Skip to content

Latest commit

 

History

History
61 lines (44 loc) · 2.39 KB

File metadata and controls

61 lines (44 loc) · 2.39 KB
summary Knowledge graph — semantic search, graph statistics, link analysis, and visualization
triggers
knowledge graph
search for
find backlinks
orphan documents
graph stats
context on_match

Datacortex Module

Purpose

Indexes Datacore knowledge bases (markdown, org-mode) and provides semantic search, link analysis, and interactive graph visualization. Shows how documents connect through wiki-links and tags. Includes temporal pulse snapshots for tracking knowledge growth over time.

Quick Start

Say "knowledge graph" to launch the interactive graph explorer.

How It Works

Indexing & Search

Reads from the shared zettel database (zettel_db.py). Provides full-text search, backlink discovery, orphan detection, and task queries across all indexed content.

Graph Visualization

D3.js force-directed graph with zoom/pan, node filtering (space, type, degree), timeline slider for pulse history, and minimap. Served via FastAPI on port 8765.

Pulse Snapshots

Timestamped graph snapshots for tracking how the knowledge base evolves. Generate, list, and diff pulses over time.

Agents & Commands

Name Type When to use
/datacortex command Interactive graph exploration with web UI
datacortex skill Knowledge graph queries and visualization
search tool Full-text search across indexed content
backlinks tool Find all documents linking to a file
orphans tool Find unlinked documents
stats tool Graph and database statistics
tasks tool Query tasks from indexed org files
patterns tool Query learning patterns
find_by_skill tool Find agents by skill keyword

Key Paths

Path Purpose
src/datacortex/ Python package (core, indexer, metrics, pulse, api, cli)
frontend/ D3.js graph visualization
config/datacortex.yaml Base configuration
config/datacortex.local.yaml User overrides (gitignored)
pulses/ Generated snapshots (gitignored)

Setup

cd /path/to/datacortex && pip install -e .

Reuses ~/.datacore/lib/zettel_db.py and zettel_processor.py -- no separate database setup needed.


This file covers structure, capability, and stable configuration. Learned behavior, user corrections, and operational preferences live as engrams -- call plur_recall_hybrid for those.