Skip to content

Document complete search-result scoring and outputs - #1230

Merged
ebhills merged 1 commit into
mainfrom
codex/document-search-scoring
Oct 11, 2026
Merged

ebhills merged 1 commit into
mainfrom
codex/document-search-scoring

Conversation

@ebhills

@ebhills ebhills commented Oct 11, 2026 •

Copy link
Copy Markdown
Collaborator

Linked issue

Refs #1227 — release-planning context: PY 1.20.5; this PR does not close the release-planning issue.

Documentation follow-up to merged #1229.

What changes

Expand docs/search_result_scoring.md from a part-code evidence guide into a complete description of compute.score_search_results, using section headings, tables, recipe examples, and worked outputs rather than an implementation walkthrough.

  • Explain the ordered inputs, URL deduplication, seven additive scoring components and their defaults, scoring categories, eligibility filters, blacklist behavior, and result ordering.
  • Describe the complete structured and formatted outputs, including pricing/provenance, scoring details, metadata, reduced part-code evidence, and the pre-reduction evidence count.
  • Distinguish numeric scoring categories from diagnostic match levels, recipe outputs from direct-Python outputs, and configurable settings from fixed heuristics.
  • Update the README link and the core function's explanatory docstring.
  • Bring the wrangle's generated schema descriptions and defaults into line with the supported input/output contract: 3–5 input column names and either one output name or a one-/two-name output list.
  • Add regression checks for schema/signature alignment, supported configuration shapes, and the guide's executable recipe, structured output, and formatted summary.

The workstream is document-search-scoring. The scoring implementation, numeric defaults, result field names, and matching/reduction behavior are unchanged.

How it was verified

Local validation on this branch:

  • Focused credential-free scoring tests: .venv/Scripts/python.exe -m pytest -c pytest-local.ini tests/recipes/wrangles/test_compute.py -q --tb=short — 36 passed.
  • Full credential-free suite via scripts/test-local.ps1 — 3,023 passed, 6 skipped, 139 deselected on Windows / Python 3.13.1 (180.29 seconds).
  • scripts/check_pytest_local_config.py — passed; 67 test files, 8 ignores, and 22 deselects checked.
  • schema/generate_recipe_schema.py — schema generation and Draft7 validation passed; generated JSON is not committed.
  • Six additional offline synthetic checks covered category/evidence differences, eligibility flags, duplicate replacement, accepted/filtered ordering, and string/list blacklist normalization.
  • The documentation test reads the examples directly from the guide and verifies them against the recipe's actual result.
  • Refreshed origin/main: zero commits behind, one focused commit ahead; git diff --check and git merge-tree --write-tree origin/main HEAD passed. Only five intended files are changed.

GitHub Actions CI run 38166137004 is in progress for commit 5392ec98; no CI pass is claimed yet. No live provider, WrangleWorks-service, deployment, or package-publication checks were performed.

Compatibility and risk

This is documentation, schema metadata, and offline test work; the scoring algorithm and runtime signatures/defaults are unchanged. Supported recipes keep their existing mappings and result shapes.

Generated schema validation now explicitly rejects input/output column counts outside the documented contract and non-string column names. Consumers relying on unsupported configuration shapes may therefore see validation errors. The guide also makes existing heuristic limitations visible; it does not change them.

No new dependencies, credentials, version bump, deployment, or release publication are included. Rollback is a revert of this focused documentation/schema/test commit; no data migration is required.

Ready-for-review checklist

  • One human delivery owner is assigned: ebhills
  • The linked issue and intended milestone are correct: Release Planning for PY 1.20.5 #1227 / v1.20, targeting PY 1.20.5
  • The branch is current with main and has no merge conflicts
  • Focused tests pass
  • New or changed behavior has direct test coverage
  • Documentation/schema/configuration is updated where applicable
  • The PR contains no unrelated changes
  • The PR description reflects the branch's current scope and latest validation
  • Required GitHub Actions checks pass
  • One primary reviewer is requested only when this PR is ready

Draft handoff: local work and validation are complete. The repository currently has five Ready PRs, so keep this PR Draft while CI is pending and the Ready queue is full. Once checks pass and a slot is available, the assignee should mark it Ready and request one primary reviewer. Opening this PR does not authorize merging.

See the pull request workflow.

@ebhills ebhills added this to the v1.20 milestone Oct 11, 2026 — with ChatGPT Codex Connector
@ebhills ebhills self-assigned this Oct 11, 2026
@ebhills ebhills mentioned this pull request Oct 11, 2026
16 of 40 tasks
@ebhills
ebhills marked this pull request as ready for review October 11, 2026 19:12
@ebhills
ebhills merged commit 8097d8d into main Oct 11, 2026
15 checks passed
@ebhills
ebhills deleted the codex/document-search-scoring branch October 11, 2026 19:13
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant