Closed david-caro closed 7 years ago
@jacquerie @kaplun @jmartinm what do you think? It's mainly to be able to have the desy spider, that directly ingests marcxml records, so we can avoid having to reimplement the 50+ spiders they currently have before we make labs the master.
so we can avoid having to reimplement the 50+ spiders
This is a fairly convincing argument...
Being fair, it's 47, but well... does not make much of a difference here.
BTW the separation would a nice time to solve https://github.com/inspirehep/inspire-next/issues/2248.
:+1:
-- David Caro david.caro@cern.ch
CERN - RCS-SIS inSPIRE-HEP High Energy Physics information system http://inspirehep.net On May 18, 2017 22:45, Jacopo Notarstefano notifications@github.com wrote:
BTW the separation would a nice time to solve #2248https://github.com/inspirehep/inspire-next/issues/2248.
— You are receiving this because you were assigned. Reply to this email directly, view it on GitHubhttps://github.com/inspirehep/inspire-next/issues/2356#issuecomment-302535951, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AAj1jATbnxEJ1UfxLI1nA3DJptZ9Rx0Mks5r7K27gaJpZM4Nfru_.
Yeah: inspire-dojson and we can be happy. Only issue (as for the inspire-jsonmerger) is that, since inspire-dojson
will depend tightly on inspire-schemas
we will have to be careful to always update inspire-schemas
and inspire-dojson
at the same time in inspire-next
. I.e. we might slightly complicate the dependency cicles, but at least it's for a very good cause of saving energy.
Now it would be a good time to do it: after https://github.com/inspirehep/inspire-next/pull/2470 we have parity between inspire-schemas
and inspire-next
.
In order to be able to use dojson easily from hepcrawl to ingest marcxml records, it would be really helpful to have dojson on it's own package so we can reuse the same code to transform the record to json.