drachs / langchain-trace-and-test

A docker environment integrated with visual studio code designed to trace and test langchain scripts

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

A docker environment integrated with visual studio code designed to trace and test langchain scripts

  1. Install the vscode remote system for working with containers: https://code.visualstudio.com/docs/devcontainers/containers
  2. Open this folder in vscode, it should build and launch your container
  3. Open a terminal and run i. jupyter notebook --allow-root
  4. Connect to the url provided by jupyter and go ham

To use langchain tracing:

  1. Kill any already running jupyter notebook
  2. Open a new terminal and run i. ./launch-with-tracing.sh
  3. The jupypter server link should appear in the terminal
  4. The langchain trace link should be: http://127.0.0.1:4173/sessions i. Run some langchain stuff i. Navigate to the tracer and hit the button labled "Session: Default" i. You should see traces here, you may have to refresh the page for new traces to show up

About

A docker environment integrated with visual studio code designed to trace and test langchain scripts


Languages

Language:Jupyter Notebook 89.3%Language:Dockerfile 8.1%Language:Shell 2.6%