Knowledge Graph internals

The Knowledge Graph (KG) is the foundation of Aucert's test generation intelligence. This page covers the full internal architecture.

Data model (protobuf)

Defined in proto/knowledge-graph.proto:

message KGQuery {
  string customer_id = 1;
  string app_id = 2;
  string query_type = 3;
  map<string, string> filters = 4;
}

message KGResponse {
  repeated KGNode nodes = 1;
  repeated KGEdge edges = 2;
}

message KGNode {
  string id = 1;
  string type = 2;
  string name = 3;
  map<string, string> properties = 4;
}

message KGEdge {
  string source_id = 1;
  string target_id = 2;
  string relationship = 3;
}

Storage: PostgreSQL + JSONB

Per ADR-005, the KG uses PostgreSQL with JSONB columns rather than a dedicated graph database (Neo4j). Rationale:

• Neo4j is operationally heavy for under 20 customers
• PostgreSQL JSONB provides flexible schema with standard tooling
• Migration path to Neo4j exists if needed at scale

Ingestion pipeline

The ingestion engine processes multiple source types:

1. AST Ingestion — Parses application code into an abstract syntax tree, extracting screen definitions, navigation paths, and state management patterns
2. API Schema Ingestion — Reads OpenAPI/Protobuf specs to map endpoint relationships
3. UI Map Ingestion — Analyzes layout files and view hierarchies
4. Historical Data — Imports past test results and bug patterns

Self-healing (Phase 2)

The KG will eventually self-heal when app structure changes break existing graph nodes. When ingestion detects a renamed screen or removed endpoint, it updates the graph rather than creating orphaned nodes.

What's next

• 5-layer deep dive — Full pipeline architecture
• Verification Cascade — Decision-making process