refactor(shared-data): Disallow empty labware geometries

[SyntaxColoring]

Overview

In labware schema v3:

• A labware has wells (of course).
• A well can refer to a geometry definition, if we have detailed information about the shape of the well's interior.
• Each geometry definition has a list of vertically-stacked sections.

Some definitions were providing a well geometry, but with an empty list of sections. This was adding an edge case to the tests and production code. This PR forbids that, so geometries either need to have real data, or they need to be omitted.

Test Plan and Hands on Testing

Just automated tests.

Changelog

• In the JSON schema, require the sections array to have at least one element.
• In any labware definitions that were supplying an empty geometry, just delete the geometry.
  • Four of these are for "generic" labware. By definition, it doesn't seem like we could ever have detailed geometry info for these.
    • definitions/3/opentrons_24_aluminumblock_generic_2ml_screwcap/3.json
    • definitions/3/opentrons_24_tuberack_generic_0.75ml_snapcap_acrylic/2.json
    • definitions/3/opentrons_40_aluminumblock_eppendorf_24x2ml_safelock_snapcap_generic_16x0.2ml_pcr_strip/2.json
    • definitions/3/opentrons_96_aluminumblock_generic_pcr_strip_200ul/3.json
  • The remaining one is definitions/3/opentrons_24_tuberack_eppendorf_2ml_safelock_snapcap_acrylic/2.json. This labware is very old, apparently predating the 4-in-1 tube rack set. It does not appear in Labware Library and does not have geometry measurements in the "LLD GEOMETRY LIBRARY" sheet. I think it's safe to say we're never going to care about getting good geometry info for it. I think it's questionable whether a schema 3 definition should exist for it at all, let alone one with innerLabwareGeometry.

Review requests

Does my thinking in the changelog above sound correct?

Risk assessment

Low.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 25.88%. Comparing base (6546cd4) to head (0ae9ad6).
Report is 2 commits behind head on edge.

Additional details and impacted files

@@            Coverage Diff             @@
##             edge   #17580      +/-   ##
==========================================
+ Coverage   25.65%   25.88%   +0.23%     
==========================================
  Files        2840     2812      -28     
  Lines      218569   216598    -1971     
  Branches    17942    17878      -64     
==========================================
- Hits        56068    56063       -5     
+ Misses     162486   160520    -1966     
  Partials       15       15              

Flag	Coverage Δ
protocol-designer	18.86% <ø> (ø)
step-generation	4.37% <ø> (-0.01%)	⬇️

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

Files with missing lines	Coverage Δ
...pentrons_shared_data/labware/labware_definition.py	80.58% <ø> (-0.08%)	⬇️

... and 41 files with indirect coverage changes

[sfoster1]

Yeah, I think that makes perfect sense to me!