Reinforcement

Introduction

A data synthesizer and masker that takes in real FHIR patient data and generate data providing certain rules.

This tool is developed with performance and customisation in mind. See below on how to customise with custom data modules.

Installation

mkdir build
cd build
cmake ..

To compile and use under current folder.

make

Or to install in system (POSIX system only),

make install

Usage

• --quiet for SILENT Mode
• --use-cache for loading cached patients.json file at a custom location, to be used with --cache-loc
• --api ${YOUR_CUSTOM_API}
• --cache-loc ${CACHE_FILE_LOCATION}
• -g ${NO OF PATIENTS TO BE GENERATED}
• -o ${OUTPUT_FILE_LOCATION}
• -h or --help for showing the help message

Sample

./GOSH_FHIRworks2020_datamaskgen --api https://localhost:5001/api/Patient -g 100 -o ./

Customisation

Customise any of the generator in the src/generator folder, or add your own class with suffix _generator.

Your customised class will be required to feed in a const reference to original patients list vector, and then generate required information based own your customised algorithm.

Specify your output of custom data set in jsonify() function in src/patient/patient.hh, and sit back to see the new data getting generated.

Performance

Build with performance in mind, Data MaskGen utilises system resources much better than the industrial standard synthea and therefore archives much better running time with similar results.

Patient Amount	Data MaskGen	Synthea
1	98% cpu, 1.836s	10% cpu, 13.067s
50	99% cpu, 1.863s	9% cpu, 15.817s
100	99% cpu, 1.917s	6% cpu, 23.035s
500	99% cpu, 2.258s	5% cpu, 30.726s
1000	99% cpu, 2.601s	4% cpu, 41.757s