| Package |
|
| Code |
|
| Tools |
|
| CI/CD |
|
| Scans |
|
| Mentions |
|
| Badge |
|
Graphinate turns your data into Graphs. Effortlessly.
It is a Python library designed to streamline the pipeline from raw data to structured Graph representations. With Graphinate, you can easily map complex data hierarchies and payloads directly to nodes and edges, creating an efficient retrieval system.
Whether you need to visualize relationships, analyze structure, or serve data via an API, Graphinate provides the tools to bring your graph to life.
Graphinate uses and builds upon the excellent NetworkX.
Define the blueprint of your graph by decorating simple functions (sync or async) that yield your data as nodes and edges, and compose models via + or GraphModel.merge().
Traverses models and coordinates lifecycle events (on_start, on_node, on_edge, on_finish) across observers via synchronous (run()) or asynchronous (arun()) execution.
Takes your graph blueprint and constructs various outputs—a queryable NetworkX graph, a D3 dictionary, a Mermaid diagram, or an executable GraphQL schema (supporting both dynamic and static schema modes).
A suite of tools to visualize your graphs, from matplotlib plots to interactive Mermaid diagrams.
Instantly serve your graph data as an interactive GraphQL API with an embedded, 100% offline 3D WebGL graph viewer requiring zero external CDNs.
A handy command-line tool to manage and interact with your graph definitions without writing boilerplate code.
- Website (including documentation): https://erivlis.github.io/graphinate
- Source: https://github.com/erivlis/graphinate
- Package: https://pypi.org/project/graphinate
Graphinate is available on PyPI:
pip install graphinateor
uv add graphinateTo install with high-performance vectorized operations (NumPy progressive enhancement):
pip install "graphinate[perf]"To install with server and offline 3D WebGL viewer support:
pip install "graphinate[server]"or combine them:
pip install "graphinate[perf,server]"Graphinate officially supports Python >= 3.11.
import graphinate
N: int = 8
# First define a GraphModel instance to hold graph definitions
graph_model: graphinate.GraphModel = graphinate.model(name="Octagonal Graph")
# Register in the Graph Model the edges' supplier generator function
@graph_model.edge()
def edge():
for i in range(N):
yield {'source': i, 'target': i + 1}
yield {'source': N, 'target': 0}
# Use the NetworkX Builder
builder = graphinate.builders.NetworkxBuilder(graph_model)
# Build the NetworkX GraphRepresentation (an nx.Graph instance)
graph = builder.build()
# Display the graph with matplotlib
graphinate.matplotlib.plot(graph, with_edge_labels=True)
# Or create a Mermaid diagram
mermaid_builder = graphinate.builders.MermaidBuilder(graph_model)
diagram: str = mermaid_builder.build()
mermaid_markdown: str = graphinate.mermaid.markdown(diagram)
# Or create a Strawberry GraphQL schema and serve it
gql_builder = graphinate.builders.GraphQLBuilder(graph_model)
schema = gql_builder.build()
graphinate.graphql.server(schema)Graphinate natively supports asynchronous generator functions (async def ... yield) and coroutines (async def ... return [...]):
import asyncio
import graphinate
from graphinate.builders import NetworkxBuilder
m = graphinate.model(name="Async Mesh")
@m.node(type_="service")
async def services():
yield "auth-service"
yield "payment-service"
async def main():
# Asynchronously build across any builder
nx_graph = await graphinate.abuild(NetworkxBuilder, m)
print(nx_graph)
asyncio.run(main())For detailed information on the command-line interface, please see the CLI Usage Guide in the official documentation.
For instructions on how to set up your development environment, run tests, and contribute to the project, please see the Development Guide in the official documentation.
For a list of the dependencies and tools that make Graphinate possible, please see the Acknowledgements page in the official documentation.
Copyright © 2023-2026 Eran Rivlis