Skip to main content

Graph retrieval

Project description

GraphRetrieval

GraphRetrieval is a Python library designed for advanced text retrieval and knowledge graph querying. It supports various models and techniques to enable efficient and accurate information retrieval from large text corpora and knowledge bases.

Installation

pip install -e git+https://github.com/jayavibhavnk/GraphRetrieval.git#egg=GraphRetrieval

or

pip install GraphRetrieval

Usage Setting Up Environment Variables

Before using the library, set up the necessary environment variables for Neo4j and OpenAI:

import os

os.environ["NEO4J_URI"] = "add your Neo4j URI here"
os.environ["NEO4J_USERNAME"] = "add your Neo4j username here"
os.environ["NEO4J_PASSWORD"] = "add your Neo4j password here"
os.environ['OPENAI_API_KEY'] = "add your OpenAI API key here"

GraphRAG

GraphRAG is used to create and query graphs based on text documents.

Example

import GraphRetrieval
from GraphRetrieval import GraphRAG

grag = GraphRAG()
grag.create_graph_from_file('add file path here')

# Query using the default A* search
print(grag.queryLLM("Ask your query here")) 

# Switch to greedy search
grag.retrieval_model = "greedy"
print(grag.queryLLM("Ask your query here"))

KnowledgeRAG

KnowledgeRAG integrates with a knowledge graph and supports hybrid searches combining structured and unstructured data.

Example

from GraphRetrieval import KnowledgeRAG
from langchain_community.graphs import Neo4jGraph

graph = Neo4jGraph()
gr = KnowledgeRAG()

# Initialize graph
gr.init_graph(graph)

# Create the graph chain
gchain = gr.graphChain()

# Query the graph chain
print(gchain.invoke({"question": "Ask your query here"}))

# Hybrid search using Neo4j vector index
gr.init_neo4j_vector_index()
gr.hybrid = True
print(gchain.invoke({"question": "Ask your query here"}))

Ingesting Data into Graph

Ingest large text data into the knowledge graph.

text = "enter text here"

from langchain_text_splitters import CharacterTextSplitter

text_splitter = CharacterTextSplitter(
    separator="

",
    chunk_size=1000,
    chunk_overlap=200,
    length_function=len,
    is_separator_regex=False,
)

docs1 = text_splitter.create_documents([text])
docs = gr.generate_graph_from_text(docs1)
gr.ingest_data_into_graph(docs)

gr.init_neo4j_vector_index()
print(gchain.invoke({"question": "Ask your query here"}))

Hybrid Search with GraphRetrieval and Knowledge Base

Combine GraphRAG and KnowledgeRAG for hybrid search.

gr.vector_index = grag
gr.hybrid = True
print(gchain.invoke({"question": "Ask your query here"}))

Image Graph RAG

Use directories of images for searching similar images.

image_graph_rag = ImageGraphRAG()
image_paths = image_graph_rag.create_graph_from_directory('/content/images')
similar_images = image_graph_rag.similarity_search('/content/images/car.jpg', k=5)

for doc in similar_images:
    print(doc.metadata["path"])
image_graph_rag.visualize_graph() # for graph visualization

Note: This is a new version without parallelization, use 0.1.5>= for parallelization.

Contributing

Contributions are welcome! Please submit a pull request or open an issue to discuss what you would like to change. License

This project is licensed under the MIT License. See the LICENSE file for details.

This README.md provides an overview of the GraphRetrieval library, installation instructions, and example usage scenarios, with the specified changes to the file path and environment variables sections.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

graphretrieval-0.1.8.tar.gz (8.6 kB view details)

Uploaded Source

Built Distribution

GraphRetrieval-0.1.8-py3-none-any.whl (8.8 kB view details)

Uploaded Python 3

File details

Details for the file graphretrieval-0.1.8.tar.gz.

File metadata

  • Download URL: graphretrieval-0.1.8.tar.gz
  • Upload date:
  • Size: 8.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.1.0 CPython/3.9.19

File hashes

Hashes for graphretrieval-0.1.8.tar.gz
Algorithm Hash digest
SHA256 1b0ac67f50da3b636a3a510c0395329d7f8ed16e2b5b22d71c0ccde665cf5469
MD5 d387969aca1827bbed39e6df8e577ac8
BLAKE2b-256 34bc219ccb263c27a4627201431393862061c2698dd20499fd2301f2613a5cd0

See more details on using hashes here.

File details

Details for the file GraphRetrieval-0.1.8-py3-none-any.whl.

File metadata

File hashes

Hashes for GraphRetrieval-0.1.8-py3-none-any.whl
Algorithm Hash digest
SHA256 0be06356d7e7a303137f77875d0f5a97328d7f7d820d17972e44c26dfa7f5047
MD5 775cba88445ba5b4e91cd7128c99bb63
BLAKE2b-256 2beaa06aa784eb3b9d66e17379185f387b6366a92fa7585012acd0f76f149010

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page