Adding new reconciliation service

136 views
Skip to first unread message

Richard Guinn

unread,
May 20, 2020, 7:39:48 AM5/20/20
to OpenRefine
Hi all

I'm trying to add a reconciliation service to OpenRefine.  I'm using Free Your Metadata.  

I've chosen the Library of Congress subject headings.

It's taking a long time and am wondering if it just takes awhile to add or if something is 'off', either on my end or such.

I tried adding the reconciliation service last night and finally shut it down and I'm trying to add again this morning.

I followed these steps from the documentation:

Then, click the RDF button in the top right corner, select Add reconciliation serviceBased on SPARQL endpoint. In the dialog, you tell OpenRefine how to access the controlled vocabulary. For LCSH, provide the following parameters:

Name
LCSH
Endpoint URL
http://sparql.freeyourmetadata.org/
Graph URI
http://id.loc.gov/authorities/subjects
Type
Virtuoso
Label properties
check only skos:prefLabel

Apologies if I'm missing something obvious --

I get to this point and it keeps 'cycling':



What can I do differently to set this up??

Thanks!
Richard

Baer,Helen

unread,
May 27, 2020, 10:08:50 AM5/27/20
to openr...@googlegroups.com

Hi Richard,

 

I had the same trouble with the FreeYourMetadata LCSH procedure. As an alternative I tried loading one of LC's dump files into OpenRefine, thinking I could do a cell.cross to reconcile, but I could never get it to load in my computing environment (managed) even though I increased memory. I ended up using "Add column by fetching URLs" to get errors from the id.loc.gov site, which told me which headings were LCSH and which were not. This solution was  inspired by Ruth Tillman Kitchin's blog post One Weird Query: Resolving LC Subject Strings to URIs Using Python (2020). Ruth goes into detail about how the LC "label" function works.

 

I'm not quite finished with the procedure because I messed up the logic with splitting off subdivisions, but it's 95% there.

 

Hope this helps!

 

Helen

--
You received this message because you are subscribed to the Google Groups "OpenRefine" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openrefine+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openrefine/2699be42-93c4-4712-a228-6253b0c2fffe%40googlegroups.com.

Richard Guinn

unread,
May 27, 2020, 10:32:42 AM5/27/20
to OpenRefine
Hi Helen and thanks very much!  I'll dive into this and see how it works.
Richard

To unsubscribe from this group and stop receiving emails from it, send an email to openr...@googlegroups.com.

Parthasarathi Mukhopadhyay

unread,
May 27, 2020, 11:40:48 AM5/27/20
to openr...@googlegroups.com
Hello friends

I don't know whether I'm addressing your issues exactly or not but sharing our experience with lc-reconcile in OpenRefine:

1. Downloaded the script from here: https://github.com/cmharlow/lc-reconcile
2. Started lc-reconcile service locally in terminal (a> pip install -r requirements.txt (first time only); b> python reconcile.py --debug; My Python is Python 2.7.12)
3. It will show: Running on http://0.0.0.0:5000/
4. Can check in browser: http://localhost:5000/ or http://0.0.0.0:5000/
   should show both LcNAf and LcSH
{
  "defaultTypes": [
    {
      "id": "Names", 
      "name": "Library of Congress Name Authority File"
    }, 
    {
      "id": "Subjects", 
      "name": "Library of Congress Subject Headings"
    }, 
    {
      "id": "LoC", 
      "name": "LCNAF & LCSH"
    }
  ], 
  "identifierSpace": "http://localhost/identifier", 
  "name": "LoC Reconciliation Service", 
  "schemaSpace": "http://localhost/schema", 
  "view": {
    "url": "{{id}}"

5. In OpenRefine select Column to reconcile and "Add standard service" and then add http://0.0.0.0:5000/ in URL box (or your IP).
5.1 After adding we need to exit before "start reconciliation"

6. Tested with OR 3.1, 3.2 and 3.3

-----------------------------------------------------------------------
Dr. Parthasarathi Mukhopadhyay
Professor, Department of Library and Information Science,
University of Kalyani, Kalyani - 741 235 (WB), India
-----------------------------------------------------------------------


To unsubscribe from this group and stop receiving emails from it, send an email to openrefine+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openrefine/02382c79-4849-458b-88e1-b4c5011543be%40googlegroups.com.

Richard Guinn

unread,
May 28, 2020, 7:42:01 AM5/28/20
to OpenRefine
Thank you, Dr. Mukhopadhyay!

Baer,Helen

unread,
May 29, 2020, 11:05:47 AM5/29/20
to openr...@googlegroups.com

Hmmm, I think I need to learn some Python!

To unsubscribe from this group and stop receiving emails from it, send an email to openrefine+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openrefine/08163295-f9f9-41a2-84ca-e024af05f242%40googlegroups.com.

Sal Hamerman

unread,
May 29, 2020, 11:36:42 AM5/29/20
to openr...@googlegroups.com
Hi all --

Thanks for all the info! 
I've tried loading Harlow's lc-reconcile service via URL (trying both of the ones provided in the Github documentation) and I always receive the following error message: "Error contacting recon service: timeout : timeout - http://lc-reconcile.cmh2166.webfactional.com/"

Has anyone else encountered this error?
Is it just required to run the service on the local terminal to get good results? I'm a Python novice but will give this method a try. 

I've had good results with this service, though: http://refine.codefork.com/reconcile/viafproxy/LC
This VIAF-LC service is great for retrieving LC name strings, but it does return the VIAF identifier rather than the LC identifier. 

Thanks, all, for your insight on methods for reconciling LCNAF and LCSH. :) 

Best,
Sarah Hamerman
Project Cataloging Specialist, Rare Books
Princeton University Library



--
Sarah "Sal" Hamerman
Poetry Cataloging Specialist, Princeton University Libraries
Twitter || @srrhhamerman




Reply all
Reply to author
Forward
0 new messages