LangForge Toolkit

To debug tests using the Playwright inspector tool, run the tests in debug mode by setting the PWDEBUG environment variable to 1.

1. Compile the extension:

   jlpm install
   jlpm build:prod

2. Install test dependencies (if not already done):

   cd ./ui-tests
   jlpm install
   jlpm playwright install
   cd ..

3. Run tests in debug mode:

   cd ./ui-tests
   PWDEBUG=1 jlpm playwright test

To remove the extension from your environment, use pip uninstall.

pip uninstall langforge

Use the langforge create command to initialize a new project. The command will prompt you for configuration details, set up a virtual environment, install necessary packages, and configure your API keys (such as OpenAI).

langforge create myapp

Navigate into your project directory and use the langforge lab command to launch JupyterLab. This environment includes predefined templates and a built-in chat interface to interact with your LangChain chains.

cd myapp
langforge lab

If you are developing the extension, follow these steps to set up a development environment. You will need NodeJS installed.

Note: jlpm is JupyterLab's pinned version of yarn. You can substitute jlpm with yarn or npm if preferred.

1. Clone the repository and enter the langforge directory.
2. Install the package in editable mode using pip install -e ".".
3. Link your development version to JupyterLab using jupyter labextension develop.
4. Build the TypeScript source using jlpm build.

# Clone the repo to your local environment
# Change directory to the langforge directory

# Install package in development mode
pip install -e "."

# Link your development version of the extension with JupyterLab
jupyter labextension develop . --overwrite

# Rebuild extension Typescript source after making changes
jlpm build

To install the LangForge extension, ensure you have JupyterLab version 3.0 or higher installed, then use pip to install the package.

Requirements

• JupyterLab >= 3.0

pip install langforge

If your tests use snapshot comparisons and you need to update the reference snapshots in the repository, follow these steps from the repository root:

1. Compile the extension:

   jlpm install
   jlpm build:prod

2. Install test dependencies (if not already done):

   cd ./ui-tests
   jlpm install
   jlpm playwright install
   cd ..

3. Run Playwright with the update flag:

   cd ./ui-tests
   jlpm playwright test -u

Tip for Pull Requests: To avoid discrepancies between local snapshots and CI snapshots, you can type please update playwright snapshots in your PR to trigger a bot on the CI to compute and commit new snapshots for you.

# From root
jlpm install
jlpm build:prod

# From ui-tests directory
cd ./ui-tests
jlpm install
jlpm playwright install
cd ..

cd ./ui-tests
jlpm playwright test -u

LangForge automatically generates a REST interface for your notebooks. To deploy your app, use the langforge serve command followed by the filename of your notebook.

langforge serve chat-creative.ipynb

To see changes in real-time, run the jlpm watch command in one terminal. This will automatically rebuild the extension whenever you save a file. In a separate terminal, run jupyter lab to start the JupyterLab server.

Note: After a rebuild, you must refresh your browser tab in JupyterLab to load the changes. It may take several seconds for the rebuild to complete.

To make debugging easier by generating source maps for JupyterLab core extensions, run: jupyter lab build --minimize=False.

# Watch the source directory in one terminal, automatically rebuilding when needed
jlpm watch

# Run JupyterLab in another terminal
jupyter lab

Install the LangForge toolkit using the langforge-ai package from PyPI.

pip install langforge-ai

The easiest way to create new integration tests is to use the Playwright code generator. This tool records your interactions with the browser and generates code automatically.

1. Compile the extension:

   jlpm install
   jlpm build:prod

2. Install test dependencies (if not already done):

   cd ./ui-tests
   jlpm install
   jlpm playwright install
   cd ..

3. Launch the code generator:

   cd ./ui-tests
   jlpm playwright codegen localhost:8888

   Note: Replace localhost:8888 with the actual URL where your JupyterLab server is running.