LangChain Zunno Integration

A LangChain integration for Zunno LLM and Embeddings, providing easy-to-use wrappers for text generation and embeddings.

Installation

pip install langchain-zunno

Quick Start

Text Generation (LLM)

Basic Usage (Returns only response text)

from langchain_zunno import ZunnoLLM

# Create an LLM instance
llm = ZunnoLLM(model_name="mistral:latest")

# Generate text
response = llm.invoke("Hello, how are you?")
print(response)

Full Response Mode (Returns complete API response)

from langchain_zunno import ZunnoLLM

# Create an LLM instance with full response
llm = ZunnoLLM(
    model_name="mistral:latest",
    return_full_response=True
)

# Get complete API response
full_response = llm.invoke("Hello, how are you?")
print(full_response)
# Returns: {"response": "...", "model_used": "...", "tokens_used": 123, ...}

Embeddings

Basic Usage (Returns only embeddings vector)

from langchain_zunno import ZunnoLLMEmbeddings

# Create an embeddings instance
embeddings = ZunnoLLMEmbeddings(model_name="mistral:latest")

# Get embeddings for a single text
embedding = embeddings.embed_query("Hello, how are you?")
print(f"Embedding dimension: {len(embedding)}")

# Get embeddings for multiple texts
texts = ["Hello world", "How are you?", "Good morning"]
embeddings_list = embeddings.embed_documents(texts)
print(f"Number of embeddings: {len(embeddings_list)}")

Full Response Mode (Returns complete API response)

from langchain_zunno import ZunnoLLMEmbeddings

# Create an embeddings instance with full response
embeddings = ZunnoLLMEmbeddings(
    model_name="mistral:latest",
    return_full_response=True
)

# Get complete API response
full_response = embeddings.embed_query("Hello, how are you?")
print(full_response)
# Returns: {"embeddings": [...], "model_used": "...", "embedding_dimension": 4096, ...}

Async Usage

import asyncio
from langchain_zunno import ZunnoLLM, ZunnoLLMEmbeddings

async def main():
    # Async LLM
    llm = ZunnoLLM(model_name="mistral:latest")
    response = await llm.ainvoke("Hello, how are you?")
    print(response)
    
    # Async embeddings
    embeddings = ZunnoLLMEmbeddings(model_name="mistral:latest")
    embedding = await embeddings.aembed_query("Hello, how are you?")
    print(f"Embedding dimension: {len(embedding)}")

asyncio.run(main())

Factory Functions

For convenience, you can use factory functions to create instances:

from langchain_zunno import create_zunno_llm, create_zunno_embeddings

# Create LLM with full response
llm = create_zunno_llm(
    model_name="mistral:latest",
    temperature=0.7,
    max_tokens=100,
    return_full_response=True
)

# Create embeddings with full response
embeddings = create_zunno_embeddings(
    model_name="mistral:latest",
    return_full_response=True
)

Configuration

LLM Configuration

• model_name: The name of the model to use
• base_url: API endpoint (default: "http://15.206.124.44/v1/prompt-response")
• temperature: Controls randomness in generation (default: 0.7)
• max_tokens: Maximum number of tokens to generate (optional)
• timeout: Request timeout in seconds (default: 300)
• return_full_response: Return complete API response instead of just text (default: False)

Embeddings Configuration

• model_name: The name of the embedding model to use
• base_url: API endpoint (default: "http://15.206.124.44/v1/text-embeddings")
• timeout: Request timeout in seconds (default: 300)
• return_full_response: Return complete API response instead of just embeddings (default: False)

Response Modes

Basic Mode (Default)

• LLM: Returns only the generated text
• Embeddings: Returns only the embeddings vector

Full Response Mode

• LLM: Returns complete JSON response with all API fields
• Embeddings: Returns complete JSON response with all API fields

Example full response for LLM:

{
  "response": "Hello! I'm doing well, thank you for asking.",
  "model_used": "mistral:latest",
  "tokens_used": 15,
  "prompt_tokens": 5,
  "completion_tokens": 10,
  "total_tokens": 15
}

Example full response for embeddings:

{
  "embeddings": [0.123, -0.456, 0.789, ...],
  "model_used": "mistral:latest",
  "embedding_dimension": 4096,
  "normalized": true
}

API Endpoints

The package connects to the following Zunno API endpoints:

• Text Generation: http://15.206.124.44/v1/prompt-response
• Embeddings: http://15.206.124.44/v1/text-embeddings

Error Handling

The package includes comprehensive error handling:

try:
    response = llm.invoke("Hello")
except Exception as e:
    print(f"Error: {e}")

Development

Installation for Development

git clone https://github.com/zunno/langchain-zunno.git
cd langchain-zunno
pip install -e ".[dev]"

Running Tests

pytest

Code Formatting

black .
isort .

License

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

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests
5. Submit a pull request

Support

For support, please open an issue on GitHub or contact us at support@zunno.ai.