ANPRE

Project to Automate NeuroPsych Record Entry. We aim to utilize AI to engineer a system for automated processing of neuropsych patient evaluation scans into spreadsheets to allow for statistical analysis and further research.

Instructions

Below are instructions for installing and using the package as a user and developer.

User Instructions

User should create the default output directories utilized in each processing stage.

$ mkdir case_outputs case_outputs case_outputs/sheets case_outputs/seg case_outputs/tables case_outputs/tables_clean

First, to select clinical summary sheet scans from full case files, run select_sheets.py:

$ python src/select_sheets.py -i data/full -o case_outputs/sheets -xl case_outputs/patient_data.csv

Next, to rectify and segment the scans into proper sections, run segment_sheets.py:

$ python src/segment_sheets.py -i case_outputs/sheets -o case_outputs/seg -t templates/spring2025.json

In order to scan sheets and enter data into CSV tables, run scan_sheets.py:

$ python src/scan_sheets.py -i case_outputs/seg -o case_outputs/tables

To process and clean CSV table data, run postprocessing.py:

$ python src/postprocessing.py -i case_outputs/tables -o case_outputs/tables_clean

Developer Instructions

For development, clone the repository and install the developer requirements in a virtual environment. Development allows for training of new models using the scripts directory.

$ pip install -r requirements_dev.txt

Install pre-commit hooks

$ pre-commit install

Run pre-commit hooks to ensure code quality

$ pre-commit run -a

Authors

1. Michal Brzus

   github: mbrzus, email: michal-brzus@uiowa.edu

2. Ashby Martin

   github: Ashby-Martin, email: ashby-martin@uiowa.edu

3. Zihe (Lucas) Wang

   github: Brocoliman, email: zihe-wang@uiowa.edu