fastccg 0.2.0.post1

Fast, minimalist, multi-model terminal-based SDK for building, testing, and interacting with LLMs via cloud APIs.

FastCCG (Fast Conversational & Completion Gateway)

FastCCG is a simple, powerful, and developer-friendly Python library for interacting with Large Language Models (LLMs). It provides a clean, unified API to work with models from leading providers like OpenAI, Google, Anthropic, and Mistral, making it easy to build, test, and deploy AI-powered applications.

🚀 Key Features

• 🔄 Unified API: Switch between different LLM providers with minimal code changes
• ⚡ Async Support: Built-in asynchronous operations for high-performance applications
• 🧠 Retrieval-Augmented Generation (RAG): Build powerful Q&A systems over your own documents
• ✨ Text Embedding: Convert text into vector representations for semantic search
• 🌊 Streaming: Real-time response streaming for interactive experiences
• 💾 Session Management: Save and restore conversation history
• 🖥️ CLI Interface: Powerful command-line tools for quick testing and interaction
• 🔧 Easy Configuration: Chainable methods for clean, readable code
• 🛡️ Error Handling: Robust error handling with custom exceptions

🏗️ Supported Providers

Provider	Models	Status
OpenAI	GPT-4o, GPT-3.5 Turbo	✅ Fully Supported
Google	Gemini 1.5 Pro, Gemini 1.5 Flash	✅ Fully Supported
Mistral	Mistral Tiny, Small, Medium	✅ Fully Supported
Anthropic	Claude 3 Sonnet	✅ Fully Supported

📦 Installation

pip install fastccg

⚡ Quick Start

import fastccg
from fastccg.models.gpt import gpt_4o

# Add your API key
api_key = fastccg.add_openai_key("sk-...")

# Initialize the model
model = fastccg.init_model(gpt_4o, api_key=api_key)

# Ask a question
response = model.ask("What is the best thing about Large Language Models?")
print(response.content)

🖥️ CLI Usage

FastCCG comes with a powerful CLI for quick interactions:

# List available models
fastccg models

# Ask a single question
fastccg ask "What is the capital of France?" --model gpt_4o

# Start an interactive chat session
fastccg chat --model gpt_4o

🧠 Retrieval-Augmented Generation (RAG)

Build a powerful question-answering system over your own documents with just a few lines of code. FastCCG handles the complexity of embedding, indexing, and context retrieval for you.

import asyncio
import fastccg
from fastccg.models.gpt import gpt_4o
from fastccg.embedding.openai import text_embedding_3_small
from fastccg.rag import RAGModel

# 1. Setup API keys and models
api_key = fastccg.add_openai_key("sk-...")
llm = fastccg.init_model(gpt_4o, api_key=api_key)
embedder = OpenAIEmbedding(api_key=api_key)

# 2. Create and configure the RAG model
rag = RAGModel(llm=llm, embedder=embedder)

# 3. Index your documents
documents = {
    "doc1": "The sky is blue during a clear day.",
    "doc2": "The grass in the park is typically green."
}

# 4. Ask a question related to your documents
async def main():
    response = await rag.ask_async("What color is the sky?")
    print(response.content)
    # Expected output will be based on the indexed context

asyncio.run(main())

# 5. Save your knowledge base for later use
rag.save("my_knowledge.fcvs", pretty_print=True)

🔄 Advanced Features

Asynchronous Operations

import asyncio

async def main():
    # Run multiple prompts concurrently
    task1 = model.ask_async("What is the speed of light?")
    task2 = model.ask_async("What is the capital of Australia?")
    
    responses = await asyncio.gather(task1, task2)
    for response in responses:
        print(response.content)

asyncio.run(main())

Streaming Responses

async def stream_example():
    async for chunk in model.ask_stream("Tell me a story"):
        print(chunk.content, end="", flush=True)

asyncio.run(stream_example())

Session Management

# Save conversation
model.save("my_session.json")

# Load conversation later
loaded_model = fastccg.load_model("my_session.json", api_key=api_key)

📚 Documentation

Comprehensive documentation is available in the docs/ directory:

• Quick Start Guide - Get up and running in minutes
• CLI Usage - Command-line interface guide
• FCVS CLI Tool - Manage .fcvs vector store files
• Embedding and RAG - Guides for embedding and RAG
• API Reference - Complete API documentation
• Supported Models - All available models and providers

🤝 Contributing

We welcome contributions! Please see our Contributing Guidelines for details.

📄 License

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

🌟 Why FastCCG?

• Developer Experience: Clean, intuitive API that just works
• Performance: Built with async-first architecture for scalable applications
• Flexibility: Easy to switch between providers and models
• Reliability: Comprehensive error handling and testing
• Community: Open source with active development and support

---

📖 Read the Full Documentation | 🚀 Get Started Now | 💬 Join the Discussion