orchestrata_ultimate.knowledge_graph
Reference › orchestrata_ultimate › knowledge_graph
Knowledge Graph module for Orchestrata.
The knowledge_graph package makes the codebase and its surrounding work a queryable graph: files contain classes, classes define methods, functions call each other, specs trace to code, and runtime events mention the code they touched. Everything it writes goes to the operational knowledge graph, the orchkg database on the orchgraph engine (a Memgraph fork), reached over Bolt. The neo4j driver names that remain are Bolt-client naming only, since Neo4j is retired. The separate shared semantic/RAG graph, the memgraph database, belongs to the semantic_graph package; only semantic_bridge reads from both. The conversation capture path writes to a third, dedicated conversation graph.
The parts, by role:
- Identity and schema: canonical_schema (merge keys, DDL, conformance), schema, ontology_mapping, plus the types in types, types_enums and types_nodes.
- Access: graph_client, config, nodes, relationships and the interfaces protocols.
- Ingestion: sync for bulk AST JSON, synchronizer with file_watcher for incremental updates, and ast_analyzer with code_parser and the extractors for live analysis.
- Read side: query_engine, search, kg_context, and context injection in context_injector with the context_pack subpackage and graph_context_validator.
- Writers and records: runtime_writer, architecture_writer, kg_design, cxp_auditor, conversation_consumer and test_results_store.
- Governance and history: pam_integration, prolog_bridge, temporal, history, versioning, anomaly_detector and portfolio.
Source files enter through sync or synchronizer, land in orchkg under canonical ids, and are read back by query_engine, kg_context and the context injector when a volk starts a task. The one server entry point is the gRPC service in server. Start with canonical_schema, then sync, then context_injector; the story is on the knowledge_graph module page.
Curated narrative
The hand-written story of this package: orchestrata_ultimate.knowledge_graph.
55 modules, 21,436 lines of code.
Subpackages
| Package | Modules | LOC |
|---|---|---|
context_pack |
6 | 1,348 |
Modules
| Module | Summary | LOC |
|---|---|---|
__init__ |
Knowledge Graph module for Orchestrata. | 5 |
_sync_queries |
Read-side query mixin for GraphSyncer. | 283 |
adr_parser |
Pure text -> dataclass parser for the ADR corpus (wave ontology-phase2). | 505 |
anomaly_detector |
Anomaly detection and correction for the knowledge graph. | 515 |
architecture_writer |
The converged architecture-ontology writer (RATIFIED, wave ontology-phase1). | 1853 |
ast_analyzer |
AST analyzer for the knowledge graph. | 121 |
canonical_schema |
Canonical Knowledge-Graph schema — the single DDL source (SPEC-106 FR-001). | 1475 |
class_extractor |
Class extraction from Python AST. | 151 |
code_parser |
Code parser using Python's ast module. | 637 |
config |
Knowledge Graph runtime configuration. | 106 |
constants |
Knowledge Graph configuration constants. | 54 |
context_injector |
Knowledge Graph Context Injector — enriches CXP packets with graph context. | 1568 |
context_validator_loop |
Validator feedback loop. | 175 |
conversation_consumer |
Conversation-capture queue consumer (SPEC-021, US-021-007). | 712 |
cxp_auditor |
Spike Auditor — graph audit trail for CXP packet routing. | 291 |
dependency_analyzer |
Dependency analysis for the knowledge graph. | 193 |
exceptions |
Knowledge Graph exception hierarchy. | 59 |
file_watcher |
File watcher with debouncing for the knowledge graph. | 270 |
function_extractor |
Function extraction from Python AST. | 125 |
graph_client |
Graph client for the knowledge graph. | 453 |
graph_context_validator |
Post-LLM response validator for graph context compliance. | 427 |
graph_node_mixins |
Mixin classes for GraphClient node operations. | 314 |
history |
Change history for the knowledge graph. | 135 |
interfaces |
Knowledge Graph protocol interfaces. | 144 |
kg_context |
Knowledge Graph code context provider. | 906 |
kg_design |
Knowledge Graph design store for spike pipeline. | 924 |
nodes |
Graph node wrappers for the knowledge graph. | 164 |
ontology_mapping |
Graph → canonical-ontology mapping — full-fidelity, nothing discarded. | 456 |
pam_integration |
PAM++ integration for the knowledge graph. | 148 |
portfolio |
Portfolio management for the knowledge graph. | 128 |
project_resolver |
Project identity resolution for knowledge-graph operations. | 122 |
prolog_bridge |
Knowledge-specific Prolog bridge. | 143 |
query_engine |
Query engine for the knowledge graph. | 435 |
relationships |
Graph relationship wrappers for the knowledge graph. | 109 |
requirement_resolution |
Canonical requirement identity resolution. | 262 |
runtime_writer |
Runtime Tier-H signal writer. | 391 |
schema |
Knowledge Graph schema definition. | 435 |
search |
Search engine for the knowledge graph. | 175 |
semantic_bridge |
Semantic-search bridge: Memgraph chunks → Neo4j file roots. | 278 |
server |
Knowledge Graph gRPC server entry point. | 76 |
signal_extraction |
Extract file paths and code entity references from runtime payloads. | 75 |
sync |
GraphSyncer — sync serialized AST JSON into knowledge graph. | 1640 |
synchronizer |
File synchronizer for the knowledge graph. | 1085 |
temporal |
Temporal versioning for the knowledge graph. | 172 |
test_results_store |
Store pytest results in the graph for impact analysis. | 297 |
types |
Knowledge Graph type definitions. | 398 |
types_enums |
Knowledge Graph enumeration types and UTC helper. | 149 |
types_nodes |
Knowledge Graph node dataclass definitions. | 418 |
versioning |
Entity versioning for the knowledge graph. | 131 |
Dependency diagram
Truncated: showing 40 of 48 modules/subpackages (most connected first) and 5 of 5 outside packages.