search

LibCrowds / libcrowds

The frontend for the LibCrowds crowdsourcing platform
MIT License
33 stars 6 forks source link

analyse.libcrowds.com redirects to another website #854

Open harryjmoss opened 5 years ago

harryjmoss commented 5 years ago

@adikeinan has pointed out to me that analyse.libcrowds.com redirects the user to another website. I've had a look in the DigitalOcean settings and it's redirecting users to 178.62.51.149, which happens to be the website of a mathematics academic. @mialondon do you know the where this used to/should be redirecting to? I'm also not sure how long this issue has been around for!

libcrowdsredirect

mialondon commented 5 years ago

@harryjmoss no idea, sorry!

harryjmoss commented 5 years ago

Changed this to point to annotations.libcrowds.com for now - this throws a 404 error (which I think is expected?) but is better than it redirecting to an external website

adikeinan commented 5 years ago

Thanks @harryjmoss! Does LibCrowds have its own 404 page? I suspect not. Improvement for the future :)

harryjmoss commented 5 years ago

Thanks @harryjmoss! Does LibCrowds have its own 404 page? I suspect not. Improvement for the future :)

Completely overlooked this, but a much better idea than I had! Done

harryjmoss commented 5 years ago

@adikeinan Just to check - did analyse.libcrowds.com previously redirect users (with admin rights) to this page?

adikeinan commented 5 years ago

Totally awesome @harryjmoss, thank you!

adikeinan commented 5 years ago

@adikeinan Just to check - did analyse.libcrowds.com previously redirect users (with admin rights) to this page?

I don't think so but unsure - will ask!

adikeinan commented 5 years ago

Checked with one curator, she doesn't think this is the old 'analyse' page but not 100% sure. I'll wait for the other curator to have a look, but she's away so it'll take a couple of weeks @harryjmoss

adikeinan commented 5 years ago

The other curator (Emma) also thinks this is not the Analysis page that she used. The analysis feature that curators used was basically a staff sorting page in which they could sort through inconsistencies in the partial matches. The staff project looked like a normal LibCrowds project but was only visible to staff. Here are example links that used to lead to staff pages but are now broken: http://www.libcrowds.com/project/category/staff/ and http://www.libcrowds.com/blplugin/pinyincardcatalogue_d1_investigate/ Any chance you could investigate this @harryjmoss ?

adikeinan commented 4 years ago

Restarting the conversation on this issue: we don't know of a way to extract records which were not matched, or only partially matched. We do have access to records matched by 2 or more people, and these could be ingested. This page, very last link - to the results analysis process: https://github.com/LibCrowds/docs/blob/master/docs/data/downloads.md could have helped but the link is broken. It looks like this could have been where we would find the relevant info on what happens with partially/un-matched results. Any ideas, @harryjmoss ?

adikeinan commented 4 years ago

To recap on the issue of the 'Analyse' section:

Prayank Kulshrestha from NIIT on the 'Page not found' (url - https://www.libcrowds.com/analyse) error: "We looked into the source code available on the GitHub. We couldn’t find either the physical file with the name “analyse” or any route path which matches the given url."

The context for this issue: CaC contributors are asked to look at a digitised card and search the OCLC WorldCat database based on some of the metadata elements printed on it to see if a record for the book already exists in some form online. If a match is found, contributors then type the Library's unique “shelfmark” as printed on the card. When 3 different contributors make the same match, a process creates a list of new records for inclusion in the British Library's Explore catalogue.

The 'Analyse' section was there to help BL curators manage cards that were only partially matched with online records. We called this area ‘staff projects’ area. It looked like a normal LibCrowds project but was only visible to BL staff. This area helped staff sort through inconsistencies in the partial matches.

Next step: NIIT to estimate cost for the re-creation of the 'Analyse' section of the platform, following these requirements:

  1. Create a mechanism for identifying and extracting all partial matches from the main project and redirecting them into a separate, staff-only project (accessed via a log-in, or visible only to user accounts with Admin rights). Partial matches may include discrepancies in the OCLC records selected as a match for a particular card, or differences in the transcription of the shelfmark. There should also be a facility for processing all user-generated comments as some may assist with the resolution of partial matches, or else require records to be amended later.

  2. Create an interface that enables BL staff to go through partially matched records and make a decision on whether a match has been found or not. It should present all relevant information required to make this decision and facilitate any necessary changes to the matched record, shelfmark, etc. In case staff created a match, it should be added to the final list of successful matches ready for QA and ingest (these lists are available on the Data page: https://www.libcrowds.com/collection/convertacard/data)

  3. Enable BL staff to download a list of any records which were not matched at all. This should combine results from both the main project and the staff project into a single batch.

  4. Enable the generation of statistics reflecting every step of the process for record-keeping and cross-checking, i.e. full matches, partial matches and non-matches following the public project AND full matches and non-matches following the staff project.

  5. Optional: Set up an email alert system to notify staff when a drawer is complete and a new batch of partial matches is ready for processing on the staff project.