search

GeoscienceAustralia / FSDF

FSDF - Foundation Base Spatial Data - Linked Data - High level models, information, etc
Other
2 stars 0 forks source link

Testing Excel to RDF to RVA publication method (FSDF/vocab-maintenance.md) #5

Open LongJenny opened 3 years ago

LongJenny commented 3 years ago

@dr-shorthair

Hi Simon,

Thanks for preparing vocab-maintenance.md, outlining the process of publishing in RVA via an Excel to RDF conversion. I found it easy to follow (even for someone new to data/vocabs like me) and well laid-out.

"FSDF Themes" has been published in RVA using this method. Can you review the published version and let me know if it all looks OK to you. Owner is Geoscience Australia - ICSM (GA - ICSM); I requested ARDC include you on this account, but if you have any trouble accessing please let me know.

Issues I've noted, and would appreciate your thoughts on, are:

  1. Hyphens and apostrophes in Excel text are not being correctly converted to RDF, and are appearing as clusters of other characters (e.g. see elevation and depth);
  2. We need to add a process step to vocab-maintenance.md, outlining how to create an RDF file for import into RVA from the webpage generated by SKOS Play! The method I used was to copy and paste text to notepad++ and save as a Turtle file, but I’m not sure if this was what was intended;
  3. None of the vocabulary metadata came across (see rows 6-15 in "basic example" from vocab-maintenance.md) - is that what you expected? I'm now assuming that content is entered in this section so it's readily available to copy and paste when uploading into whatever vocabulary tool is being used. I'll be updating my template according to RVA requirements, but would appreciate your confirmation that contents from this section are transferred manually (and maybe we need to note this somehow in the process document?)
  4. In the case of multiple altLabels, how should these be entered? Comma-separated?? Not an issue with the FSDF Themes vocab, but something we’ll need to be able to do for other entries…

FYI, I've also updated vocab-maintenance.md to clarify a couple of things. If this wasn't appropriate please re-revise:

(a) Code line 33: added clarification 'Note in this instance "pt" refers to place type and is specific to this example only' (b) Code line 75: added clarification …'(row 25 in this example)', as it wasn't clear to me what "Data Structure Row" referred to

Cheers, Jenny (LongJenny)

dr-shorthair commented 3 years ago

@LongJenny can you share the input (Excel) file please? Put it in https://github.com/GeoscienceAustralia/FSDF/tree/master/codelists

(Do you know how to work with Git or GitHub?)

LongJenny commented 3 years ago

Nope, sorry, I’m a GitHub newby. But dragging and dropping seems to have worked and hopefully if I’ve put it in the wrong place or named it incorrectly you can fix that? Anyway, it’s there now…

dr-shorthair commented 3 years ago

OK - I found the following bug:

LongJenny commented 3 years ago

Great, thanks Simon. I’ll have a look at it when I’m back in the office tomorrow

dr-shorthair commented 3 years ago

I also changed its name to something much easier – FSDF-themes.xlsx

dr-shorthair commented 3 years ago

Thx for the tweaks to the instructions BTW - this is how it should be done!

LongJenny commented 3 years ago

@dr-shorthair

Good pick up on the missing forward slash, and thanks for guidance re naming conventions for URI prefix code as well as filename.

Are you able to give me some feedback on all four of the numbered items I put in my original message in this Issue? e.g. it's still returning odd characters in the place of hyphens and apostrophes...

dr-shorthair commented 3 years ago

I've just run SKOS-Play! conversion again and I'm not seeing the problems 1, 3.

  1. theme:elevation-and-depth is as expected
  2. yes - just copy and paste the output from SKOS-Play!
  3. all the vocabulary metadata is there, associated with the ConceptScheme
  4. Commas can be used to separate URIs, but not values in a string by default anyway, as comma is a valid character that could appear in the middle of a string.

However, you can add another configuration option to force it to split a string into multiple values - see https://skos-play.sparna.fr/play/convert#split Another way to get multiple values is to insert additional lines for the additional labels. All the 'properties' of a URI get associated with just the one entry.

I've tested the following which illustrates both approaches, and it creates three skos:altLabel values.

image

LongJenny commented 3 years ago

Thanks for your thorough input - much appreciated.

This is what I'm getting instead of apostrophe (API view from RVA as I find that easier to comprehend): image

Sounds possible problem is my end only - I'll check on my computer at home and raise it with GA if displaying correctly on my device, but won't update this issue post immediately as I'm off now until next Thursday (loooong weekend!) If I can't resolve it I may check back with you late next week.

I hope you enjoy your weekend.

Jenny

dr-shorthair commented 3 years ago

Ah – I see. That is something deep in character sets and encoding. If you replace the apostrophe (which is a funny character) with a single quote (which is part of the ASCII character set) then it works (I just tested).

Lucky you for long weekend. We still can’t go further than 25km from home, and didn’t feel like we could plan while our lockdown was in place. We have lost so many holidays in the last 18 months, as we also lost 2020 summer holiday in Kangaroo Island due to the bushfires.