test(robot-server): Test compatibility with databases created in v6.0

[SyntaxColoring]

Overview

This adds a basic test for persistence backwards compatibility, which includes backwards compatibility in our SQL database. The test makes sure that the current robot-server code can read some resources that were created by a v6.0.1 robot-server.

Closes RSS-102.

This is a humongous diff—sorry! Most of it is opaque fixture data. The actual code changes are fairly small.

Changelog

• Copy a pre-populated persistence directory into the repository, intended to be used as an opaque test fixture. I created it by manually issuing HTTP requests to a v6.0.1 dev server and saving its persistence directory. See the added readme for details.

• Add a test that runs a dev server underpinned by that persistence directory, issues a bunch of HTTP GET requests to it, and makes sure none of the requests come back with an obvious error.

  For simplicity, we only check the responses in a very shallow way. Basically, we just check that the HTTP status is successful, and that the response data isn't empty. I believe this is sufficient to catch the bugs that we've been prone to so far.

Review requests

• Will these tests give us sufficient confidence to rework our persistence code according to the plan in RSS-98?
• Try making a change in robot-server, api, or shared-data that would break backwards compatibility. For example, try making an Optional field in a persisted Pydantic model non-Optional. Then, run make -C robot-server test—does it catch the problem?
  • They definitely catch the bug that was fixed by refactor(api): add opentrons.protocol_engine.types alias to PipetteName #11422, so that's cool.
• Do you have any ideas for future improvements? If so, please record them in RSS-100.

Risk assessment

No risk to production code. Changes are only to tests.

[SyntaxColoring]

This looked like a mistake in the fixture. Both pipettes were being loaded on the right mount, which caused simulation and run problems. Elsewhere in this file, within designerApplication, this pipette is said to be on the left mount, so I changed this to match.

[codecov]

Codecov Report

Merging #11448 (569754e) into edge (715b2a0) will decrease coverage by 0.08%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             edge   #11448      +/-   ##
==========================================
- Coverage   74.31%   74.22%   -0.09%     
==========================================
  Files        2072     2083      +11     
  Lines       57377    57468      +91     
  Branches     5539     5550      +11     
==========================================
+ Hits        42638    42655      +17     
- Misses      13482    13555      +73     
- Partials     1257     1258       +1     

Flag	Coverage Δ
app	74.20% <ø> (-0.44%)	⬇️
protocol-designer	45.81% <ø> (ø)
shared-data	86.44% <ø> (ø)
step-generation	88.39% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
app/src/organisms/ChangePipette/Instructions.tsx	76.92% <0.00%> (-0.86%)	⬇️
app/src/App/index.tsx	100.00% <0.00%> (ø)
app/src/molecules/Modal/index.tsx	100.00% <0.00%> (ø)
app/src/atoms/StatusLabel/index.tsx	100.00% <0.00%> (ø)
.../python/opentrons_shared_data/pipette/dev_types.py	100.00% <0.00%> (ø)
app/src/atoms/buttons/index.tsx
app/src/atoms/Toast/Toast.stories.tsx	0.00% <0.00%> (ø)
app/src/atoms/buttons/SubmitPrimaryButton.tsx	100.00% <0.00%> (ø)
app/src/organisms/CalibrationStatusCard/index.tsx	66.66% <0.00%> (ø)
...p/src/pages/Devices/CalibrationDashboard/index.tsx	80.00% <0.00%> (ø)
... and 10 more

[mcous]

A few little thoughts but this seems like a pretty vast improvement in our regression detection abilities for these APIs

[TamarZanzouri]

Awesome work Max! Thank you!