mptyl/ThothII
1
0
Fork 0
Code Issues 25 Pull Requests Actions Packages Projects Releases 1 Wiki Activity

fix(harness): recover schema table name typos

Browse Source
This commit is contained in:
marcopan
2026-07-20 17:53:24 +02:00
parent 3b52c8c08c
commit ec9b12dff4
2 changed files with 145 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files
harness/.pi/extensions/gate/__tests__/gate_schema_linking.test.js
+84 -9
View File
@@ -4,8 +4,9 @@
//
// HARNESS ADAPTATIONS (no existing test invokes a tool's execute() end-to-end):
// 1. tht-gate.js uses ESM `import { execFileSync } from "node:child_process"`. Under
// CommonJS `require(esm)` the named binding tracks the module namespace, so patching
// `child_process.execFileSync` BEFORE requiring tht-gate.js is observed by tht().
// CommonJS `require(esm)` the named binding is snapshotted at load time. A single
// delegating stub is installed ONCE before requiring tht-gate.js; each test swaps the
// inner handler via `_handler` so every stub is visible to the frozen ESM binding.
// 2. tht()'s loadEnvFromDotenv calls `require("node:fs")`; that `require` is undefined
// in the ESM module scope when loaded via require(esm) (in production Pi loads the
// extension as CommonJS, where require is native). We install a globalThis.require
@@ -33,12 +34,20 @@ const CATALOG = {
],
};
// Adaptation #1: the ESM import of execFileSync is frozen at require(GATE) time.
// A single delegating function is installed before the first load; each test swaps
// the inner _handler and clears it in finally — no save/restore of cp.execFileSync.
let _handler = null;
const _origExecFileSync = cp.execFileSync;
cp.execFileSync = function delegatingStub(file, args, opts) {
if (_handler) return _handler(file, args, opts);
return _origExecFileSync.call(cp, file, args, opts);
};
test("reviewer_schema_linking records table/column decisions and syncs schema_linking", async () => {
const calls = [];
const inputs = [];
// Adaptation #1: stub the shell BEFORE requiring tht-gate.js.
const origExecFileSync = cp.execFileSync;
cp.execFileSync = (file, args, opts) => {
_handler = (file, args, opts) => {
calls.push(args.join(" "));
inputs.push(opts?.input);
if (args[0] === "phase" && args[1] === "meta")
@@ -119,13 +128,79 @@ test("reviewer_schema_linking records table/column decisions and syncs schema_li
assert.match(result.content[0].text, /Schema linking registrato/);
} finally {
cp.execFileSync = origExecFileSync;
_handler = null;
}
});
test("reviewer_schema_linking auto-corrects typos in table names via fuzzy matching", async () => {
const calls = [];
const inputs = [];
_handler = (file, args, opts) => {
calls.push(args.join(" "));
inputs.push(opts?.input);
if (args[0] === "phase" && args[1] === "meta")
return JSON.stringify({ phases: [{ num: 4, id: "F4" }] });
if (args[0] === "phase" && args[1] === "show") return "Fase corrente: 4\n";
// The typo name fails:
if (args[0] === "schema" && args[1] === "columns" && args[2] === "dim_patientt") {
const err = new Error("Command failed");
err.stderr = "ERRORE: tabella 'dim_patientt' non trovata nel catalogo";
throw err;
}
// The correct name works:
if (args[0] === "schema" && args[1] === "columns" && args[2] === "dim_patient")
return JSON.stringify(CATALOG);
// catalogTableNames calls schema render:
if (args[0] === "schema" && args[1] === "render")
return "CREATE TABLE dim_patient (\n cod_paz bigint\n);\n";
return "";
};
try {
const gate = require(GATE);
const { createFakePi } = require("./fake_pi_runtime.js");
const { pi, ctx, tools } = createFakePi();
ctx.cwd = "/nonexistent-thothii-test-cwd-fuzzy"; // distinct key → fresh cache
gate.default(pi);
let capturedDescriptor = null;
ctx.ui.input = async (title) => {
capturedDescriptor = JSON.parse(title);
return JSON.stringify({
id: capturedDescriptor.id,
kind: "schema-linking",
tables: [{ id: "dim_patient", enacted: true, columns: ["cod_paz"] }],
});
};
const tool = tools.get("reviewer_schema_linking");
const result = await tool.def.execute(
"call-1",
{
session: "s1",
title: "Schema linking",
tables: [{ id: "t-pat", name: "dim_patientt", kind: "promote", suggested_columns: ["cod_paz"] }],
},
null,
null,
ctx,
);
// The gate should have auto-corrected "dim_patientt" → "dim_patient".
