processing error using distance and species range as constraints

[gladwinj]

{
"action": "runJob",
"response": {
"uuid": "778a8433-1841-49df-bceb-5eeec8f4a020",
"job": "generate_scores",
"created": "2025-07-24T14:52:20.720463-07:00",
"status": "failure",
"outputs": "{}"
}
}

[klinkow]

This has been observed on production when a distance constraint is defined before a species constraint, but not the other way around. Here are the Sentry details: https://conservation-biology-institute.sentry.io/issues/6768812377?project=5558587

This error does not occur in a development environment.

[klinkow]

The species constraints data on production is impossible to replicate. On production, toggling the constraints contextual layer shows a styling different from what would result from using publish_netcdf command. I'm not sure what the source files were for constraints data on production.

The only differences between constraints.py on the development vs master branches are formatting changes. This suggests code changes aren't responsible for this working in development, but failing on production. However, it's possible that code changes to ncdjango and trefoil could potentially be responsible for this discrepancy in functionality.

My guess is that the constraints data is the reason this is failing on production, but working in development. I'm also guessing that it shouldn't fail on production - that we should rewrite the code to handle a wider range of constraints data, including that currently being used on production. However, I'm not able to develop a fix without a copy of production constraints data (and/or the import process to publish it as a service if the process used on production makes the data or how it's served substantively different from using publish_netcdf).

@nikmolnar , thoughts?

[nikmolnar]

On production, toggling the constraints contextual layer shows a styling different from what would result from using publish_netcdf command. I'm not sure what the source files were for constraints data on production.

The only difference is styling. I think we just ran publish_netcdf, then edited the styling in the database.

[nikmolnar]

This has been observed on production when a distance constraint is defined before a species constraint, but not the other way around. Here are the Sentry details: https://conservation-biology-institute.sentry.io/issues/6768812377?project=5558587

This error does not occur in a development environment.

Since using these two constraints in combination is a somewhat uncommon use case, let's wait until the new data & code are deployed and see if this is still a problem.

[briannafair]

Trying to replicate this on prod, but not seeing any issues. My steps:

1. Find Seedlots: Choose point
2. Climate scenarios default
3. Choose Distance constraint on Step 7, and set Range
4. Back to Step 5, choose Trait, then Select a species from dropdown
5. Step 6 is autofilled
6. "Run Tool"
7. Results as expected

I've tried different species, different distance ranges, etc. with no change in a successful run.

@nikmolnar Am I missing a testing scenario that results in this error?

[nikmolnar]

@briannafair The issue is with distance constraint plus species constraint, i.e., staying in the "Constraints" step, not going back to step 5. However, I realize now we can't reproduce this problem since we removed species constraints when we updated the climate data, until our collaborators get us updated species range datasets. Sorry for not thinking of that sooner!