feat(opt): three efficiency levers for NL→SQL workflow
Lever 1: Join-graph via FK logics in annotations + suggest-fks command
- TableAnnotation.foreign_keys field stores curated logical FKs (DWH has no FK constraints)
- tht schema suggest-fks: mine from approved SQL, heuristics (time_key → dim_time),
same-name discovery + explicit --assume flag for multi-owner PKs
- mschema renders 【Foreign keys】 section populated; validation in merge.py
- SKILL.md F4 now reads FKs from mschema-text, no custom data_time_key logic
Lever 2: Context-pack consolidation at kickoff (tht search pack)
- Single embedding of question, reused for schema + evidence + solved searches
- One command: tht search pack <question> --session <id> → retrieval_pack.md
- Graceful degradation when Ollama/vector store unreachable (exit 0, empty sections)
- SKILL.md F1 prescribes as first call; reduces model thinking turns via pre-retrieval
Lever 3: Phase-summary recap v2 auto-construction from session ledger
- tht session show --json includes full decisions ledger
- tht phase meta --json exports 'emits' (substantive decision types per phase)
- Gate appends deterministic 【Decisioni registrate in questa fase】 section (appendLedgerSection)
- Model authors only summary + checks; recap table comes from persisted state (exact by construction)
- SKILL.md Disciplina 6: brief model output, gate fills the rest
Tests: 358 Python (including 10 FK + 3 pack + 1 session-ledger tests) + 111 JS gate tests, all pass.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
@@ -151,10 +151,32 @@ function enrichCteResultColumns(payload, planTables, getColumns) {
|
||||
return { ...payload, columns };
|
||||
}
|
||||
|
||||
// Contract C bis. Appends a deterministic "decisions of this phase" section built
|
||||
// from the session ledger, so the model's recap can stay thin (Discipline 6 exactness
|
||||
// comes from persisted state, not model prose). `decisions` is session-show's ledger
|
||||
// dump; `emits` is the phase's substantive decision-type list from workflow.yaml.
|
||||
// No matching decisions -> data returned unchanged. Pure: returns a NEW object.
|
||||
function appendLedgerSection(data, decisions, emits) {
|
||||
const emitSet = new Set(emits || []);
|
||||
const rows = (decisions || []).filter((d) => d && emitSet.has(d.type));
|
||||
if (rows.length === 0) return data;
|
||||
const section = {
|
||||
title: "Decisioni registrate in questa fase (dal ledger)",
|
||||
items: rows.map((d) => ({
|
||||
label: `${d.type}: ${d.subject}`,
|
||||
value: d.detail || "",
|
||||
kind: "decision",
|
||||
rationale: d.rationale || "",
|
||||
})),
|
||||
};
|
||||
return { ...data, sections: [...(data.sections || []), section] };
|
||||
}
|
||||
|
||||
module.exports = {
|
||||
memoizeGetColumns,
|
||||
enrichCtePlanV2,
|
||||
buildCteResultV2,
|
||||
enrichCteResultColumns,
|
||||
enrichPhaseSummaryV2,
|
||||
appendLedgerSection,
|
||||
};
|
||||
|
||||
Reference in New Issue
Block a user