ParlaMint-RO corpus builder

This repository contains the code required for building the ParlaMint-RO corpus.

Each script in the root of this repository represents a step in the processing pipeline. Theese scripts can be classified as follows:

Prepreprocessing scripts

• build-speakers-list.py - iterates through session transcripts in JSON format and builds a list of unique speaker names, which is then saved to a CSV file.
• classify-speakers.py - iterates through session transcripts in JSON format and classifies speakers into MPs and invited speakers; the lists are saved in CSV format.

Corpus building script

The script to build the corpus is build-corpus.py.