Skip to content

Latest commit

 

History

History
36 lines (23 loc) · 2.06 KB

File metadata and controls

36 lines (23 loc) · 2.06 KB

CHANGELOG

All notable changes to CormorantCast will be documented here.


[2.4.1] - 2026-03-18

  • Hotfix for the pathogen bulletin parser choking on NACA-formatted PDF exports — turned out to be a charset issue that only showed up with certain regional authority templates (#1337)
  • Fixed withdrawal period tracker not advancing correctly when an antibiotic course was logged across a DST boundary (yes, really)
  • Minor fixes

[2.4.0] - 2026-02-03

  • Overhauled the outbreak prediction window algorithm to weight dissolved oxygen variance more aggressively during high-density stocking phases — early testing on a few partner farms suggests we're catching columnaris risk signals 18–24 hours earlier than before (#892)
  • Biosecurity audit trail now exports directly to the AQUAVETPLAN checklist format; should save certifiers a bunch of manual transcription work
  • Added support for pulling telemetry from YSI ProDSS sensors over the REST bridge (took way longer than it should have, their API docs are a mess)
  • Performance improvements

[2.3.2] - 2025-11-14

  • Stocking density graph was rendering blank on Safari when the date range crossed a month boundary — couldn't reproduce it for weeks, finally got a screen recording from a user (#441)
  • Tightened up the antibiotic withdrawal countdown notifications; they were sometimes firing a day late if you had multiple concurrent treatments running on different pond segments
  • Minor fixes

[2.3.0] - 2025-09-29

  • First pass at the regional pathogen bulletin ingestion feed — currently supports bulletins from CABI Aquaculture Compendium and a handful of state-level extension services, more sources coming once I figure out a sustainable parsing strategy
  • Certification paperwork templates updated to reflect the 2025 revision to FAO technical guideline 5.4; old templates still available if your authority hasn't caught up yet
  • Rewrote the telemetry ingest queue from scratch after it started dropping readings under sustained load from multi-site deployments; should be considerably more stable now (#881)