langchain-deepai 0.0.4

LangChain integration for DeepAI API with chat, images, and speech capabilities

LangChain DeepAI Integration

A comprehensive LangChain integration for the DeepAI API, providing seamless access to chat completions, image generation, text-to-speech, and speech-to-text capabilities.

Features

• 🗨️ Chat Completions: Multiple specialized models for different use cases
• 🎨 Image Generation: Text-to-image with various artistic styles
• 🔊 Text-to-Speech: High-quality speech synthesis
• 🎤 Speech-to-Text: Accurate audio transcription
• ⚡ LangChain Compatible: Full integration with LangChain ecosystem
• 🔄 Async Support: Both sync and async operations
• 📚 Well Documented: Comprehensive docstrings and examples

Installation

pip install langchain-deepai

For image processing capabilities:

pip install langchain-deepai[image]

For development:

pip install langchain-deepai[dev]

Quick Start

1. Set up your API key

export DEEPAI_API_KEY="your-deepai-api-key"

Or pass it directly:

from langchain_deepai import ChatDeepAI

chat = ChatDeepAI(api_key="your-deepai-api-key")

2. Basic Chat Example

from langchain_deepai import ChatDeepAI
from langchain_core.messages import HumanMessage

# Initialize chat model
chat = ChatDeepAI(
    model="standard",
    chat_style="chatgpt-alternative"
)

# Send a message
messages = [HumanMessage(content="Hello! How are you today?")]
response = chat.invoke(messages)
print(response.content)

3. Specialized Models

from langchain_deepai import DeepAIMath, DeepAICode

# Math-focused model
math_chat = DeepAIMath()
response = math_chat.invoke([HumanMessage(content="Solve: 2x + 5 = 15")])

# Code-focused model  
code_chat = DeepAICode()
response = code_chat.invoke([HumanMessage(content="Write a Python sorting function")])

4. Image Generation

from langchain_deepai import ImageDeepAI, generate_image

# Using the class
image_gen = ImageDeepAI()
result = image_gen.generate(
    prompt="A beautiful sunset over mountains",
    model="text2img",
    width=512,
    height=512
)

# Save the image
with open("sunset.jpg", "wb") as f:
    f.write(result["image_data"])

# Using the convenience function
result = generate_image(
    prompt="A cyberpunk cityscape at night",
    model="cyberpunk-generator"
)

5. Text-to-Speech

from langchain_deepai import TextToSpeechDeepAI

tts = TextToSpeechDeepAI()
result = tts.synthesize(
    text="Hello, this is a test of text-to-speech synthesis.",
    voice="default"
)

# Save audio file
with open("speech.wav", "wb") as f:
    f.write(result["audio_data"])

6. Speech-to-Text

from langchain_deepai import SpeechToTextDeepAI

stt = SpeechToTextDeepAI()
result = stt.transcribe("audio_file.wav")
print(result["text"])

Available Models and Styles

Chat Models

• standard: General purpose conversational AI
• math: Mathematics and problem-solving specialized
• online: Web-aware model with current information
• code: Programming and development focused

Chat Styles

• chatgpt-alternative: Default conversational style
• ai-code: Programming focused
• mathematics: Mathematical reasoning
• professional: Business communication
• creative: Creative and imaginative
• casual: Relaxed and informal
• goku: Enthusiastic character style
• gojo_9: Confident character style

Image Models

• text2img: Standard text-to-image generation
• fantasy-world-generator: Fantasy themed
• cyberpunk-generator: Cyberpunk aesthetic
• renaissance-painting-generator: Renaissance art style
• abstract-painting-generator: Abstract art
• impressionism-painting-generator: Impressionist style

Advanced Usage

Custom Chat Styles

from langchain_deepai import ChatDeepAI

# Professional business communication
business_chat = ChatDeepAI(
    model="standard",
    chat_style="professional"
)

# Creative writing assistant
creative_chat = ChatDeepAI(
    model="standard", 
    chat_style="creative"
)

Session Management

chat = ChatDeepAI()
chat.set_session_id("user-123")

# All messages in this session will maintain context
response1 = chat.invoke([HumanMessage(content="My name is Alice")])
response2 = chat.invoke([HumanMessage(content="What's my name?")])

Error Handling

from langchain_deepai import ChatDeepAI
import logging

# Enable logging
logging.basicConfig(level=logging.INFO)

try:
    chat = ChatDeepAI(api_key="invalid-key")
    response = chat.invoke([HumanMessage(content="Hello")])
except Exception as e:
    print(f"Error: {e}")

Batch Processing

from langchain_deepai import ImageDeepAI

image_gen = ImageDeepAI()

prompts = [
    "A serene lake at sunrise",
    "A bustling city street at night", 
    "A peaceful forest in autumn"
]

results = image_gen.generate_multiple(prompts, model="text2img")

Integration with LangChain

Chains

from langchain.chains import LLMChain
from langchain.prompts import PromptTemplate
from langchain_deepai import ChatDeepAI

# Create a chain
prompt = PromptTemplate(
    input_variables=["topic"],
    template="Write a short story about {topic}"
)

chain = LLMChain(
    llm=ChatDeepAI(chat_style="creative"),
    prompt=prompt
)

result = chain.run(topic="time travel")

Agents

from langchain.agents import initialize_agent, Tool
from langchain_deepai import ChatDeepAI, ImageDeepAI

def generate_image_tool(prompt: str) -> str:
    image_gen = ImageDeepAI()
    result = image_gen.generate(prompt)
    return f"Image generated and saved. URL: {result['image_url']}"

tools = [
    Tool(
        name="ImageGenerator",
        func=generate_image_tool,
        description="Generate images from text descriptions"
    )
]

agent = initialize_agent(
    tools=tools,
    llm=ChatDeepAI(),
    agent="zero-shot-react-description"
)

API Reference

ChatDeepAI

Main chat model class with full LangChain compatibility.

Parameters:

• model_name: Model to use ("standard", "math", "online", "code")
• chat_style: Conversation style
• api_key: DeepAI API key
• max_tokens: Maximum tokens to generate
• temperature: Response randomness
• session_id: Session identifier for context

Specialized Models

• DeepAIMath: Mathematics-focused chat model
• DeepAICode: Programming-focused chat model
• DeepAIOnline: Web-aware chat model
• DeepAIProfessional: Business communication model
• DeepAICreative: Creative writing model
• DeepAICasual: Casual conversation model

ImageDeepAI

Image generation with multiple artistic styles.

Methods:

• generate(): Generate single image
• generate_multiple(): Generate multiple images
• generate_and_save(): Generate and save to file

TextToSpeechDeepAI

High-quality text-to-speech synthesis.

Methods:

• synthesize(): Convert text to speech
• synthesize_and_save(): Convert and save audio
• synthesize_multiple(): Batch TTS processing

SpeechToTextDeepAI

Accurate speech-to-text transcription.

Methods:

• transcribe(): Transcribe audio to text
• transcribe_multiple(): Batch STT processing

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

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

Support

For support and questions:

• Check the documentation
• Open an issue
• DeepAI API documentation: https://deepai.org/docs