Meeting Minutes for 07/02/2019
Discovery Committee Meeting
July 2, 2019
Minutes
- 2019.06 deployment delayed to 07.09.19 due to holiday
- Sierra API Extract
- Pascal rewrote the API Extract process
- Added new error checking processes
- Put it in action on the Marmot, LION, and Aurora’s test servers
- Records are added to Pika through two process :
- First, preferred method: Sierra generates a MARC file for pika to download
- But, there is a limit to the number of records generated with a time-period (probably an hour)
- Second, more common: Reconstruct a MARC file for pika using calls for bibliographic and item data
- Generated record will contain the majority of the data records generated in the normal process, but not all.
- Will contain any data required for use in Pika
- Special handling is required for the Bib-level Fixed Fields (typically the 998 tag in generated records)
- Pika fetches the fields used in processing but ignores fields not used by Pika
- Locations, matType, bCode3 (bib-suppression)
- Special handling for Item record variable fields
- Item-level call numbers, as well as item-level volume information
- ILS econtent item level URLs
- Marmot only: item level econtent descriptor (non-MARC field g)
- First, preferred method: Sierra generates a MARC file for pika to download
- Bibliographic Changes
- Now, the ILS will be getting bibliographic changes throughout the day
- Sometimes this requires updating the record grouping (Changes to the title, author or format determination)
- Record Grouping during the day
- When a record’s grouping changes during the day, the time between the grouping change and the reindexing of that change, causes “ghost search results.” The old grouping will be reflected in searches, but viewing the specific record page with the changed grouping will not show.
- Apparent delay in the API for Aurora Public library
- Their server clock was behind by 5 minutes
- Built a tool to test for this delay
- No other Sierra sites exhibited a significant delay
- Migration Process
- Get Updated API key from all Sierra sites
- With access to the API version 5 and expanded permissions
- For July, move pika test sites onto API Extract
- Test to uncover any site-specific issues - some additional testing by your site’s catalogers will help
- For August, move pika production sites onto the API Extract, but retain the nightly full-export
- After an evaluation period by the library, the Scheduler job for the full export can be turned off
- If your library only has Scheduler in order to implement the full export, you can then cancel your Scheduler subscription
- But retain the create list query for the full export for occasions with a manual full export is needed
- Get Updated API key from all Sierra sites
- Sierra API Key and Sierra API Extract Migration
- Process and other considerations
- Call for assistance with testing (Sierra libraries only)
- Things needed
- Mark a record for re-extract
- Mark an ILS indexing profile for full re-index
- When there is not a nightly export, settings changes that should have an effect on all search results will need to cause all ILS records to be reindexed during the full update process
- e.g. Loan Rule/Loan Rule Determiner updates
- e.g. Indexing setting changes
- When there is not a nightly export, settings changes that should have an effect on all search results will need to cause all ILS records to be reindexed during the full update process
- Full extraction of all Sierra records
- Without a full export, any mistakes will accumulate over some interval of time
- Overdrive extract
- Time limit for processing
- Or; decouple from ILS extraction/reindexing cycle
- Overdrive on its own extract, then reindex cycle
- Sierra on its own extract, then reindex cycle with priority over the Overdrive cycle
User Rating Stars Discussion
- Currently, anything that is rated 3 stars and above will display in a filter: User Rating: Three Star & Up
- Pacal wondered if the 2 star rating should be changed to 2 star and below, or 1 star only, or is 2 star and up just fine
- Pascal explained that the way the facet populates is every grouped work has an average. They take all the group ratings values and create an average value. These average ratings are used to populate the User Rating facet
- User ratings and user tagging are all on the group work level
- The group decided that it would be best to have each star represent a single star rating and not a group of star ratings
- A question was asked about being able to select multiple ratings
- Pascal responded that this functionality is not available at this time, since it is using the “and” filter
Demo of new features, bug fixes, and documentation for upcoming release
Discussion of development priorities for the next release
Current Release
- Sierra API Extract
- Pascal will handle the extract pieces that he laid out earlier
- Sierra API Patron Integration
- Chris is working on the patron interactions
- Horizon ROA Driver
- Chris is working on self-registration
Next Quarter
- Record Grouping Adjustments
- Pascal will be working on this in a few more releases
- Pascal would like to take three releases to perfect the second version of the algorithm, before putting it into production
- It will change what a group work ID means, so he wants to capture as much in one change with production
- Control of Econtent Manifestation Display in Search Results and Grouped Work Views
- Chris will be working on this in a few more releases
- Programming Language Updates
- Chris will handle the details of the updates the affects the code in numerous ways
Short Term
- Clean Up Search Bugs
- These are broken out into the Long Term section
- Long Term
- Stability Version - Search: Advanced Search
- Stability Version - Search: Language
- Stability Version - Search: Display and Interface
- Stability Version - Search: Format Determination
- Stability Version - Record Grouping: Language
- Stability Version - Record Grouping: Improved handling for grouping category: Movie
- Stability Version - Record Grouping: New grouping category: Graphic Novel
- Stability Version - Search: Combo Handling
Other discussion topics
- Revisiting Code Contribution process
- Any additional questions or feedback, please contact Marmot
- Sierra API Key and Sierra API Extract Migration
- Process and other considerations
- Call for assistance with testing (Sierra libraries only)
- An email will be sent out asking Sierra Discovery Partners to create a new API key, because the keys are tied to a specific version
- Pika may not have access to any new things in the new API version
- This might help fix things we thought were a problem
- This will make the Sierra picker uniform as we go through this migration extract process
- Ashley will be working with libraries on testing on their Pika test sites
- Create a couple of dummy records to see how long it takes them to show up on your test site
- Create two records with the same title to see if they group or regroup after making edits
- Test to see how long it takes for edits to a record to display in your test site
- The typical cataloging changes that you have to wait overnight to see in Pika, should take a couple of hours with the API extract
- Delete a bib to see how long it takes to no longer be on your Pika test site
- Create a couple of dummy records to see how long it takes them to show up on your test site
- Just evaluate the preformation of the new extract process
Next Meeting is Tuesday, August 6th
Meeting Date:
Tuesday, 2019, July 2
Documentation Type:
Meeting Minutes
Link to File:
Committees:
Discovery Committee