Unoplat-CodeConfluence - The Universal Code Context Engine

🎯 Vision

Unoplat-CodeConfluence aims to be the definitive solution for extracting, understanding, and providing precise code context across repositories and domains. By combining deterministic code grammar with state-of-the-art LLM pipelines, we achieve human-like understanding of codebases in minutes rather than months. Our graph-based architecture ensures relationships and context are preserved at every level.

🌟 Why Unoplat-CodeConfluence?

• Deterministic Understanding: Built on ArchGuard and Tree-sitter for reliable, language-agnostic code parsing
• Smart Summarization: Bottom-up code analysis from functions to entire codebases, preserving context at every level
• Graph based Embedding: Embed codebases' functions using SOTA embeddings to enable semantic search and retrieval
• Enhanced Onboarding: Intuitive, interconnected documentation helps new team members understand complex codebases quickly
• Graph-Based Intelligence: Query and explore codebases through natural, graphical relationships between components
• Deep Dependency Insights: Comprehensive parsing of package managers as well as any other related metadata reveals true project structure and relationships.
• Integration Ready: Designed to work seamlessly with your existing development tools and workflows

🤝 For Knowledge Engine and AI Coding Assistants

Unoplat-CodeConfluence will provide:

• Precise Context API: Get reliable, deterministic code understanding through:
  • Bottom-up code summarization from functions to systems
  • Graph-based querying on SOTA embeddings with deterministic code grammar for natural code exploration.
  • Deep package and dependency analysis

🔍 OSS Atlas Initiative

Our UnoplatOssAtlas project is designed to dramatically accelerate contributor onboarding and productivity in open-source projects. By providing deep, contextual understanding of popular repositories, we will help developers in the following:

• Accelerate Onboarding: Understand complex codebases in minutes instead of months
• Boost Contribution Velocity: Make meaningful contributions faster with deep contextual insights
• Navigate Complex Systems: Easily understand dependencies, patterns, and architectural decisions
• Learn Best Practices: Study and adopt patterns from well-established open-source projects

This initiative demonstrates our commitment to:

• Empowering the open-source ecosystem by reducing barriers to contribution
• Showcasing practical applications of our context extraction capabilities
• Supporting sustainable open-source development through better understanding and reducing time to value for oss projects.

🚀 Getting Started

Ready to enhance your development workflow? Check out our Quick Start Guide We ran our code graph capability on Onyx for its python backend and here is result

Project Status

ALPHA

Programming Language Support

Language	In-POC	Alpha	Beta	Stable
Python	✓	✓	✓
Java	✓
TypeScript
Go

Tasks Status

Task	Research	POC	Released
Code Grammar	✓	✓	✓
Integration With Workflow Orchestrator	✓	✓	✓
Data Modelling for Code Grammar	✓	✓	✓
Insertion into Graph Database	✓	✓	✓
Data Modelling for Code Summarisation	✓	✓
SOTA LLM Pipelines For Code Grammar Summarisation/Reports	✓	✓
SOTA Embeddings on Codebase Functions	✓
Automatic Documentation	✓
GraphRag based Query Module
SDK for integration

Code Grammar

Python

Feature	Beta	Stable	Limitations
Package Parsing	✓
Package Metadata Parsing - Poetry and PIP and UV	✓
Inheritance	✓
Function Parsing	✓
Class Parsing	✓
Procedural Code Parsing	✓
Global Variable Parsing	✓
Function Call Parsing	✓		Certain function calls are not detected.
Class Variables/Instance Variables Parsing	✓		Not available per class/procedure. But available per function across functions.
Function Local Variable Parsing	✓
Function Return Type Parsing	✓		Return type is not captured properly.
Figuring out dependent internal classes	✓
Import Segregation	✓
Sorting functions within a class/procedure based on dependency	✓		Circular dependencies/recursion will not work as topological sort is used.

Note: For detailed limitations and resolutions that we are working towards refer to Limitations_Resolutions.md

Credits/heroes/supporters/Inspirations

These are the people/entities because of which this work has been possible. Unoplat code confluence would not exist without them.

1. Phodal from Chapi and ArcGuard
2. [Ishaan & Krrish from Litellm]([email protected] / [email protected])
3. Omar Khattab
4. Joao Moura from crewai
5. Vipin Shreyas Kumar
6. Danswer
7. Continue
8. OpenDevin
9. Devon
10. Apeksha
11. Argilla
12. Jeremy Howard

Maintainers

1. Jay Ghiya
   • Contact: [email protected]
2. Vipin Shreyas Kumar
   • Contact: [email protected]

Book a call with us - Cal Link

Community Channel

UnoplatCodeConfluence Discord Channel

X handle

x