kpwhri/dqt_api

Simple backend API for a data query tool.

About

Backend using flask with cherrypy or tornado server.

This manages a data cohort, allowing users to make requests of the data. See the dqt-client for more info on frontend use cases.

Doco

Setup

1. Create a base directory where logging and configuration information can be stored, and put this in config.py file.

    BASE_DIR='path/to/local/storage'  # location for logs, etc.
    SECRET_KEY=b''  # generate this with os.urandom(24), any number is fine
    SQLALCHEMY_DATABASE_URI = 'sqlalchemy-connection-string'  # see # see: http://docs.sqlalchemy.org/en/latest/core/engines.html
    ORIGINS = ['*']  # origins to allow for cors
    # optional
    LOG_KEY=b''  # if you want to encrypt old log files, os.urandom(32)
    MASK = 5  # mask values smaller than this number
    AGE_STEP = 5  # increment for age graph
    AGE_MAX = 90  # max age to show (otherwise inferred from data)
    AGE_MIN = 60  # minimum age to show (otherwise inferred from data)
    COHORT_TITLE = 'ACT'  # short name for population under review
   

2. Install requirements with pip install -r requirements.txt

3. Create database (run create database data_query_tool;)

4. Build tables:

   • python manage.py --method create --config /path/to/config.py
   • python manage.py --method createuserdata --config /path/to/config.py

5. Load data using load_csv.py

   • This isn't meant to be a general purpose load script, so it will require some modification on your part
   • You can also auto-fill 100 subjects by using:
     • python manage.py --method load --count 100 --config /path/to/config.py

6. Run the application with python dqt_api --config /path/to/config.py.

   • Instead of adding --config, you can specify the environment variable FLASK_CONFIG_FILE

7. Navigate to http://127.0.0.1:8090 for test page.

8. First request from client will take longer (be patient) as indexes are being built.

   • When reloading data, delete the dump.pkl file in BASE_DIR