This is a tool for extracting the nominations from Senate records using machine learning.
To begin with my computer does not currently make it easy to install tensorflow so I had to use conda - something I had not done before.
Using an ARM based Mac - I had to use the following instructions to get tensorflow installed Apple Link
Multiple models were orignially trained but here is an example that works for older 1991 senate records.
To begin
Unzip the training data and remove the zip files.
Train the model
python train.py
Use the model
python process.py
This will process each page from the 1991 Senate hearing on judicial nominations and attempt to extract out disclosures intout the ouptut directory.