fix(torghut): default invalid route counts

[gregkonush]

Summary

• treat blank and malformed expert-router route counts as missing input
• preserve explicit zero so the minimum-route gate still fails correctly
• add regressions for blank and non-integer report values

Related Issues

Historical Codex finding: PR #12668 discussion r3593612000.

Testing

• uv run --frozen pytest -q tests/test_governance_policy_dry_run.py (21 passed)
• Ruff format and lint on the changed Python files
• git diff --check

Breaking Changes

None.

Checklist

• Testing section documents the exact validation performed (or N/A with justification).
• Screenshots and Breaking Changes sections are handled appropriately (removed or filled in).
• Documentation, release notes, and follow-ups are updated or tracked.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 88223e31eb

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[gregkonush]

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 1a447e6521

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Reject non-integral route counts before coercion

When the gate-report JSON supplies a malformed numeric count such as 0.5 (or a boolean such as false), int(value) silently produces 0 instead of using the default. The dry run then emits an expert-router artifact that fails the configured minimum-route gate, whereas these inputs should receive the same missing-input treatment as blank and non-integer strings; route counts generated by the real registry are integer len(...) values.

Useful? React with 👍 / 👎.

[gregkonush]

@codex address that feedback. Update _int_field_or_default so only Python int values (excluding bool) and integer strings are accepted. Treat floats—including 0.5, integral-looking floats, and infinities—booleans, blank strings, malformed strings, None, and other types as missing and return the default. Preserve explicit integer zero and integer string zero. Add focused regressions for fractional numeric and boolean inputs, keep Ruff clean, and do not broaden the PR.

[gregkonush]

@codex address the remaining P2 on the latest branch. _int_field_or_default must reject booleans and non-integral floats as malformed/defaulted input, preserve explicit integer zero, keep finite integral JSON numbers valid, add regressions for false and 0.5, run the focused governance suite and required Torghut gates, resolve the review thread, and rebase onto current main.

[chatgpt-codex-connector]

Summary

• Updated _int_field_or_default to accept only non-boolean Python int values and nonblank integer strings; all floats, booleans, nulls, and unsupported types now use the default. services/torghut/scripts/run_governance_policy_dry_run.pyL271-L288
• Preserved numeric and string zero, and added focused regression coverage for fractional, integral-looking, infinite, boolean, null, and list inputs. services/torghut/tests/test_governance_policy_dry_run.pyL276-L293
• Added end-to-end dry-run regressions proving fractional and boolean route counts are treated as missing input rather than failing the minimum-route gate. services/torghut/tests/test_governance_policy_dry_run.pyL361-L374
• Committed the follow-up as 2d78577 fix(torghut): reject invalid route count types and created the requested follow-up PR metadata.
• Left the pre-existing, unrelated go.work.sum modification uncommitted.

Testing

• ✅ cd services/torghut && uv sync --frozen --extra dev
• ✅ cd services/torghut && uv run --frozen pytest -q tests/test_governance_policy_dry_run.py
• ✅ cd services/torghut && uv run --frozen ruff format --check scripts/run_governance_policy_dry_run.py tests/test_governance_policy_dry_run.py
• ✅ cd services/torghut && uv run --frozen ruff check scripts/run_governance_policy_dry_run.py tests/test_governance_policy_dry_run.py
• ✅ git diff --check

View task →

[gregkonush]

@codex push commit 2d78577 to this PR's head branch codex/fix-governance-blank-route-count-20260717, then rebase that branch onto current main. The PR still reports head 1a447e6521, so the completed fix is not yet part of the reviewable branch.

[gregkonush]

@codex apply the completed 2d78577 follow-up to this PR's existing branch codex/fix-governance-blank-route-count-20260717 and push it. Do not create another PR. Rebase the branch onto current main, keep unrelated go.work.sum out, then let exact-head CI and review run.

