mirror of
https://github.com/kuhyx/testsAndMisc.git
synced 2026-11-11 18:14:03 +01:00
Krzysztof kuhy Rudnicki
f58caea2ab
store.py mixed the writes that move a file through its lifecycle with the read-only projections over the same table. The reads move to _store_queries.py: claimable_values, ready_items, ready_item, counts, pending_downloads, known_md5s, state_of, attempts_for. claim_next goes with them despite being a write. It is the one write that has to be a single atomic transaction against exactly those reads, so splitting it from them would have left BEGIN IMMEDIATE on one side of a module boundary and its SELECT on the other. _scalar becomes scalar: record_files still counts rows before and after its insert, so store imports it back. The dependency runs one way only. store.py re-exports through __all__, so downloader.py, cli.py and test_store.py keep reaching everything at store.* unchanged -- all 15 names verified to resolve after ruff's autofix pass. store.py 382 -> 202, _store_queries.py 232. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011w2Pj4Ek3ZNMQVYcFHgXDY
36 lines
2.8 KiB
JSON
36 lines
2.8 KiB
JSON
{
|
|
"intent": "Bring python_pkg/wsg_grabber/store.py under the 250-line file cap by separating the read-only projections over the file index from the lifecycle writes, keeping every store.* name reachable for downloader.py, cli.py and test_store.py.",
|
|
"scope": [
|
|
"python_pkg/wsg_grabber/store.py",
|
|
"python_pkg/wsg_grabber/_store_queries.py (new)",
|
|
"Non-goal: no behaviour change, no SQL change, no test changes"
|
|
],
|
|
"changes": [
|
|
"Moved the read-only projections into _store_queries.py: claimable_values, ready_items, ready_item, counts, pending_downloads, known_md5s, state_of, attempts_for.",
|
|
"Moved claim_next there too -- it is a write, but it is the one write that must be a single atomic transaction against exactly those reads, so splitting it from them would have put a BEGIN IMMEDIATE on one side of a module boundary and its SELECT on the other.",
|
|
"Renamed the shared helper _scalar -> scalar because it is now imported across a module boundary; record_files still needs it (it counts rows before and after the insert), so store.py imports it back. The dependency runs store -> _store_queries only, no cycle.",
|
|
"store.py re-exports the moved names through __all__.",
|
|
"store.py 382 -> 202 lines; _store_queries.py 232. Both under the cap."
|
|
],
|
|
"verification": [
|
|
{
|
|
"command": "PYTHONPATH=. python3 -c \"from python_pkg.wsg_grabber import store; [getattr(store, n) for n in store.__all__]\"",
|
|
"result": "pass",
|
|
"evidence": "Printed 'store OK: 15'. Re-run after ruff's autofix pass removed 8 dead imports, confirming the __all__ entries were not among what it stripped."
|
|
},
|
|
{
|
|
"command": "pre-commit run --files python_pkg/wsg_grabber/store.py python_pkg/wsg_grabber/_store_queries.py",
|
|
"result": "pass",
|
|
"evidence": "After the autofix pass settled, the only remaining failure was the expected 'no staged evidence artifact' from this file not yet existing; ruff, ruff format, mypy, pylint and the coverage-gated pytest all pass. Final confirmation is the commit itself, which runs the full hook set including jscpd."
|
|
}
|
|
],
|
|
"risks": [
|
|
"An earlier attempt at this split was silently reverted: `git push --dry-run` ran pre-commit, which stashed unstaged files and restored the pre-edit store.py, leaving the new module orphaned. The split was redone and the files staged before running the gate. Stage before running pre-commit on a work-in-progress tree.",
|
|
"claim_next sits in a module named _store_queries despite being a write; the docstring says why. Splitting it from the reads it transacts against would be worse."
|
|
],
|
|
"rollback": [
|
|
"git revert the commit; store.py returns to its 382-line form and _store_queries.py is deleted.",
|
|
"After rollback, run python3 -m pytest python_pkg/wsg_grabber/tests/ -q to confirm the suite still passes."
|
|
]
|
|
}
|