recent update at cdli
Émilie Pagé-Perron
Dear all,
I updated the cdli software Sunday morning with a few additional tweaks during the day. Unfortunately, we have inadvertently introduced a small issue in editing metadata but this will be fixed in the next few days.
Here are the changes to expect:
==== Search ===
Lars has refactored the search features in order to resolve virtually all issues we were encountering with the search so far. I only list a sample of fixed issues and new features, but don’t hesitate to check it out and ask questions!
For now, we are only releasing the new search as a shadow of the current search in case we find unexpected issues. Because of that, I would be grateful if you could spend 20 minutes or more, trying out the new search and search options here:
https://cdli.mpiwg-berlin.mpg.de/new-search
https://cdli.mpiwg-berlin.mpg.de/new-search/advanced
https://cdli.mpiwg-berlin.mpg.de/new-search/settings
If you follow a link to search in the main menu, it will bring you back to the main search so make sure you see “new-search” in the url when testing.
You can also try on tablet and mobile. Please report any issues you find in the search behavior and in the display of the results at cdli-s...@orinst.ox.ac.uk (preferably one email per separate issue).
We will fix all reported issues and replace the current search with this new version as soon as possible.
Solved issues and new features:
Filters are shown alphabetically but period in period order
Many new filters added to restrict search more granularly (only entries with photos, only entries with no genre, etc)
The share search button was fixed and selected filters are now also shared
Empty search is now possible, you can search for nothing and get ALL results like before
Text search now also works using unicode and other conventions characters (eg. š and sh instead of sz), and for selected field that contain transliterated data, (eg. proveniences, etc) eg. s can be used instead of š.
Ported the transliteration features of multiple terms search, single line search, case sensitive search
Wildcard and literal search are now possible (use * and double quotes)
Sign value permutation search is working as expected with added features (search for any signs or sign values and all values for all these signs appearing in the same sequence will be searched for and results highlighted; try it from advanced search).
Add a series of P, Q or S numbers in a url to fetch those entries as search results (eg. https://cdli.mpiwg-berlin.mpg.de/P123456,P123457,P123454 or https://cdli.mpiwg-berlin.mpg.de/Q000001,P000001,S000002)
Partial searches now return expected results (eg museum number “UM 29-15-155” for “UM 29-15-155 (cast)” and excavation number ”W 20274” for “W 20274, xxx” ; these also work perfectly using the “identification numbers” field in simple search)
Text search results highlighting improved with surface and column information displayed
Searching for a seal number or a composite number will bring both the seal and impressed artifacts, or the composite text and its witnesses
Sorting is now possible in simple and compact displays (yay!)
When searching text, illogical results are not returned (eg. do not return 'ku3' for 'ku')
Search settings are saved properly and all of them have effect
New search settings added, eg: show results in compact mode by default
Faster search and rendering (1000 results work fine)
Searching for specific text strings, genres, or others would sometimes bring up an error, now they don't
Excavation number was missing from the advanced search, it’s now in there
There is a small issue with the width of the columns in the compact view of search results, this will be resolved soon.
==== Editing ===
There is now a “concatenate data” switch when uploading metadata changes. With this option enabled, one can add publications data, additional materials or others to the existing data, without having to also provide the existing data. This means existing data is not overwritten for those fields. When trying it out, double-check the result before submitting to the review queue. (Added by EPP)
Submitting bulk annotations sometimes triggered an error which Lars has now fixed.
When looking at the review queue to accept or decline a change in text, the text diff was showing the new text only, now showing the diff so it’s easier to evaluate the change. (Fixed by Lars)
When merging publications duplicates, the exact reference field is now pre-filled from the proper publication and not the artifact’s primary publication. (by EPP)
=== Data ===
I have uploaded the MTAAC dataset of manually morphologically annotated texts. This is a set of almost 400 Ur III administrative texts which now will have a “textual annotation” accordion on their artifact page. The annotations have been available for some time on GitHub but will not be accessible through the cdli site (also for download from the site when these artifacts are in your search results). I’ll be periodically adding more such annotation data as we have other sources we used during the project. (Added by EPP)
A quick thanks to our editors which have adapted to the new website and keep improving the dataset: R. Firth, A. Jagersma, and R. Rattenborg in particular, also my co-directors of course (J Dahl and B Lafont) but thanks also to individuals who have opened an account and started submitting changes to the crowdsourcing queue like J de Ridder, A. Anderson, L-S Zimmermann, D Sánchez Muñoz, and A Deloucas! Using this method saves us a lot of time and reduces the risk of errors. Thank you! (We’ll also announce this in the wider-reaching newsletter sometime after Xmas).
=== Other ===
Link to cdli wiki had been added to the resources pages and menu. (EPP)
Lars has strengthened the security of the API.
Different small fixes were applied to various pages (Lars)
Aditi fixed a sneaky bug that introduced extra information before the output in exported data.