ai-spector-graph

Browser SDK for ai-spector traceability graphs. Read-only — load JSON, query subgraphs, run impact analysis.

Your backend fetches repo files and returns JSON. Your frontend passes that JSON into this SDK.

Install

npm install ai-spector-graph

Quick start

import { ProjectSession, parseImpactRules, DEFAULT_IMPACT_RULES } from "ai-spector-graph";

// 1. Fetch JSON from your API
const [graph, knowledge, registry] = await Promise.all([
  fetch("/api/files/traceability.graph.json").then((r) => r.json()),
  fetch("/api/files/knowledge.json").then((r) => (r.ok ? r.json() : null)),
  fetch("/api/files/section-registry.json").then((r) => (r.ok ? r.json() : null)),
]);

// 2. Build project session
const project = ProjectSession.fromBundle({
  graph,
  knowledge,
  registry,
  impactRules: DEFAULT_IMPACT_RULES,
});

// 3. Graph + knowledge + labels
project.graph.query("UC-01", { depth: 2 });
project.knowledgeCoverage();           // in-graph ✓/✗ per analyze row
project.sectionLabel("sec.srs.en.02"); // "2. Actors"

Documentation

| Guide | Description | |-------|-------------| | Integration guide | Start here — backend API, React setup, recipes, checklist | | API reference | GraphSession / ProjectSession methods | | Expansion roadmap | Future data sources |

JSON files your API should serve

| Repo file | Purpose | |-----------|---------| | .ai-spector/graph/traceability.graph.json | Required — main graph | | schemas/rules.impact.json | Impact analysis (or use DEFAULT_IMPACT_RULES) | | .ai-spector/.docflow/analysis/knowledge.json | Optional — analyze stats + in-graph coverage | | .ai-spector/registry/section-registry.json | Optional — section headings for UI | | .ai-spector/docflow.config.json | Optional — languages | | .ai-spector/.docflow/state.json | Optional — index/merge timestamps | | .ai-spector/.docflow/translation-queue/pending.json | Optional — translation dashboard |

No HTTP in this package

Fetch is your app's job. The SDK only accepts parsed JSON.