Yurik added a comment.
@NavinoEvans I agree - feel free to take my implementation (which was already
working for any CSV-style inputs), and extend/adapt it. Ideally, it should be
merged upstream to the Blazegraph, so it should support any kind of CSVs. It
may make sense to have either some s
NavinoEvans added a comment.
Just a bit of info on why I've nominated this for the 2020 hackthon:
There is currently a major issue with storing statistical data in Wikidata,
which would be solved if we could upload the data to Commons as Tabular Data
files.
It's proving unsustainable to u
Smalyshev added a comment.
Its not a must have requirement, but if there are no performance or other major disadvantages,
Complexity is one. I'm a big fan of YAGNI :) I'm not 100% against supporting it, but doubtful whether it's actually needed.
I am not sure what would be the best way to split W
Yurik added a comment.
@Smalyshev the reason I made type as a string is to allow additional parsing parameters, e.g. ?start tabular:startDate 'date:-mm-dd'
MultiSearchIterator and binding params
Correct, most usages would be static, but in theory it might be possible to supply URL or other
Smalyshev added a comment.
Some thoughts on the implementation:
Parameter matching seems to be backwards:
?url tabular:url 'uri' .
?type tabular:type 'string' .
If I understand it right, 'uri', 'string', 'integer' etc. are types. As such, there's a limited set of them and they should be predicat
Yurik added a comment.
The first version of this feature has been implemented in Sophox -- see docs. At this point, it supports any GET request that returns CSV-style data (parsable by Java's CSVParser, with many parameters).
If @Smalyshev has any spare time to review the code at https://github.c
Yurik added a comment.
@Lucas_Werkmeister_WMDE I agree - I am planning to implement this feature for both WDQS and Sophox QS. For WDQS, it should only support tabular datasets, or possibly other respected sources.TASK DETAILhttps://phabricator.wikimedia.org/T181319EMAIL PREFERENCEShttps://phabricat
Lucas_Werkmeister_WMDE added a comment.
I don’t think we should allow integration with datasets in arbitrary URLs, for the same reasons that we don’t allow federation to arbitrary SPARQL endpoints (DoSing a website from WMF’s servers, license issues, …). Let’s restrict this to tabular data on Commo