ThatAIguyA world of ideas

THE CONSCIOUSNESS ATLAS
Read. Experiment. Follow the evidence.

ConsciousnessLearn · Explore · Joseph Anady

Janady13/missionmatch / db/queries/evaluations.sql

Source identity, revision, and recovered contents.

GitHub sourceSource text recovered

Source identity

FieldRecorded value
Source typeGitHub source
Original recordJanady13/missionmatch / db/queries/evaluations.sql
DateNot specified in source
RepositoryJanady13/missionmatch
Revision720d47fae476b0d38477ae4401697d372b1ea2bc
Branchfeature/opportunity-to-outcome
Path or archive memberdb/queries/evaluations.sql
Original SHA-256f452e792192065ba90452352e29e0620706e39a28b779100b737f1ee6f3206a7
Git blob SHA6cd31191ddfec0559f2deb67528a645647fec261
Reading-copy SHA-256f452e792192065ba90452352e29e0620706e39a28b779100b737f1ee6f3206a7

Original source ↗

Every recorded occurrence

These 2 records contain identical original source bytes. Their paths and revisions remain part of the archive.

Original recordBranch or archiveRevision / identifierOriginal location
Janady13/missionmatch / db/queries/evaluations.sqlfeature/opportunity-to-outcome720d47fae476b0d38477ae4401697d372b1ea2bcOpen original
Janady13/missionmatch / db/queries/evaluations.sqlfeature/aioe-upgrade2178e7298d0e5d9af83afa12402fd587f0ba6ed8Open original

Recovered source transcript

-- name: UpsertCandidateEvaluation :exec -- Upsert the latest screenMATCH qualification for one (candidate, requirement) pair. A fresh screen -- overwrites the prior row so there is exactly one current evaluation per pair, which missionMATCH -- reads instead of recomputing. evaluated_at is preserved as the first time the pair was screened; -- updated_at moves to now() on every rescreen. INSERT INTO candidate_evaluations ( organization_id, candidate_id, requirement_id, total_score, band, clearance_score, skills_score, certs_score, customer_score, experience_score, location_score, explanation, matched_skills, gap_summary, model_version, evaluated_by, weight_profile_version ) VALUES ($1,$2,$3,$4,$5,$6,$7,$8,$9,$10,$11,$12,$13,$14,$15,$16,$17) ON CONFLICT (organization_id, candidate_id, requirement_id) DO UPDATE SET total_score = EXCLUDED.total_score, band = EXCLUDED.band, clearance_score = EXCLUDED.clearance_score, skills_score = EXCLUDED.skills_score, certs_score = EXCLUDED.certs_score, customer_score = EXCLUDED.customer_score, experience_score = EXCLUDED.experience_score, location_score = EXCLUDED.location_score, explanation = EXCLUDED.explanation, matched_skills = EXCLUDED.matched_skills, gap_summary = EXCLUDED.gap_summary, model_version = EXCLUDED.model_version, evaluated_by = EXCLUDED.evaluated_by, weight_profile_version = EXCLUDED.weight_profile_version, updated_at = now(); -- name: GetCandidateEvaluation :one SELECT * FROM candidate_evaluations WHERE candidate_id = $1 AND requirement_id = $2; -- name: ListEvaluationsForRequirement :many -- The persisted screenMATCH qualifications for one requirement, joined to the candidate name and -- availability, best first. missionMATCH "who can support this mission today" reads this STORED -- artifact rather than rescore the pool. Both tables are RLS scoped, so the join cannot cross tenants. SELECT ce.candidate_id, ce.total_score, ce.band, ce.evaluated_at, c.full_name, c.available FROM candidate_evaluations ce JOIN candidates c ON c.organization_id = ce.organization_id AND c.id = ce.candidate_id WHERE ce.requirement_id = $1 ORDER BY ce.total_score DESC, ce.candidate_id;

Search equations, builds, manuscript text, and recovered sources.