OntoLoader

This package is under development. You can take an early peak at your own risk :)

OntoLoader is a package to extract the logical structure of some world. In its final version, it will provide tools to identify entities, events and situations in a world, as well as associated tagged data and vector representations.

Installation

We recommend installing this package in a virtual environment. If you do not have virtualenv installed, you can get it in the following way:

sudo apt update
sudo apt install python3-setuptools
sudo apt install python3-pip
sudo apt install python3-virtualenv

Then, create a directory for ontoloader, make your virtual environment and install the package:

mkdir ontoloader
cd ontoloader
virtualenv env && source env/bin/activate
pip install git+https://github.com/possible-worlds-research/ontoloader.git

Data collection

You will need some data to play with. To get you started, we recommend grabbing some Wikipedia data using the WikiLoader package. For instance, let us grab the plots of some action novels to play with:

from wikicategories.wikicategories import WikiCatProcessor

lang = 'en'

catprocessor = WikiCatProcessor(lang)

categories = ['Action novels']
catprocessor.get_category_pages(categories)
catprocessor.get_page_content(categories, doctags=False, tokenize=False, lower=False, sections=['Plot'])

The corpus we obtain will be saved at data/en/categories/Action_novels/linear.plot.txt. It contains the plot summaries of the books tagged as 'Action novel' in Wikipedia.

Getting character 'bounding boxes'

If you are familiar with 'bounding boxes' from the computer vision literature, you will recognise them as parts of an image that uniquely identify a particular entity. We are expanding the concept here and introduce 'text bounding boxes', which are snippets of text related to a single entity. The snippets have undergone coreference resolution so that the pronouns referring to the entity are resolved in the output.

from ontoloader.ontoloader import OntoLoader

ontoloader = OntoLoader()
resolved,idx = ontoloader.coref('data/en/categories/Action_novels/linear.plot.txt')
ontoloader.bounding_box_people(resolved,idx)