LangChain

Comet provides seamless integration with LangChain, allowing you to easily log and trace your LangChain-based applications. By using the CometTracer callback, you can automatically capture detailed information about your LangChain runs, including inputs, outputs, and metadata for each step in your chain.

You can check out the Colab Notebook if you'd like to jump straight to the code:

Getting Started

To use the CometTracer with LangChain, you'll need to have both the opik and langchain packages installed. You can install them using pip:

pip install opik langchain langchain_openai

In addition, you can configure Opik using the opik configure command which will prompt you for the correct local server address or if you are using the Cloud platfrom your API key:

opik configure

Using CometTracer

Here's a basic example of how to use the CometTracer callback with a LangChain chain:

from langchain.chains import LLMChain
from langchain_openai import OpenAI
from langchain.prompts import PromptTemplate
from opik.integrations.langchain import OpikTracer

# Initialize the tracer
opik_tracer = OpikTracer()

# Create the LLM Chain using LangChain
llm = OpenAI(temperature=0)

prompt_template = PromptTemplate(
    input_variables=["input"],
    template="Translate the following text to French: {input}"
)

llm_chain = LLMChain(llm=llm, prompt=prompt_template)

# Generate the translations
translation = llm_chain.run("Hello, how are you?", callbacks=[opik_tracer])
print(translation)

# The CometTracer will automatically log the run and its details to Comet

This example demonstrates how to create a LangChain chain with a CometTracer callback. When you run the chain with a prompt, the CometTracer will automatically log the run and its details to Comet, including the input prompt, the output, and metadata for each step in the chain.

Settings tags and metadata

You can also customize the CometTracer callback to include additional metadata or logging options. For example:

from opik.integrations.langchain import OpikTracer

opik_tracer = OpikTracer(
    tags=["langchain"],
    metadata={"use-case": "documentation-example"}
)

Accessing logged traces

You can use the created_traces method to access the trace IDs collected by the CometTracer callback:

from opik.integrations.langchain import OpikTracer

opik_tracer = OpikTracer()

# Calling Langchain object

traces = opik_tracer.created_traces()
print([trace.id for trace in traces])

This can be especially useful if you would like to update or log feedback scores for traces logged using the CometTracer.

Advanced usage

The CometTracer object has a flush method that can be used to make sure that all traces are logged to the Comet platform before you exit a script. This method will return once all traces have been logged or if the timeout is reach, whichever comes first.

from opik.integrations.langchain import OpikTracer

opik_tracer = OpikTracer()
opik_tracer.flush()