Currently adding a new hydrophone requires code changes in a bunch of different places. Currently hydrophone metadata is hard coded in:
- deploy\configmap.yaml
- demos\hls_reader.py
- deploy-aci.yaml
- deploy\*.yaml
- misc\make_tsv_from_json.py
- src\LiveInferenceOrchestrator.py
- src\globals.py
- src\model\datautils.py
- src\PrepareDataForPredictionExplorer.py
Orcasite already has the metadata and the danger is it could get out of sync if a database change is made without changing the InferenceSystem code. Instead of requiring code changes, the metadata can be dynamically looked up via the https://live.orcasound.net/api/json/feeds API, using the docker container namespace to find the entry with the matching slug. The NotificationSystem is already doing that, but the InferenceSystem is not.
Currently adding a new hydrophone requires code changes in a bunch of different places. Currently hydrophone metadata is hard coded in:
Orcasite already has the metadata and the danger is it could get out of sync if a database change is made without changing the InferenceSystem code. Instead of requiring code changes, the metadata can be dynamically looked up via the
https://live.orcasound.net/api/json/feedsAPI, using the docker container namespace to find the entry with the matching slug. The NotificationSystem is already doing that, but the InferenceSystem is not.