Skip to content

agold-rh/healthcare-nlp-visualizer-demo

 
 

Repository files navigation

Healthcare NLP Visualizer Demo

The demo application is a Node.js and React.js system to visualize the Google Cloud Healthcare Natural Language API. You can upload your own sample medical text to visualize the output such as medical dictionaries, entity extraction and relationships, context assessment and more. We have also provided sample texts for a a medical record, research paper and lab form.

screencast

Prerequisites

  1. A GCP Project with billing and the Healthcare NLP API enabled.
  2. Complete the Healthcare NLP How-to Guide.
    • Note: For the purposes of this demo, grant the new service account "Viewer", not "Owner".
  3. Familiarity with Google Cloud Functions and Vue.js.

Set Up Instructions

Important

This demo will create a cloud function that is executable by ANYONE on the Internet. Be sure to delete it when your demonstration is complete.

Backend

The HTTP Cloud Function can be found in the /analyzeDocument directory. Please note, this code is NOT meant for production use.

  1. grep the /analyzeDocument directory for the string REPLACE.
    • You will need to replace these placeholders with real values.
  2. Download the service account key for your project.
  3. Deploy the Cloud Function, you can follow the instructions here.
    • See the sample invocation in util/deploy_cf.sh
  4. Copy the endpoint for your Cloud Function.

Frontend

The Vue.js app is found in the /app directory.

  1. grep the /app directory for the string REPLACE.
    • You will need to replace these placeholders with real values.
  2. cd app/
  3. Paste your Cloud Function endpoint in to the placeholder in index.html.
  4. Start a local server of your choice and open the application in your browser.
    • See the sample server in util/server.py.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • JavaScript 49.4%
  • HTML 31.9%
  • Shell 7.2%
  • CSS 6.8%
  • Python 4.7%