Documentation

We assume you need a textual representation of a specific section of the judgment document, for instance The Facts.

We also assume that you do not want to use SQL but rather a python library such as pandas.

1. Locating the proper section in the judgment document

First, you need to load the file containing all the information about all cases: build/unstructured/cases.json.

The file contains a list of dictionary where each dictionary represents all the information about a case.
The content of documents associated to a given case is under the key content. The field content contains itself a dictionary of documents related to a case. The judgment representation has the form of a tree where each node is:

{
	'content': "paragraph_or_section_title"
	'elements': []
}

For first level nodes, corresponding to the main sections of the judgment, there is an attribute section_name to identify specific sections.

It is then enough to find the node with the attribute section_name": "facts".

2. Reconstructing the text for the specific section

With the content of the section, if you need just the full text, you need to recursively reconstruct it.

This can be done by recursively aggregating each node using the following function: