Flask web application

Language: Python 2.7

The externally facing component of the Willow backend which serves as the point of communication for the front-end (Angular app instances) and acts as a mediator, relaying data operation requests received from clients to Celery workers. The Flask application also handles importing and exporting of datasets.

Dependencies

• Flask as the web application framework
• Flask-SocketIO for enabling support for WebSocket connections with SocketIO
• Flask-SQLAlchemy for persistently associationg SocketIO client IDs with Willow sessionIDs using SQLite

WebSocket requests

The Flask web application supports all of the requests detailed in the WebSocket API, and relays each requests to the appropriate Celery task function.

HTTP endpoints

POST /upload

Import a new dataset into Willow and initialize a session

Data Params

Content-Type: multipart/form-data

• file=[File]

  CSV, JSON or Excel dataset (maximum size 1GB)
• initialSkip=[Integer], optional

  specify first n lines to skip (for CSV and Excel files)
• sampleSize=[Float], optional

  define the proportion of the dataset for random sampling (between 0 and 1)
• seed=[Integer], optional

  seed value for random sampling, use for replicable sampling

Response

Content-Type: application/json

{
        'success': [boolean],
        'sessionID': [30 character long hexadecimal string]
}

Note

sessionID outputted in response must be saved, as it is necessary for referring to the uploaded dataset in future HTTP and WebSocket requests

GET downloadJSON/<sessionID>

Initiate a dataset file download in JSON format.

e.g. GET downloadJSON/617646cb1e421f72b7e742dbdbd4cb

Response

Content-Disposition: attachment

Content-Type: application/json

GET downloadCSV/<sessionID>

Initiate a dataset file download in CSV format.

e.g. GET downloadCSV/617646cb1e421f72b7e742dbdbd4cb

Response

Content-Disposition: attachment

Content-Type: text/csv

POST /celeryTaskCompleted

Notify a connected client over WebSocket that a requested operation has been completed.

Caution

This endpoint is used internally for communication with Celery workers, and should not be considered part of the HTTP API.

Data Params

Content-Type: application/json

The JSON in the body of the POST request is forwarded as is to the corresponding WebSocket client connected to the Flask application.