Skip to content
Snippets Groups Projects

added option to change History.squeeze behavior

Merged David Schäfer requested to merge flag-aggregation into develop
All threads resolved!

We currently squeeze multiple History columns into one by choosing the last column. This makes sense in many cases, but, as I got a request for alternative aggregation schemes from the IAGOS-project, obviously not in all. This MR implements the option to select from several squeezing options, currently "last" (out current implementation), "min" (select the smallest flag per row), "max" (select the highest flag per row).

As I didn't really find a good place, where this rather special behavior, that furthermore requires quite some insight into the saqc-machinery, should be controlled, it is now/currently set by a module global variable.

Any thoughts @palmb and @luenensc ?

Merge request reports

Pipeline #205152 passed

Pipeline passed for a5ab4192 on flag-aggregation

Test coverage 77.00% (0.00%) from 1 job

Merged by David SchäferDavid Schäfer 1 year ago (Mar 21, 2024 12:18pm UTC)

Loading

Pipeline #205434 passed

Pipeline passed for f861f59f on develop

Test coverage 77.00% (0.00%) from 1 job

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
    • Resolved by Bert Palm

      Shouldn't this be done in the Translators, than in the History?

      The History represents our internal state of a history. Since we dropped every relation between the columns/flags (remember we once had only write worst flags etc), there should be no need for another aggregation. Moreover the current implementation x.ffill(axis=1).iloc[:, -1] just return the last value or nan as or policies (i know this again) dictates.

      The squeeze already has a raw option to implement custom stuff like for example aggregations and i think within the translators is the right place for that.

      i currently dont see the need for this MR, but i'm open to listen.

      Edited by Bert Palm
  • Bert Palm
  • Bert Palm
  • David Schäfer added 1 commit

    added 1 commit

    • c91652f4 - Apply 1 suggestion(s) to 1 file(s)

    Compare with previous version

  • I would like to unblock this MR...

  • David Schäfer resolved all threads

    resolved all threads

  • Can I get an approval?

  • David Schäfer added 3 commits

    added 3 commits

    Compare with previous version

  • Peter Lünenschloß approved this merge request

    approved this merge request

  • merged

  • David Schäfer mentioned in commit f861f59f

    mentioned in commit f861f59f

  • Please register or sign in to reply
    Loading