Some natural language processing of the Dail and Seanad of individual members / political parties. Do things like see the difference between the words people/parties use along with predicting what political party a person will align with given a sample of their speech.
Poppler - sudo apt install libpoppler-cpp-dev
PDFToHTML - sudo apt install pdftohtml
pip install oireachtas-nlp
After installation run pull_debates for a while to get some data. Do this for as long as you want. The longer the better
Also run pull_members to get member data.
Look at the help of any of the below for a better description of how to use it.
oir_sounds_like
This generates a classifier with which you can see what a pice of text beongs to (a member or a party)
oir_belong
This will go through members and determine if they speak like the party they belong to (or not and which one they do sound like they belong to).
oir_word_usage_by
This will show the difference in words used between members or parties.
oir_sentiment
This will show how positive / negative members / parties are.
oir_lexical_diversity
This shows how diverse the word usage is of members / parties.
oir_lexical_difficulty
This shows how difficult the language is of members / parties. Many methods of calculation available.
oir_generate_text
Generate sentences based on the content of a party / member. You can also give prefixes and have the sentence finished with generated text. TODO: With this is another command for pretrained models oir_generate_text_from_model