NLP Search

Description

This function package provides the following feature to be used in a workflow:

• Search existing incidents and find the top similar ones.

In addition, this package contains a command line tool used to build a Natural Language Process (NLP) model. Please refer to the User Guide for instructions on building the model and running predictions.

System Requirements

• Resilient platform version 32 or later
• Ability to connect to a Resilient platform with HTTPS on port 443 and 65001
• The minimum set of Resilient API permissions if using an API key account:
  • Incidents.Read
  • Edit Org Data
  • Functions.Read

Package Dependencies

• sklearn 0.19.2
• pandas 0.23.3
• numpy 1.12.1
• scipy 1.1.0
• Resilient Circuits version 32 or later

Installation

Install this package with 'pip', or run python setup.py install

Setup

Create app.config by running resilient-circuits config -c or update app.config with resilient-circuits config -u.

The app.config file needs the following configuration values, in addition to those in the appropriate [resilient] section for connecting to your Resilient platform:

[fn_machine_learning_nlp]
#
# Required. The (absolute) path to the folder of the saved NLP model
#
model_path=path_of_the_saved_models

#
#   Advanced configuration
#-------------------------
# Use the followings to optimize the performance of a NLP model
#
# Number of features for NLP word2vec model.
num_features=50

Customize

Run with: resilient-circuits customize to install function definitions and sample workflows to the Resilient platform.

Start

Start this function with: resilient-circuits run

Build a new NLP model

Use the command line tool:

res-ml build_nlp

to build a machine learning model

Rebuild the NLP model

Same as building a new model.

As a rule, it is recommended that customers rebuild their model once a week.

There are more advanced options. Please refer to the User Guide.

Uninstall

pip uninstall fn-machine-learning-nlp

Troubleshooting

If res-ml build_nlp fails due to a LookupError caused by SSL errors and NLTK Downloader unable to obtain specific resources, you can run the file bin/nltkdownload.py and download everything for NLTK. This will resolve any future SSL errors when trying to access NLTK resources.