Protocol Deviation Digital Healthcare Product
  • INTRODUCTION
    • Purpose
    • Scope
    • Deployment
  • USER GUIDE
    • User Login
    • User Roles and Permissions
      • Administration
        • User Management
        • Team Management
        • Study Management
        • Audit Trail
    • Data Files
      • Data File Requirements
      • Spreadsheets
        • Using the Spreadsheet for AI Categorisation
        • Using the Spreadsheet for Manual Categorisation
    • Data Upload
      • Audit Trail
      • Data Trail
      • Data Categorisation
    • Data Visualisation
      • Team Visualisations
      • Study Visualisations
    • Maintenance and Support
      • Troubleshooting for users
      • Support contact information
  • TECHNICAL DOCUMENTATION
    • Web App
      • Design
        • Database
        • Backend
        • Frontend
      • Installation
    • AI Classifier
      • User documentation
      • Administrator documentation
      • Code documentation
    • Developer Guide
      • Requirements
      • Installation
      • Builds
      • Database
Powered by GitBook
On this page
  • Administrator documentation
  • Description
  • Requirements
  • Installation
  • Startup
  • Testing
  1. TECHNICAL DOCUMENTATION
  2. AI Classifier

Administrator documentation

PreviousUser documentationNextCode documentation

Last updated 1 month ago

Administrator documentation

The Protocol Deviation (PD) Classifier is a microservice system designed to handle requests over a specified server and port.

Description

The PD classifier accepts only "POST" requests. For details on querying the system, please refer to the User's Guide. The PD classifier is built on Uvicorn, an Asynchronous Server Gateway Interface (ASGI) web server implementation for Python.

Requirements

  • Python 3.7+

  • pip

Installation

  1. Download or clone the repository to your local machine.

  2. In the base directory of the cloned repository, navigate to the pdai directory : 

cd pdai

  1. Run the pip command with the parameters below:

pip install .

CAUTION

Please note that the last character in the command is a period (.)

Once all dependencies are installed, the PD classifier is ready to use.

Startup

To start the system, run the following command:

python pdai/app.py <listening_address> <listening_port>

Note

Depending on your operating system (OS), it may be possible to integrate this command as an OS service. Requirements will vary based on the host server’s configuration.

The parameters <listening_address> and <listening_port> specify the server address and the IP port that will listen for incoming requests. Additionally, the classifier only processes POST requests made to the /prediction endpoint. Calls to any other path or using a different request method will return an error message.

If these parameters are not provided, the server will listen on the host's IP address and port 8088.

CAUTION

Parameters must be provided in the specified order. If you need to run the service on a different IP address but use the default port, only the first parameter should be specified. If the service needs to run on a different port while using the default IP address, both parameters must be provided. Failure to adhere to these requirements may result in an error.

The service will continue running and will send messages to the console indicating its activity:

Starting application on 0.0.0.0:8088! INFO: Started server process [5226]
INFO: Waiting for application startup. INFO: Application startup complete. 
INFO: Uvicorn running on http://0.0.0.0:8088 (Press CTRL+C to quit) 
INFO: 127.0.0.1:52985 - "POST /prediction HTTP/1.1" 200 OK 
INFO: 127.0.0.1:53569 - "POST /prediction HTTP/1.1" 200 OK

The PD classifier will stop when the CTRL+C keys are pressed or when the operating system prompt returns to the user for any other reason. A message like the following may appear.

INFO: Shutting down INFO: Waiting for application shutdown. 
INFO: Application shutdown complete. 
INFO: Finished server process [5226]

The status of the PD classifier can be monitored by checking the process number indicated in the line INFO: Started server process [process_number], where the process_number corresponds to the identifier assigned to the server.

Testing

To validate the service, you can send a query to the server as outlined in the user's guide.