

We need a discussion (maybe a thread/issue on pylexibank) of how to handle this, as bibtex-adding is one of the things that annoy me when adding datasets.

But the easiest way would be: use "evobibasbibtex" only if there's no "sources.bib", and if it's there, leave the sources.bib in raw, and have them for manipulating the bibtex, in case this is needed.

Similar to what I do now: I use iso2glottocode once, but then I add the language file in etc/languages.csv along with geo-coordinates from glottolog, as this makes the installation run faster...

[SimonGreenhill]

Yeah, I like the idea of just including the sources.bib file directly in lexibank, as this makes it more selfcontained and source information is fundmental.

Also note that getEvobibAsBibtex happily returns this if it can't find the reference, so we need to visually inspect the sources.bib anyway:

@misc{Zgraggen1980NA,
Note={missing source}

}

(the Madang dataset for example has been showing source=100% although it's actually 0 :) )

[chrzyki]

Oki, will change this! On 29 Oct 2019 11:38, Simon J Greenhill <[email protected]> wrote: Yeah, I like the idea of just including the sources.bib file directly in lexibank, as this makes it more selfcontained and source information is fundmental. Also note that getEvobibAsBibtex happily returns this if it can't find the reference, so we need to visually inspect the sources.bib anyway: @misc{Zgraggen1980NA, Note={missing source} } (the Madang dataset for example has been showing source=100% although it's actually 0 :) ) — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub<#1?email_source=notifications&email_token=ABDYZCYTSQCTOKDQ6FK5FGDQRAHDPA5CNFSM4JGGATEKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOECQAO6I#issuecomment-547358585>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABDYZCYQ6HHCQ3WAQB25GZ3QRAHDPANCNFSM4JGGATEA>.

[chrzyki]

Done in 1a65b8d.