fix(app-testing): only specify labware that is used when running from tests

[DerekMaggio]

Overview

Fixes https://opentrons.atlassian.net/browse/EXEC-582

I ran into an issue when adding PL protocols to snapshot testing.
We were passing all custom labware in the labware folder to python -m opentrons.cli.analyze. Subsequently, when you add another custom labware, every protocol has the custom labware added to its command, meaning it ends up in the analysis. This causes every snapshot to need to be updated when you add a new custom labware.

This PR slightly modifies how we build the analyze command when we use generate_analyses_from_test. Instead of passing all of the custom labware, we only pass the custom labware that the Protocol object specifies. Doing this ensures that adding a new custom labware does not every protocol to change.

NOTE: The functionality of analyze_against_image is maintained as passing all the custom labware.

Test Plan

• Verify bug
  • Run snapshots and make sure they all pass
  • Without any changes to the logic in generate_analyses.py, add new custom labware to the labware folder. The added custom labware should not be specified in any of the ran protocols
  • Run snapshot tests and verify that they all fail because new labware was added to analysis output
• Remove previously added custom labware
• Verify fix
  • Add changes to generate_analyses.py
  • Run snapshot tests.
  • Verify all tests fail ONLY because any extraneous labware has been removed.
  • Verify that any protocols that use custom labware still have them in their analyses
  • Update snapshots on branch
  • Add the same labware that you used to verify the bug
  • Rerun the snapshot tests against the new snapshots and confirm that all tests pass (are not changed by adding that labware)

Review requests

None

Risk assessment

Low