Proposal: merge the issue-66-steve-python3.12 branch into main

[saramsey]

By my count, the issue-66-steve-python3.12 branch has ten improvements vs. the main branch code:

1. Code cleanup in plover.py so it passes ruff, mypy, vulture, and most pylint checks
2. Code cleanup in main.py so it passes ruff, mypy, vulture, and most pylnit checks
3. The updated index-building code in plover.py is faster and uses less memory (see KG2.10.3c Plover is failing five unit tests #66 for benchmarks against the code in main)
4. Enhanced runtime configurability via the PLOVER_OTEL and PLOVER_DEBUG_SNAPSHOTS environment variables. Open Telemetry turned off by default.
5. Enhanced support for running PloverDB outside of Docker, on a development machine.
6. Updates so that PloverDB can host KG2.10.3 (which has a slightly changed schema from KG2.10.2).
7. Additional unit tests and some enhancements to existing unit tests.
8. Additional capabilities for memory profiling and debugging, so PloverDB can (during index build) compute and display object sizes (something we would only do in testing/debugging).
9. Improved logging messages in terms of clarity.
10. Improved PEP8 code formatting and modernized type hints.

[saramsey]

I have tested the code in this branch to build the KG2.10.2 graph (and verified that PloverDB can host it and pass its unit tests) and to build the KG2.10.3 graph (and verified that PloverDB can host it and pass its unit tests). I've also tested both values for the PLOVER_OTEL environment variable (i.e., true and false), to make sure PloverDB works as expected in both cases.

[saramsey]

OK, I did a full build of KG2.10.2/PloverDB using the latest code (untouched and untweaked!) from the issue-66-steve-python3.12 branch, on ploverdev8x.rtx.ai and tested it using unit tests. It passes all 38 unit tests in the issue-66-steve-python3.12 branch.

Further, here is confirmation that it is hosting KG2.10.2:

  "code_info": "HEAD: issue-66-steve-python3.12; Date: 2026-03-02",
  "endpoint_build_nodes": {
    "kg2c": {
      "biolink_version": "4.2.5",
      "category": "biolink:InformationContentEntity",
      "description": "This Plover build was done on 2026-03-02 18:30:50.972394 from input files 'https://kg2webhost.rtx.ai/kg2c-2.10.2-v1.0-nodes.jsonl.gz' and 'https://kg2webhost.rtx.ai/kg2c-2.10.2-v1.0-edges.jsonl.gz'.",
      "name": "Plover deployment of infores:rtx-kg2"
    }
  }
}

[saramsey]

I don't want to jinx it, but it looks like this branch's code can host KG2.10.2 or KG2.10.3, with no problems.

[saramsey]

Just to be on the safe side, I'm doing another test of a full build of the "fat" KG2.10.3/PloverDB (i.e., including the edges' publications_info property data) on the ploverdev8x.rtx.ai instance....

[saramsey]

OK, testing of the "fat" KG2.10.3 graph (including publications_info) using the PloverDB code in the issue-66-steve-python3.12 branch is complete. All unit tests passed.

[saramsey]

Might be worth checking whether TLS still works with the code and configs in the issue-66-steve-python3.12 branch. I would expect that it would, but have not yet tested it. I've only done my testing with http:// queries so far.