assert.equal(capturedDescriptor.tables[0].name, "dim_patient");
assert.match(result.content[0].text, /Schema linking registrato/);
// The persisted decisions use the corrected name.
const batchIdx = calls.findIndex((c) => c === "decision add-batch --session s1 --doc -");
assert.ok(batchIdx !== -1, "expected atomic decision batch");
const batch = JSON.parse(inputs[batchIdx]);
assert.equal(batch[0].subject, "dim_patient");
} finally {
_handler = null;
}
});
test("reviewer_schema_linking returns a textResult (not a throw) when a proposed table is not in the catalog", async () => {
const origExecFileSync = cp.execFileSync;
cp.execFileSync = (file, args) => {
_handler = (file, args) => {
if (args[0] === "phase" && args[1] === "meta")
return JSON.stringify({ phases: [{ num: 4, id: "F4" }] });
if (args[0] === "phase" && args[1] === "show") return "Fase corrente: 4\n";
@@ -167,6 +242,6 @@ test("reviewer_schema_linking returns a textResult (not a throw) when a proposed
assert.match(result.content[0].text, /badtable/);
} finally {
cp.execFileSync = origExecFileSync;
_handler = null;
}
});
harness/.pi/extensions/tht-gate.js
+61 -4
View File
@@ -367,6 +367,50 @@ function currentPhase(ctx, session) {
return m ? parseInt(m[1], 10) : 1;
}
// --- fuzzy table-name matching (prevents multi-turn debug loops on typos) ---
function levenshtein(a, b) {
if (a === b) return 0;
const la = a.length, lb = b.length;
if (!la) return lb;
if (!lb) return la;
let prev = Array.from({ length: lb + 1 }, (_, i) => i);
for (let i = 1; i <= la; i++) {
const cur = [i];
for (let j = 1; j <= lb; j++) {
cur[j] = Math.min(prev[j] + 1, cur[j - 1] + 1, prev[j - 1] + (a[i - 1] !== b[j - 1] ? 1 : 0));
}
prev = cur;
}
return prev[lb];
}
const _catalogNamesCache = new Map();
function catalogTableNames(ctx) {
const key = ctx.cwd ?? "";
if (_catalogNamesCache.has(key)) return _catalogNamesCache.get(key);
let names;
try {
const raw = tht(ctx, ["schema", "render", "--format", "mschema-text"]);
names = [...raw.matchAll(/^CREATE TABLE (\S+)/gm)].map((m) => m[1]);
} catch {
names = [];
}
_catalogNamesCache.set(key, names);
return names;
}
function closestTableName(ctx, name, maxDist = 3) {
const names = catalogTableNames(ctx);
let best = null, bestDist = maxDist + 1;
for (const n of names) {
if (Math.abs(n.length - name.length) > maxDist) continue;
const d = levenshtein(n, name);
if (d < bestDist) { bestDist = d; best = n; }
}
return best;
}
// tht phase advance --auto: exit 6 = not ready / needs human (silent no-op), others propagated.
// Used for the fire-and-forget auto-advance of the auto phases (F2 memory, F6 cte) after a
// reviewer_decide: it only advances when the phase is auto-eligible (zero substantive
@@ -978,10 +1022,23 @@ export default function (pi) {
try {
cat = JSON.parse(tht(ctx, ["schema", "columns", t.name, "--json"]));
} catch (e) {
const msg = (e.stderr || e.message || String(e)).toString().trim();
return textResult(
`Tabella '${t.name}' non caricabile dal catalogo (${msg}). Proponi solo tabelle presenti nel catalogo (usa 'tht schema render' / 'tht search' per verificarne i nomi).`,
);
// Fuzzy recovery: the model often misspells Italian table names
// (e.g. "abellazione" vs "ablazione"). Auto-correct if a close
// catalog match exists, preventing a multi-turn debug spiral.
const fix = closestTableName(ctx, t.name);
if (fix) {
try {
cat = JSON.parse(tht(ctx, ["schema", "columns", fix, "--json"]));
t.id = fix;
t.name = fix;
} catch { /* fall through to error */ }
}
if (!cat) {
const msg = (e.stderr || e.message || String(e)).toString().trim();
return textResult(
`Tabella '${t.name}' non caricabile dal catalogo (${msg}). Proponi solo tabelle presenti nel catalogo (usa 'tht schema render' / 'tht search' per verificarne i nomi).`,
);
}
}
if (!Array.isArray(cat.columns)) {
return textResult(`Catalogo per '${t.name}' non contiene colonne valide. Verifica con 'tht schema columns ${t.name} --json'.`);