SmartGraph
SmartGraph is a powerful Python library for building stateful, multi-component applications with Large Language Models (LLMs). Built on top of the ReactiveX for Python (reactivex) framework, SmartGraph provides a reactive, flexible, and maintainable system for creating complex data processing pipelines.
Think of SmartGraph as "Svelte for backend LLM components" - it offers a simple, reactive approach to building complex LLM-powered applications, similar to how Svelte simplifies frontend development.
Features
- Declarative and reactive framework for defining workflows
- Support for both simple linear and complex branching workflows
- Powerful state management capabilities
- Multi-component support with easy integration of LLMs and pre-built toolkits
- Compilation step for graph validation and runtime configuration
- Comprehensive error handling and logging
Installation
Install SmartGraph using pip:
pip install smartgraph
Quick Start
Here's an example of how to use SmartGraph to create a search assistant that can answer questions using web search results:
import asyncio
import os
from dotenv import load_dotenv
from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
from smartgraph import ReactiveSmartGraph
from smartgraph.components import CompletionComponent
from smartgraph.tools.duckduckgo_toolkit import DuckDuckGoToolkit
from smartgraph.logging import SmartGraphLogger
# Load environment variables
load_dotenv()
# Set up logging
logger = SmartGraphLogger.get_logger()
logger.set_level("INFO")
# Create a graph and pipeline
graph = ReactiveSmartGraph()
pipeline = graph.create_pipeline("SearchAssistant")
# Initialize DuckDuckGoToolkit
ddg_toolkit = DuckDuckGoToolkit()
# Add CompletionComponent with DuckDuckGoToolkit
completion = CompletionComponent(
name="SearchAssistant",
model="gpt-4o-mini",
api_key=os.getenv("OPENAI_API_KEY"),
toolkits=[ddg_toolkit],
system_context="You are a helpful assistant that can search the internet for information.",
)
pipeline.add_component(completion)
# Compile the graph
graph.compile()
# Create FastAPI app
app = FastAPI()
class SearchQuery(BaseModel):
query: str
@app.post("/search")
async def search(search_query: SearchQuery):
try:
result = await graph.execute("SearchAssistant", {"message": search_query.query})
return {"response": result.get("ai_response", "No response generated.")}
except Exception as e:
logger.error(f"An error occurred: {str(e)}")
raise HTTPException(status_code=500, detail="An error occurred during processing")
if __name__ == "__main__":
import uvicorn
uvicorn.run(app, host="0.0.0.0", port=8000)
This example sets up a FastAPI application with a /search endpoint. When a POST request is sent to this endpoint with a search query, SmartGraph processes the query using the DuckDuckGoToolkit to search the web and the CompletionComponent to generate a response based on the search results.
To run this example:
- Save the code in a file (e.g.,
main.py). - Set up your environment variables (OPENAI_API_KEY) in a
.envfile. - Install the required dependencies (
smartgraph,fastapi,uvicorn,python-dotenv). - Run the script:
python main.py - Send a POST request to
http://localhost:8000/searchwith a JSON body like{"query": "What is the capital of France?"}.
This example demonstrates how SmartGraph can be used to create a simple API for a search assistant, showcasing its ability to integrate different components and toolkits in a reactive pipeline.
Documentation
For more detailed information on how to use SmartGraph, please refer to our documentation.
Core Concepts
- ReactiveSmartGraph: The main class representing the entire graph structure.
- Pipeline: A sequence of connected components that process data.
- ReactiveComponent: The base class for all components in SmartGraph.
- CompletionComponent: A component for integrating Large Language Models.
- Toolkits: Pre-built components for common tasks like web searches and memory management.
Advanced Features
- Custom component creation
- Complex branching workflows
- Asynchronous API integration
- Caching and retry mechanisms
- Input validation
- Graph visualization
Contributing
We welcome contributions! Please see our contributing guide for details on how to get started.
License
SmartGraph is released under the MIT License. See the LICENSE file for more details.
Support
If you encounter any issues or have questions, please file an issue on the GitHub issue tracker.
Acknowledgements
SmartGraph is built on top of the excellent ReactiveX for Python library. We're grateful to the ReactiveX community for their work.
Metadata
Release files for smartgraph 0.2.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| smartgraph-0.2.0.tar.gz | 22.1 kB | Without provenance | Details for smartgraph-0.2.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| smartgraph-0.2.0-py3-none-any.whl | Python 3 | none | any | Details for smartgraph-0.2.0-py3-none-any.whl |
Total release size: 48.7 kB
Release files / smartgraph-0.2.0.tar.gz
| Download URL | smartgraph-0.2.0.tar.gz |
|---|---|
| Size | 22.1 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
c8c85f6eeff9db423f2f98a57a4c9ebcf2d3327c8e5d77da08843a7d2731ba1d
|
|
BLAKE2b-256 checksum How to use checksums |
fbb10bda93e911e29d5361a67971de2e5c7a1f1e12ba07e478d98da283c011c3
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.5.1 CPython/3.11.9 Linux/6.5.0-1025-azure
|
Release files / smartgraph-0.2.0-py3-none-any.whl
| Download URL | smartgraph-0.2.0-py3-none-any.whl |
|---|---|
| Size | 26.6 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
78979c35c1f096a5c229b8b3da4adf57384dde10553f64d8af9093644f3f68ff
|
|
BLAKE2b-256 checksum How to use checksums |
56c1ce64901cf771bed5975a50f14469b2e92860c80bb40a4f6c73713b666858
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.5.1 CPython/3.11.9 Linux/6.5.0-1025-azure
|
Security for smartgraph 0.2.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.