Workflow for installing in-progress integration seems broken

[tstirrat15]

I'm working on the SpiceDB integration in #2538. I'm trying to run my check locally with a setup that I previously got working with the openmetrics plugin: https://github.com/tstirrat15/spicedb-metrics-adventure/tree/main

The branch associated with the in-progress plugin is here: https://github.com/tstirrat15/spicedb-metrics-adventure/tree/spicedb-plugin

I'm setting things up by:

# Build the integration
ddev release build spicedb

# Copy the wheel into the compose workspace from above
cp ~/forks/datadog/integrations-extras/spicedb/dist/datadog_spicedb-1.0.0-py3-none-any.whl .

# Build the image
docker build -t datadog .

# Run the setup
docker compose up

I'm getting this log message:

datadog-1  | 2024-12-02 22:21:44 UTC | CORE | ERROR | (pkg/collector/scheduler.go:212 in getChecks) | Unable to load a check from instance of config 'spicedb': Python Check Loader: unable to import module 'spicedb': No module named 'spicedb'; Core Check Loader: Check spicedb not found in Catalog

and if I attempt to run the check directly:

Error: no valid check found

If I query the integration from within teh container I get:

~ agent integration show datadog-spicedb
Package datadog-spicedb:
Installed version: 1.0.0

I've verified that the package is installed on the agent and I can query it; I haven't been able to make heads or tails of the loader error message. I don't see where in the code it's deciding to look for a package named spicedb; I don't see where it's constructing the name of the requested package; I can't see where the installed package's module name would be defined.

Is there something obvious I'm missing?