Closed wsalesky closed 8 years ago
@davidamichelson @suellenstringer-hye First pass at Gazetteer data transform. There is a record for each, and also a record that for all data, was not sure what format was best. Please let me know how it looks. I did some spot checking but I'm sure we will need to do some tweaking.
Thanks Winona,
Much appreciated! ---Suellen
From: Winona Salesky [mailto:notifications@github.com] Sent: Friday, March 04, 2016 10:17 AM To: srophe/Linked-Data Cc: Stringer-Hye, Suellen Subject: Re: [Linked-Data] Write transform (#1)
@davidamichelsonhttps://github.com/davidamichelson @suellenstringer-hyehttps://github.com/suellenstringer-hye First pass at Gazetteer data transform. There is a record for each, and also a record that for all data, was not sure what format was best. Please let me know how it looks. I did some spot checking but I'm sure we will need to do some tweaking.
— Reply to this email directly or view it on GitHubhttps://github.com/srophe/Linked-Data/issues/1#issuecomment-192340466.
Great! Let me know if/when we need to make adjustments.
Look forward to looking at this!
Hi Winona,
We are running into validation errors on all-places.ttl and many of the individual place .ttl files.Some of them seem to be Xquery errors (according to Cliff) but others are maybe errors with the data itself. Do you have a validator you can use to correct these errors? We started trying to do it ourselves but maybe there is an automated method? Plus we wouldn't want to have to validate each time we regenerate. Stardog won't let us upload unvalidated files.
Thanks much, Suellen and Alex
Sorry! I will take a look. What do you use to validate .ttl? -Winona
We used
https://www.npmjs.com/package/turtle-validator
(pretty sure...Cliff took us there through a back door but we are 99% sure that's what we used)
---Suellen
Great thanks. Sorry about the errors, was trying to push the data before I left for vacation!
No, we're glad you did...some of the individual files were fine so we are happy to have had a chance to upload them as proof of concept.
Okay, here is the second version. If I have time later tonight I may try loading them into a triple store. I corrected the XQuery errors, and handled all the validation errors reported by the above turtle validator. (Mostly quotes within labels.)
Seems to load okay in my version of StarDog.
Thanks, @wsalesky. We look forward to trying it out.
Will use XSLT. It processes faster then xquery in my experience with eXist.
@davidamichelson We should discuss if we want this run as a cron job and stored in the db? Stored on github?