examples/

Details

Valid go.mod file
Redistributable license
Tagged version
Stable version
Learn more about best practices

Repository

github.com/liliang-cn/cortexdb

Links

Open Source Insights

README ¶

CortexDB Examples

The examples are organized by the current public architecture:

go run ./examples/01_core
go run ./examples/02_rag
go run ./examples/03_memoryflow
go run ./examples/04_knowledge_graph
go run ./examples/05_graphflow
go run ./examples/06_tools_mcp
go run ./examples/07_importflow
go run ./examples/08_self_knowledge_graph

01_core

Core embedded vector storage through pkg/cortexdb.

Opens a single-file DB
Uses Quick() for vector add/search
Uses collections for namespacing

Use this when you want the smallest CortexDB integration surface.

02_rag

High-level durable knowledge APIs through pkg/cortexdb.

SaveKnowledge
no-embedder lexical SearchKnowledge
entity/relation metadata for graph-aware retrieval

Use this for RAG and app knowledge storage.

03_memoryflow

Agent memory workflow through pkg/memoryflow.

transcript ingest
recall
layered wake-up context
diary write/read
transcript reconstruction
optional Hindsight recall strategy plugin

Use this for chat/session/agent memory.

04_knowledge_graph

Embedded RDF/KG workflow through pkg/cortexdb and pkg/graph.

RDF triples
SPARQL property paths and subqueries
incremental RDFS inference
SHACL-lite validation

Use this when you need raw knowledge graph semantics.

05_graphflow

Corpus-to-graph workflow through pkg/graphflow.

canonical extraction schema
deterministic build/analyze/report
graph.json, GRAPH_REPORT.md, and HTML export

Use this when you need to turn documents or model extraction output into an inspectable graph.

06_tools_mcp

Toolbox/MCP-aligned APIs through db.GraphRAGTools().

lists the available tool definitions
calls tools in-process with JSON payloads
demonstrates optional lexical semantic-router tool selection
shows the same shapes that can be exposed over MCP

Use this for agents and external LLM orchestration.

07_importflow

Structured-data import through pkg/importflow.

parses a CSV (or MySQL/PostgreSQL dump) source
builds a MappingPlan routing columns to RAG and knowledge-graph sinks
Plan/Run/AutoImport flow with optional AI-assisted mapping

Use this to bootstrap a CortexDB file from existing structured data.

08_self_knowledge_graph

Dogfooding demo: CortexDB maps itself.

splits docs/PROJECT_OVERVIEW.md into sections
extracts entities/relations with an LLM (local Ollama or any OpenAI-compatible endpoint via EXTRACT_BASE_URL/EXTRACT_API_KEY)
builds, analyzes, and exports the graph (JSON, report, HTML viz)
qa_test.go answers natural-language questions deterministically from graph edges

Use this as the end-to-end graphflow reference.

Rule of Thumb

Need raw vectors / collections?       -> 01_core
Need RAG knowledge storage/search?    -> 02_rag
Need chat/session memory?             -> 03_memoryflow
Need RDF/SPARQL/RDFS/SHACL?           -> 04_knowledge_graph
Need corpus-to-graph/report/export?   -> 05_graphflow
Need agent tool/MCP integration?      -> 06_tools_mcp
Need to import CSV / SQL dumps?       -> 07_importflow
Want the full graphflow E2E demo?     -> 08_self_knowledge_graph

Directories ¶

Path	Synopsis
01_core
02_rag
03_memoryflow
04_knowledge_graph
05_graphflow
06_tools_mcp
07_importflow
08_self_knowledge_graph Dogfooding demo: turn docs/PROJECT_OVERVIEW.md (CortexDB's own project overview) into a knowledge graph using CortexDB's graphflow workflow, with an LLM extractor.	Dogfooding demo: turn docs/PROJECT_OVERVIEW.md (CortexDB's own project overview) into a knowledge graph using CortexDB's graphflow workflow, with an LLM extractor.
09_connector Demo: desensitize a CSV through the connector privacy gate, then import to RAG.	Demo: desensitize a CSV through the connector privacy gate, then import to RAG.
10_support_brain Customer-support "agent brain" — an end-to-end CortexDB demo over a REAL database (Postgres or MySQL).	Customer-support "agent brain" — an end-to-end CortexDB demo over a REAL database (Postgres or MySQL).
11_unified_brain Unified support brain — a complex, multi-source CortexDB application.	Unified support brain — a complex, multi-source CortexDB application.
12_incident_agent Incident-analysis agent — a complex CortexDB example that REQUIRES an LLM (a chat/generation model, not an embedding model).	Incident-analysis agent — a complex CortexDB example that REQUIRES an LLM (a chat/generation model, not an embedding model).
13_scale_analytics Scale + analytics — a comprehensive, larger-volume CortexDB example.	Scale + analytics — a comprehensive, larger-volume CortexDB example.
14_semantic_rag Semantic RAG — CortexDB with a real embedding model (vector search), plus an LLM answer.	Semantic RAG — CortexDB with a real embedding model (vector search), plus an LLM answer.
kg_e2e Command kg_e2e is a runnable, fully-printed end-to-end walkthrough of the CortexDB RDF / Knowledge Graph stack through the public pkg/cortexdb facade.	Command kg_e2e is a runnable, fully-printed end-to-end walkthrough of the CortexDB RDF / Knowledge Graph stack through the public pkg/cortexdb facade.

?	: This menu
/	: Search site
f or F	: Jump to
y or Y	: Canonical URL