LangChain Opper

A LangChain integration package for Opper AI that provides chat models and provider utilities with seamless LangGraph compatibility.

Features

• 🔗 LangChain Integration: Drop-in replacement for other LangChain chat models
• 📊 Built-in Tracing: Automatic span creation and management using Opper's tracing
• 📈 Metrics Support: Easy metric collection for evaluation and monitoring
• 🏗️ Structured Outputs: Native support for Pydantic models with with_structured_output()
• 🔧 Flexible Configuration: Customizable instructions, models, and task names
• 🚀 Easy Setup: Simple installation and configuration

Installation

pip install langchain-opperai

Setup

Set your Opper API key as an environment variable:

export OPPER_API_KEY="your_opper_api_key_here"

Quick Start

Basic Chat Model

from langchain_opperai import ChatOpperAI
from langchain_core.messages import HumanMessage

# Initialize the chat model
llm = ChatOpperAI(
    task_name="chat",
    model_name="anthropic/claude-3.5-sonnet",
    instructions="You are a helpful AI assistant.",
)

# Use it like any LangChain chat model
messages = [HumanMessage(content="What is the capital of France?")]
result = llm.invoke(messages)
print(result.content)  # "The capital of France is Paris."

Structured Output

from langchain_opperai import ChatOpperAI
from pydantic import BaseModel, Field

class Joke(BaseModel):
    setup: str = Field(description="The joke setup")
    punchline: str = Field(description="The joke punchline")

# Create a structured model
llm = ChatOpperAI()
structured_llm = llm.with_structured_output(Joke)

# Get structured output
joke = structured_llm.invoke("Tell me a joke about cats")
print(f"Setup: {joke.setup}")
print(f"Punchline: {joke.punchline}")

Using the Provider

from langchain_opperai import OpperProvider
from pydantic import BaseModel, Field

class Response(BaseModel):
    answer: str = Field(description="The answer")
    confidence: float = Field(description="Confidence score 0-1")

# Initialize provider
provider = OpperProvider()

# Create models
chat_model = provider.create_chat_model(
    task_name="chat",
    instructions="You are a helpful assistant.",
)

structured_model = provider.create_structured_model(
    task_name="structured_chat",
    instructions="Provide structured responses.",
    output_schema=Response,
)

# With tracing
trace_id = provider.start_trace("conversation", "User asks about Python")
result = chat_model.invoke("What is Python?")
provider.end_trace("Provided Python explanation")

LangGraph Integration

from langgraph.graph import StateGraph, END
from langgraph.graph.message import add_messages
from langchain_core.messages import HumanMessage
from langchain_opperai import OpperProvider
from typing import TypedDict, Annotated

class State(TypedDict):
    messages: Annotated[list, add_messages]

def chat_node(state: State):
    provider = OpperProvider()
    llm = provider.create_chat_model(
        task_name="langgraph_chat",
        instructions="You are a helpful assistant in a graph workflow.",
    )
    
    response = llm.invoke(state["messages"])
    return {"messages": [response]}

# Build graph
workflow = StateGraph(State)
workflow.add_node("chat", chat_node)
workflow.set_entry_point("chat")
workflow.add_edge("chat", END)

app = workflow.compile()

# Run
result = app.invoke({
    "messages": [HumanMessage(content="Hello!")]
})

API Reference

ChatOpperAI

A LangChain chat model that integrates with Opper AI.

Parameters:

• task_name (str): Name for the Opper task (default: "chat")
• model_name (str): Model to use (default: "anthropic/claude-3.5-sonnet")
• instructions (str): Instructions for the model
• opper_client (Optional[Opper]): Opper client instance
• parent_span_id (Optional[str]): Parent span ID for tracing

Methods:

• with_structured_output(schema): Create a model that returns structured output
• invoke(messages): Generate a response to the given messages
• ainvoke(messages): Async version of invoke

OpperProvider

A provider class for creating Opper chat models and managing traces.

Methods:

• create_chat_model(**kwargs): Create a new ChatOpperAI
• create_structured_model(task_name, instructions, output_schema, **kwargs): Create a structured model
• start_trace(name, input_data): Start a new trace
• end_trace(output_data): End the current trace
• add_metric(span_id, dimension, value, comment): Add a metric to a span

Environment Variables

• OPPER_API_KEY: Your Opper API key (required)

Requirements

• Python ≥ 3.9.2
• langchain-core ≥ 0.3.0
• opperai ≥ 1.0.0
• pydantic ≥ 2.0.0

License

MIT

Contributing

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

Links

• Opper AI
• LangChain
• LangGraph