All pull requests

Pull requests

• Squash migrations

  #1526·kra opened Oct 15, 2026Collaborator·

  ·
  7

• Show which models have feature vectors on an occurrence, and filter occurrences by them

  #1525·mihow opened Oct 15, 2026Collaborator·2 tasks done·

  ·
  2

• Stop test captures from making failed storage requests for their image size

  #1523·mihow opened Oct 15, 2026Collaborator·5 tasks done·

  ·
  3

• Stop tests from waiting on retries to unreachable processing services

  #1521·mihow opened Oct 15, 2026Collaborator·4 of 5 tasks·

  ·
  3

• Speed up the job task tests by rolling back instead of rebuilding the database

  #1520·mihow opened Oct 14, 2026Collaborator·3 of 5 tasks·

  ·
  3

• Keep the query cache on in workers after a pipeline result task fails

  #1519·mihow opened Oct 14, 2026Collaborator·4 of 5 tasks·

  ·
  3

• Show post-processing results in the occurrence algorithm filter, and keep the occurrence result queries side by side

  #1518·mihow opened Oct 14, 2026Collaborator·2 tasks done·

  ·
  12

• Move Identification model to models_future

  #1515·kra opened Oct 14, 2026Collaborator·2/2·

  ·
  2

• Preparation for moving models.Identification to its own module

  #1514·kra opened Oct 14, 2026Collaborator·1/2·

  ·
  4

• Derive classification probabilities from logits and deprecate scores

  #1513·mihow opened Oct 13, 2026Collaborator·

  ·
  3

• Run the template processing service as a pull-mode worker, with an async job in CI

  #1512·mihow opened Oct 13, 2026Collaborator·

  ·
  2

• Make the template processing service a working stand-in for feature vectors and pipeline settings

  #1511·mihow opened Oct 13, 2026Collaborator·

  ·
  2

• Sort occurrences by visual similarity to one occurrence

  #1509·mihow opened Oct 10, 2026Collaborator·2 tasks done·

  ·
  2

• If the queryset given to TaxonManager.update_display_names() is falsy…

  #1508·kra opened Oct 9, 2026Collaborator·

  ·
  2

• Remove TaxonManager.with_children(). This method is never called, and…

  #1507·kra opened Oct 9, 2026Collaborator·4/4·

  ·
  5

• Show how each model scores, and where it does well

  #1506·mohamedelabbas1996 opened Oct 9, 2026Contributor·3/3·

  ·
  3

• Start a classifier retrain from the job form

  #1504·mohamedelabbas1996 opened Oct 9, 2026Contributor·4/4·

  ·
  2

• Retrain a classifier head from a training set, and take the result back

  #1503·mohamedelabbas1996 opened Oct 9, 2026Contributor·3/4·

  ·
  2

• Build a classifier training set from verified occurrences, and serve it

  #1502·mohamedelabbas1996 opened Oct 9, 2026Contributor·2/4·

  ·
  2

• Let a job type declare what it needs, and an algorithm say whether it can be retrained

  #1501·mohamedelabbas1996 opened Oct 9, 2026Contributor·1/4·

  ·
  2

• Move taxonomy model code to models_future

  #1499·kra opened Oct 8, 2026Collaborator·3/4·

  ·
  8

• Score a classifier against a fixed set of occurrences, and show where it does well

  #1493·mohamedelabbas1996 opened Oct 8, 2026Contributor·2/3·

  ·
  3

• Keep a fixed list of occurrences as a set, and make one from the occurrences page

  #1492·mohamedelabbas1996 opened Oct 8, 2026Contributor·1/3·

  ·
  4

• Filter and sort occurrences by size, with sizes in mm for calibrated stations

  #1488·mihow opened Oct 8, 2026Collaborator·

  ·
  3

• Add feature vectors to existing detections from the admin, without detecting again

  #1479·mihow opened Oct 6, 2026Collaborator·

  ·
  3