Installing H2O AutoDoc

H2O AutoDoc Requirements

  • Python 3.6, 3.7 or 3.8

  • H2O-3 or higher

  • A supported H2O model. Supported models include:

    • Deep Learning

    • Distributed Random Forest (including Extremely Randomized Forest)

    • Generalized Linear Model

    • Gradient Boosting Machine

    • Stacked Ensembles

    • XGBoost

H2O AutoDoc Setup & Installation

  1. Install the required dependencies for h2o_autodoc.

    • Pandoc

      Pandoc Note:

# example using HomeBrew
brew install pandoc
# example for Ubuntu
dpkg -i pandoc*.deb
  1. Download and install the H2O AutoDoc.

  • Run the following to pip install latest H2O AutoDoc:

pip install h2o_autodoc
  • Alternatively, run the following to pip install a specific version of H2O AutoDoc:

pip install h2o_autodoc==1.0.5
  1. Create your first H2O AutoDoc - follow mini-tutorials in Creating & Configuring H2O AutoDoc section.

H2O AutoDoc License Key

The H2O AutoDoc Python package requires a license key. The license key can be specified through several options:

  • Through a Config class parameters:

    • license_file: the file system location for the license file

    • license_text: the license text

  • Through an Environment variable:

    • ‘AUTODOC_LICENSE_FILE’: the location of file containing the license key

    • ‘AUTODOC_LICENSE_KEY’: the license key

  • Through License file in standard location:

    • /license/license.sig: the file containing the license key

    • ~/.h2o_autodoc/license.sig: the file containing the license key on user home

Setting License Key Example

The following example shows how to set your license key as the license_text string parameter in the Config class.

# H2O AutoDoc package imports
from h2o_autodoc import Config
from h2o_autodoc import render_autodoc

# specify the full path to your H2O AutoDoc Report
output_file_path = 'full/path/to/your/autodoc/report_H2O3.docx'
config = Config(output_path=output_file_path, license_text="your license key here")

# generate an H2O AutDoc report for your model and train dataset
render_autodoc(h2o=h2o, config=config, model=model, train=train)