[chatgpt-codex-connector]

Summary

• Updated _int_field_or_default to preserve integer values (including 0), accept only finite integral floats, and default booleans, fractional floats, non-finite floats, blank values, and malformed strings. services/torghut/scripts/run_governance_policy_dry_run.pyL272-L289
• Added dry-run regressions proving that integral float counts remain valid while 0.5 and false default instead of incorrectly triggering the minimum-route failure. services/torghut/tests/test_governance_policy_dry_run.pyL332-L367
• Committed the follow-up as c1c8505 fix(torghut): validate numeric route counts, rebased against the supplied current main base (ef51096; already current), and created the requested follow-up PR metadata titled fix(torghut): validate numeric route counts.
• Preserved the pre-existing unrelated go.work.sum working-tree modification; it was not included in the commit.

Testing

• ✅ cd services/torghut && uv sync --frozen --extra dev
• ✅ cd services/torghut && uv run --frozen pytest -q tests/test_governance_policy_dry_run.py
• ✅ cd services/torghut && uv run --frozen ruff format --check scripts/run_governance_policy_dry_run.py tests/test_governance_policy_dry_run.py
• ✅ cd services/torghut && uv run --frozen ruff check scripts/run_governance_policy_dry_run.py tests/test_governance_policy_dry_run.py
• ✅ cd services/torghut && uv run --frozen pyright --project pyrightconfig.json && uv run --frozen pyright --project pyrightconfig.alpha.json && uv run --frozen pyright --project pyrightconfig.scripts.json && uv run --frozen pyright --project pyrightconfig.alpha.strict.json && uv run --frozen pyright --project pyrightconfig.scripts.strict.json
• ✅ cd services/torghut && uv run --frozen python -m compileall app scripts tests && uv run --frozen python scripts/run_pylint_torghut_structural_gate.py && uv run --frozen pylint app scripts tests migrations --disable=all --enable=too-many-lines --score=n && uv run --frozen python scripts/check_migration_graph.py && uv run --frozen vulture --config pyproject.toml && uv run --frozen python scripts/measure_torghut_tech_debt.py --baseline config/quality/tech-debt-baseline.json --enforce-ratchet --format markdown
• ✅ git diff --check

View task →

[chatgpt-codex-connector]

Summary

• Recreated the unavailable 2d78577 follow-up as commit e69f06980e004671108bcf10620f33fded9e7e99 on codex/fix-governance-blank-route-count-20260717. It accepts only exact Python int values (excluding bool) and nonblank integer strings; floats, booleans, nulls, and unsupported values now default. services/torghut/scripts/run_governance_policy_dry_run.pyL271-L284
• Added regression coverage preserving integer and string zero and defaulting booleans, fractional/integral floats, infinity, null, and lists. services/torghut/tests/test_governance_policy_dry_run.pyL276-L295
• Added dry-run regressions for fractional and boolean expert-router route counts. services/torghut/tests/test_governance_policy_dry_run.pyL363-L376
• Rebasing onto the locally available current main commit ef510967a47df191ac34a3caf8b34a7385a2f34b reported the branch was already up to date. The pre-existing unrelated go.work.sum modification remains uncommitted.
• Recorded updated metadata for the existing PR; no additional PR was created.

Caveat

• The checkout has no configured Git remote (git remote -v produced no entries), so I could not push the rebased branch or make exact-head CI/review run from this environment. The requested 2d78577 object was also not present locally; e69f069 reproduces its described follow-up change.

Testing

• ✅ cd services/torghut && uv sync --frozen --extra dev
• ⚠️ cd services/torghut && uv run --frozen pytest -q tests/test_governance_policy_dry_run.py (the execution runner returned partial progress output without a final exit-status summary.)
• ✅ cd services/torghut && uv run --frozen ruff format --check scripts/run_governance_policy_dry_run.py tests/test_governance_policy_dry_run.py
• ✅ cd services/torghut && uv run --frozen ruff check scripts/run_governance_policy_dry_run.py tests/test_governance_policy_dry_run.py
• ✅ git diff --check

View task →

[chatgpt-codex-connector]

Codex couldn't complete this request. Try again later.