This is an in-development tool for test purposes only. This tool is not currently tested for clinical use.
Please see requirements.txt for installation of required dependencies.
This repository provides scripts for the development of a semantic textual similarity model to link common symptom inventories used in TBI research. It also contains codes for development of a web tool.
The code for the development of linked STS scores across inventories is available in the script STS_score.py. The crosswalk model using these STS scores is available in crosswalk_symptom_inventories.py.
app.py implements a local browser version of the online web tool using the command shiny run app.py. After any changes to the web tool, please run shiny run --reload --launch-browser your_dir/app.py and reload the browser, where "your_dir" is the folder of the application.
Full documentation of the tool properties, accuracy, and development are available online in the associated publication: TBD
A shiny app hosting this application is available here
Instructions for the web tool:
- Select the inventory input and output from the dropdown menu. Then input the numbers into the fields (the ranges are specified below the inventory name).
- Run conversion: Click the "Convert table" button to generate the converted scores.
- Download converted data: There is also a 'Download Conversion' button to download the converted table as a csv.