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

Linearrag

Linearrag is a Python research codebase for constructing and using graph-based retrieval-augmented generation systems over large corpora. It documents relation-free graph construction using lightweight entity recognition and semantic linking, with a runnable example and configurable retrieval options.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Linearrag — 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 Linearrag when you want a codebase for prototyping or studying graph-based retrieval over a corpus and can supply its local models, dataset, and API configuration.

Who it is for

  • Researchers studying GraphRAG construction and retrieval.
  • Developers prototyping RAG systems over large document collections.
  • Practitioners comparing retrieval approaches and construction-time LLM token use.

Documented uses

  • Prepare a graph-based retrieval workflow using a dataset, embedding model, spaCy model, and LLM configuration.
  • Explore relation-free graph construction intended to avoid LLM token costs during graph construction.
  • Run the documented multi-hop dataset example and optionally select vectorized retrieval when a strong GPU is available.
Before you choose

Requirements & limits

The repository is public and lists a GPL-3.0 license.

  • The documented workflow requires local Python dependencies, a spaCy model, dataset files, an embedding model, and OpenAI API configuration.
  • The README prefers Python 3.9.
  • The optional vectorized retrieval mode is described for strong GPUs; BFS iteration is the stated alternative.
  • The medical dataset requires a separate scientific/biomedical spaCy model.

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

Run the documented 2wikimultihop example

Use the repository's example command with its stated dataset and model settings.

Before you start

  • The setup prerequisites described in the repository README
  • A configured spaCy model, embedding model, dataset, and LLM model
  • GPU acceleration is relevant only if enabling the optional vectorized retrieval mode

Steps

  1. Set SPACY_MODEL to en_core_web_trf and EMBEDDING_MODEL to model/all-mpnet-base-v2.

  2. Set DATASET_NAME to 2wikimultihop and LLM_MODEL to gpt-4o-mini.

  3. Set MAX_WORKERS to 16.

  4. Run python run.py with the documented --spacy_model, --embedding_model, --dataset_name, --llm_model, and --max_workers arguments.

  5. Enable --use_vectorized_retrieval only when using the optional GPU-oriented retrieval path; otherwise use the documented BFS iteration approach.

Expected outcome

The repository's example workflow is invoked with the listed dataset and model configuration.

Guide limitations

  • The README does not make the optional vectorized mode a requirement and recommends it only with a strong GPU.
  • The command and setup are documented instructions, not an execution or compatibility guarantee.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Retrieval-augmented applications

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
LinearragPrepare a graph-based retrieval workflow using a dataset, embedding model, spaCy model, and LLM configuration.Linux, macOS, WindowsThe repository is public and lists a GPL-3.0 license.
LangsearchCrawl website or local data for a knowledge base.CLI, Self-hosted, Server
Hybrid Search RagCombine document retrieval and language-model generation in a configurable workflow.
Evidence and transparency

Sources & review

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