whizzml_kernel

whizzml_kernel is a simple Jupyter kernel that allows you to run WhizzML code in a Jupyter Notebook.

Installation

Node.js and npm are needed for the kernel to work properly. If you need information to install those, please refer to https://docs.npmjs.com/downloading-and-installing-node-js-and-npm. The rest of dependencies will automatically be installed when installing whizzml_kernel from PyPI::

    pip install whizzml_kernel
    npm install xmlhttprequest
    python -m whizzml_kernel.install

WhizzML credentials

In order to access BigML using WhizzML, you'll need to provide your credentials. Currently, the credentials are extracted from the existing environment variables: BIGML_USERNAME, BIGML_API_KEY, BIGML_DOMAIN, BIGML_API_VERSION, BIGML_PROJECT and BIGML_ORGANIZATION. The first two are mandatory, while the rest can be missing. If so, BIGML_DOMAIN will be set to https://bigml.io, the BIGML_API_VERSION will be set to the latest version and no project or organization will be used. In order to set or change those, you'll need to use the %env magic command, as you would using the Python kernel.

Using the WhizzML kernel

Notebook: The New menu in the notebook should show an option for a WhizzML notebook.

Console frontends: To use it with the console frontends, add --kernel whizzml to their command line arguments.

The goal of this kernel is providing a REPL to use WhizzML. As an example, you can try basic WhizzML operations, like defining a variable:

defining a function:

calling the function on two numbers:

creating a source in BigML from some data:

Summarizing and storing your WhizzML

Some magics have been added to the currently available Python magics as utilities that can help you store your WhizzML code and create a script from it. The steps to do so are:

• Saving the notebook, so that the corresponding .ipynb file contains the current version of the code.
• Using the %wstore magic command in a separated input cell. The command will call nbconvert to extract from the stored notebook the WhizzML code and will create a .whizzml file with it. After that, a new cell will be added to your notebook containing the %%wscript magic command and the JSON needed to define the name, description, inputs and outputs of the script. Inputs and outputs will be pre-populated by analizing the WhizzML variables, but their type and defaults should be modified at will.
• Running the %%wscript command with the correctly formatted JSON will create a script from the code and the metadata.

If you want to see the WhizzML code retrieved from your input cells, you can use the magic command %wsource, that will output that for you.

Also, if you want your environment (credentials, stored WhizzML code, etc.) to be removed, you can use %wreset that will clear all the stored information without affecting the code in the input cells of your notebook.