Add missing nodes to various places in InferenceSystem

[dthaler]

Eventually these should all pull from https://live.orcasound.net/api/json/feeds but the first step is to try adding the missing ones manually.

[micya]

"Update the last line of the Dockerfile to point to the new config file"

In theory, each container should have its own dockerfile. In practice, the original owners used one dockerfile and manually changed the last line before executing docker build.

I do not know how to test this PR.

You should test the individual docker images by running docker run locally and verifying that the logs look as expected (does not crash).

I am not sure if the github CI/CD workflows exercise these scripts or not.

No CI/CD is currently setup for the inference system. We manually deploy to kubernetes using the instructions here.

Although we no longer deploy on Azure Container Instances, it is good practice to keep the ACI deployment config up to date (as you have done in this PR) in the unlikely chance that Azure fixes the stability issues and we move back to ACI.

[dthaler]

"Update the last line of the Dockerfile to point to the new config file"

In theory, each container should have its own dockerfile. In practice, the original owners used one dockerfile and manually changed the last line before executing docker build.

What's confusing in the README is that it implies that the PR should update the Dockerfile. But it sounds like that's not the case. It sounds like the README should be updated to say to not check in the Dockerfile changes, but to do it locally, once for each hydrophone being added, etc. per the explanation you gave in comments above.

[micya]

"Update the last line of the Dockerfile to point to the new config file"

In theory, each container should have its own dockerfile. In practice, the original owners used one dockerfile and manually changed the last line before executing docker build.

What's confusing in the README is that it implies that the PR should update the Dockerfile. But it sounds like that's not the case. It sounds like the README should be updated to say to not check in the Dockerfile changes, but to do it locally, once for each hydrophone being added, etc. per the explanation you gave in comments above.

Your understanding is correct. The proper fix is to add a dockerfile per location.

[dthaler]

You should test the individual docker images by running docker run locally

I tried following the instructions to do so but couldn't even get far enough to try. Filed issue #156 which is blocking my ability to test locally.

[micya]

I will approve once this PR is tested as noted in #151 (comment).

[dthaler]

I will approve once this PR is tested as noted in #151 (comment).

Testing is currently blocked on #156

[micya]

Obsoleted by #197.

[dthaler]

Obsoleted by #197.

Only partially. There are some things in this PR that are not in 197. I will rebase.

[dthaler]

Obsoleted by #197.

Only partially. There are some things in this PR that are not in 197. I will rebase.

I rebased. @micya Please re-review.

• Updated README.md so that line 179 need not be modified every time we add a hydrophone
• Updated hls_reader.py to add the commas since the checked-in file is invalid syntax (as reported by CodeQL) and to add the remaining streams
• Updated LiveInferenceOrchestrator.py to alphabetize the list, and also to correct the "name" of Orcasound Lab per issue Haro Strait vs Straight #145
• Updated PrepareDataForPredictionExplorer.py to add the missing steams
• Updated globals.py to alphabetize the list

[micya]

No behavior change. Redeployment not required.