atomspace  by opencog

In-memory knowledge representation database with query engine and graph rewriting

Created 10 years ago
936 stars

Top 39.1% on SourcePulse

GitHubView on GitHub
Project Summary

The OpenCog AtomSpace is an in-memory hypergraph database and graph rewriting system designed as the central knowledge representation component for Artificial General Intelligence (AGI) systems. It offers a flexible and powerful alternative to traditional graph databases, enabling complex pattern matching, graph execution, and dynamic data flows for advanced AI research and development.

How It Works

The AtomSpace utilizes a "metagraph" structure, which is a more efficient and flexible representation than standard graphs. Its core innovation lies in treating queries themselves as graphs, allowing for meta-programming and rule-based systems. It supports advanced features like inverted searches, executable graphs (Atomese), typed graph elements, and mutable key-value stores attached to each graph element, enabling sophisticated data manipulation and processing pipelines.

Quick Start & Requirements

  • Install: cd opencog; mkdir build; cd build; cmake ..; make -j4; sudo make install
  • Prerequisites: Boost C++ utilities, CMake (v3.0.2+), Guile (v3.0+), cxxtest.
  • Optional: Cython, Haskell, OCaml, PostgreSQL.
  • OS: GNU/Linux (some BSDs may work).
  • Docs: AtomSpace, Atomese

Highlighted Details

  • Treats queries as first-class graphs for meta-programming.
  • Supports "Atomese," a language where graphs are executable abstract syntax trees.
  • Each graph element (Atom) can host a mutable key-value store for dynamic data.
  • Provides advanced pattern matching including unordered sets, choices, globs, and negation.

Maintenance & Community

The project is part of the larger OpenCog ecosystem, with active development and research focus. Community interaction is primarily through the OpenCog project channels.

Licensing & Compatibility

The README does not explicitly state a license. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The AtomSpace has a steep learning curve due to its unconventional design and APIs. It is primarily a research platform, and while mature, it is still under active development for advanced features like distributed processing and cross-system bridges. It does not support non-Linux operating systems.

Health Check
Last Commit

2 days ago

Responsiveness

1 day

Pull Requests (30d)
2
Issues (30d)
4
Star History
11 stars in the last 30 days

Explore Similar Projects

Starred by Chip Huyen Chip Huyen(Author of "AI Engineering", "Designing Machine Learning Systems") and Elvis Saravia Elvis Saravia(Founder of DAIR.AI).

llm-graph-builder by neo4j-labs

0.4%
4k
LLM app builds Neo4j graphs from unstructured data
Created 2 years ago
Updated 2 days ago
Feedback? Help us improve.