OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
RAG & Knowledge Systems

Autoschemakg

Autoschemakg is a Python framework for constructing knowledge graphs from unstructured text. It uses large language models to extract entity and event triples, induces a graph schema through conceptualization, and includes knowledge-graph completion. Its atlas-rag package also contains retrieval and vector-storage components.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Autoschemakg — OpenViST project profile in RAG & Knowledge Systems.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use AutoSchemaKG when a project needs a documented Python workflow for turning text into a schema-induced knowledge graph, or when exploring graph-based retrieval and multi-hop QA evaluation.

Who it is for

  • Developers building knowledge-graph construction pipelines from documents.
  • Researchers exploring schema induction, graph completion or multi-hop question-answering evaluation.
  • Teams preparing PDF, Markdown, JSON or multilingual text for graph extraction.

Documented uses

  • Extract entity and event relationships from unstructured text as graph triples.
  • Generate a knowledge-graph schema through conceptualization rather than supplying a predefined schema.
  • Apply graph completion and explore retrieval or vector-storage components.
  • Evaluate multi-hop question answering using the provided notebook and benchmark instructions.
Before you choose

Requirements & limits

The public GitHub repository lists an MIT license. The README documents installing the atlas-rag package with pip and gives a separate optional install for NV-embed-v2 support.

  • The documented construction example uses a large language model; the README shows a Transformers text-generation pipeline and a commented OpenAI client configuration.
  • NV-embed-v2 support calls for transformers version >=4.42.4 and
  • The repository separates core package modules, examples, evaluation directories and Neo4j utilities, so construction, evaluation and hosting are distinct parts of the workflow.
  • The README describes PDF input through a preparation route: convert PDF to Markdown, then Markdown to JSON for AutoSchemaKG processing.
  • The project provides notebooks for a full graph-construction pipeline, billion-scale graph usage and multi-hop QA evaluation. No specific operating-system guarantee is established.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install the atlas-rag package

Install AutoSchemaKG's atlas-rag package for knowledge-graph construction.

Before you start

  • Python environment capable of installing packages with pip.

Steps

  1. Run the command below.

    Documented command
    pip install atlas-rag

  2. If you need NV-embed-v2 support, install the optional package with the command below; this specifies a transformers version constraint of >=4.42.4 and <=4.47.1.

    Documented command
    pip install atlas-rag[nvembed]

Expected outcome

The atlas-rag package is installed; the optional installation includes the documented NV-embed-v2 support dependency.

Guide limitations

  • Installation alone does not construct a graph; the repository points to its full pipeline notebook for construction usage.
  • The NV-embed-v2 installation has a documented transformers version constraint.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
9 October 2026
Original website title
HKUST-KnowComp/AutoSchemaKG · Official homepage title · retained in its source language
How we review sourcesSuggest a correction