feat: complete catalog-driven preprocessing
Publish documentation / publish (push) Successful in 2m12s

This commit is contained in:
Codex
2026-09-06 17:49:35 +02:00
parent 8707ae1d46
commit cffa60772e
141 changed files with 5898 additions and 3015 deletions
+35 -5
View File
@@ -2,7 +2,7 @@ import { createHash } from "node:crypto";
import { normalize } from "node:path";
export interface CanonicalEffectiveConfig {
schemaVersion: 2;
schemaVersion: 3;
dwh: CanonicalDwhConfig;
vector: CanonicalVectorConfig;
embedding: CanonicalEmbeddingConfig;
@@ -21,7 +21,10 @@ export interface CanonicalDwhConfig {
}
export interface CanonicalVectorConfig {
collection: string;
collections: {
reference: string;
memory: string;
};
dimensions: number;
distance: string;
}
@@ -120,7 +123,10 @@ function buildVectorConfig(rendered: Record<string, unknown>): CanonicalVectorCo
if (!vector) {
throw new TypeError("effective config is missing vector resources");
}
const collection = requireString(vector, "collection");
const collections = asRecord(vector.collections);
if (!collections) {
throw new TypeError("effective config is missing vector collections");
}
const semanticIndex = asRecord(rendered.semantic_index);
const vectorStore = semanticIndex ? asRecord(semanticIndex.vector_store) : undefined;
const dimensions = vectorStore
@@ -129,7 +135,14 @@ function buildVectorConfig(rendered: Record<string, unknown>): CanonicalVectorCo
const distance = vectorStore
? requireString(vectorStore, "distance")
: (optionalString(vector, "distance") ?? "cosine");
return { collection, dimensions, distance };
return {
collections: {
reference: requireString(collections, "reference"),
memory: requireString(collections, "memory"),
},
dimensions,
distance,
};
}
function buildEmbeddingConfig(rendered: Record<string, unknown>): CanonicalEmbeddingConfig {
@@ -169,7 +182,7 @@ export function buildCanonicalEffectiveConfig(renderedConfig: unknown): Canonica
throw new TypeError("effective config requires a rendered configuration object");
}
return {
schemaVersion: 2,
schemaVersion: 3,
dwh: buildDwhConfig(rendered),
vector: buildVectorConfig(rendered),
embedding: buildEmbeddingConfig(rendered),
@@ -220,3 +233,20 @@ export function configFingerprint(renderedConfig: unknown): string {
export function inputFingerprint(workspaceId: string, renderedConfig: unknown): string {
return sha256(effectiveConfigIdentity(workspaceId, renderedConfig));
}
/**
* Fingerprint every non-Catalog input consumed by complete preprocessing. The immutable Git
* revision covers the workspace descriptor and its revision-pinned Evidence tree; the effective
* configuration identity covers the Catalog-derived DWH binding and semantic runtime contract.
*/
export function preprocessingInputFingerprint(
workspaceId: string,
workspaceRevision: string,
renderedConfig: unknown,
): string {
return sha256(JSON.stringify({
workspaceId,
workspaceRevision,
effectiveConfigIdentity: effectiveConfigIdentity(workspaceId, renderedConfig),
}));
}