Meeting Notes for 19 Feb

General

• e-Lab Next Steps:
  • We will receive input from Jeanne. She will look at the e-Labs and give us feedback.
  • We will work on improving the Drupal list by adding more details to each item and comments.
• Edit is working with database queries to retrieve some stats from the user and datacatalog tables. Some of the questions to be answered:
  • last login date per user, count for logbook entries, plots per user, etc.

Cosmic

• Benchmark-comments:
  • Ken reported a delay between selecting a benchmark and being able to used for uploads. We need to follow up with Jeff Rodriguez and get the steps from the workshop in Mexico.
  • Blessing code needs a change: we have to replace greater by 'greater or equal' and less by 'less or equal' so we allow zero values to work OK. Edit made the change to the code, rolling out this weekend. Edit will send a list of files that 'failed to be blessed' so Tom can check which files failed to bless due to the problem in the algorithm.
  • Bob will look at the list of 'manually blessed' files. 31 of them shouldn't have been blessed. Then we will decide what to do about this. In the meantime, were removing the tool to manually bless from the blessing charts page. Rolling out this weekend.
  • Ken brought up the fact that the benchmarks show by detector rather than by user. We are not changing this at the moment.
  • Benchmark documentation/tutorials: we are still working with new documentation and tutorials. Hopefully, the latest will be rolled out with all the new updates this weekend.
  • Tom will call Peter Grach to give him more information about access to blessed/unblessed data.
• Stats and analyses:
  • Discussed how the current count for analyses works. We need to rewrite this portion and save the analysis stats in the database. For now, low priority.
• Analyses-speed:
  • Edit will run tests on 13 comparing analyses with/without static .thresh file to be able to compare speed/completion, etc.
• Analyses-test:
  • Edit will create some automatic testing for Bob to be able to launch the studies as a regression test.