1. SaaS Specification

• Build an API to feed a json object to a python script and get a json object back
• The API must be able to check credentials either with an API key or registered users
• Build functionality to generate and share API keys, and register users from a landing page (http://paraknowledge.ai)
• Build capability to monitor usage:
  • how many API calls
  • determine the pricing tier based on api-key or user-registration
  • determine the number of samples analyzed per api call (number of patients in the input json object)
• Build capability to charge account based on pricing tier and usage
• Have option for generating invoices

Required software

API's codebase must work on Python 3.9.16 exactly.

In order to simulate the intended software installation, versioning and usage, please install the zcor_dummy python package. It is necessary to simulate the loading and usage of the DUMMY_PREDICTOR.pickle

pip install --no-deps git+https://github.com/zeroknowledgediscovery/zcor_dummy.git

IMPORTANT :: the --no-deps argument above will ensure the consistency within the zcor_dummy requirements at the expense of the dependency consistency within the repo you install zcor_dummy to. Install it in a clean python environment or proceed with caution.

Sample dummy programs

Please take a look at the python script (run_dummy_model.py) provided to have an idea about the interface necessary. We next describe the IO specification

Data IO Specification

The input/output operations will be handled by the python script. An example, called run_dummy_model.py, will be provided for reference.

Example usage with provided data:

python3 run_dummy_model.py -p DUMMY_PREDICTOR.pickle -i sample_data.json -o sample_predictions.json -v True
python3 run_dummy_model.py -p DUMMY_PREDICTOR.pickle -i sample_data.json -v True

If -o if not specified, the script will print the json object into the STDOUT.

The output should be logged in a user history.

Input/Output Format

We expect the json input for the patient data in the following .json format:

[
    { 
        'patient_id': 'P00101606306',
        'sex': 'F',
        'location': 'Chicago, IL, 60601',
        'age': 0,
        'birth_date': '01-01-2006',
        'race': 'More than one race',
        'ethnicity': 'Hispanic or Latino',
        'income': '80000',
        'occupation': 'School Teacher',
        'family_relation': 'Mother',
        'DX_record': [
            {'date': '07-31-2006', 'code': 'Z38.00'},
            {'date': '07-31-2006', 'code': 'P59.9'},
            {'date': '08-07-2006', 'code': 'Z00.129'},
            {'date': '08-07-2006', 'code': 'P59.9'},
            {'date': '08-07-2006', 'code': 'P59.9'},
            {'date': '08-29-2016', 'code': 'J01.90'}
        ],
        'RX_record': [
            {'date': '10-29-2011', 'code': 'rxLDA017'},
            {'date': '05-16-2015', 'code': 'rxIDG004'},
            {'date': '08-08-2015', 'code': 'rxIDG004'},
            {'date': '06-04-2016', 'code': 'rxIDD013'},
        ],
        'PROC_record': [
            {'date': '02-05-2007', 'code': '90723'},
            {'date': '11-05-2007', 'code': 'J1100'},
            {'date': '11-05-2007', 'code': '99214'},
        ]
    }
]

The output will be of the following .json format:

[
    {'patient_id': 'P00101606306', 'predicted_risk': 0.0042816491258671, 'decision': 0, 'confidence': 0.0},
    {'patient_id': 'P00101606307', 'predicted_risk': 0.02346271, 'decision': 1, 'confidence': 0.8},
    {'patient_id': 'P00101606308', 'predicted_risk': 0.01765892, 'decision': 0, 'confidence': 0.5},
    {'patient_id': 'P00101606309', 'predicted_risk': 0.00239817, 'decision': 1, 'confidence': 0.9}
]

2. Frontend Design and Optimization

We have a preliminary example here: (http://paraknowledge.ai) This needs to be augmented or modified to provide the user registration, and potentially show usage and invoice to users after they register and login