The Monarch Hippo

Introduction

This is a search browser written in React JS to provide a user interface for Phenomics backend services (Phantom). The browser's core functionalities are:

1: Search biomedical entities in PubMed abstracts annotated by Phantom. Currently there are 4 types of entities supported (GENE, PHENOTYPE, DISEASE, DRUG)

2: Once search brings back the results, displaying all abstracts with highlighted annotations, PubMedID, Title and Author information.

3: Collecting all other annotations present in search abstracts and treat them as a filter.

4: Feedback for annotation quality

5: Subscription for search items

Following are dev and prod URL for backend:

Dev: https://phenomics-dev.kccg.garvan.org.au/search-engine/

Prod: https://phenomics.kccg.garvan.org.au/search-engine/

Setup

Environment

Node and NPM are required to build and run this app locally.

On Mac this can be done achieved using homebrew

brew update

brew install node

Once done, you can check the versions to make sure installation is done properly:

node -v

npm -v

Build

Build process is fairly simple, a backend service url is needed with build command.

REACT_APP_BASE_URL=https://phenomics-dev.kccg.garvan.org.au/search-engine/ npm run build

This will create a build folder, which is the complete app and can be accessed with index.html page.

Deployed Application

Currently the application is deployed on an ECS cluster and can be accessed with following URLS: Production: https://hippo.monarchinitiative.org/#/ Alias: https://phenomics.kccg.garvan.org.au/hippo/#/

Dev: https://phenomics-dev.kccg.garvan.org.au/hippo/#/

It is advisable to move the deployments to Monarch servers in order to keep them running seamlessly.

Backend APIs

Here are a bunch of backend APIs along with sample input/output:

Only prod links are listed here, the host prefix should be

https://phenomics.kccg.garvan.org.au/search-engine/

1 Auto Complete

GET /query/autocomplete/{infix}

Input: A case insensitive string to fetch matches from all 4 ontologies.

Response:

[
  {
    id: "string",
    rank: 0,
    text: "string",
    type: "string"
  }
];

2 Search

POST /query/

Input:

params: pageSize, pageNo, includeHistorical

body:

{
  "filterItems": [
    "string"
  ],
  "queryId": "string",
  "searchItems": [
    "string"
  ]
}

Example: https://phenomics.kccg.garvan.org.au/search-engine/query/?pageSize=10&pageNo=1&includeHistorical=false

  "searchItems": [
    "1100"
  ]
}

Response:

{
  "articles": [
    {
      "annotations": [
        {
          "endIndex": 0,
          "highlights": [
            {
              "description": "string",
              "feedbackId": "string",
              "id": "string",
              "link": "string",
              "standard": "string",
              "text": "string",
              "type": "string"
            }
          ],
          "startIndex": 0
        }
      ],
      "articleAbstract": "string",
      "articleRank": 0,
      "articleTitle": "string",
      "authors": [
        {
          "foreName": "string",
          "initials": "string",
          "lastName": "string"
        }
      ],
      "datePublished": "string",
      "language": "string",
      "pmid": 0,
      "publication": {
        "isoAbbreviation": "string",
        "issnNumber": "string",
        "issnType": "string",
        "title": "string",
        "validKey": true
      }
    }
  ],
  "filters": [
    {
      "articleCount": 0,
      "filteredArticleCount": 0,
      "id": "string",
      "rank": 0,
      "text": "string",
      "type": "string"
    }
  ],
  "pagination": {
    "includeHistorical": true,
    "pageNo": 0,
    "pageSize": 0,
    "totalArticles": 0,
    "totalPages": 0
  },
  "query": {
    "filterItems": [
      {
        "id": "string",
        "text": "string",
        "type": "string"
      }
    ],
    "queryId": "string",
    "searchItems": [
      {
        "id": "string",
        "text": "string",
        "type": "string"
      }
    ]
  }
}

3 Subscription

POST /subscription

GET /subscription

GET /subscription/{subscriptionId}

DELETE /subscription/{subscriptionId}

Input:

body:

{
  "digestFrequencyInDays": 0,
  "emailId": "string",
  "query": {
    "filterItems": [
      "string"
    ],
    "queryId": "string",
    "searchItems": [
      "string"
    ]
  },
  "searchName": "string",
  "subscriptionId": "string"
}

4 Feedback

POST /feedback

Input:

Feedback ID is provided with each annotation in the articles result. Value of feedback can be "good" or "bad".

body:

{
   "feedback": "string",
   "feedbackId": "string"
}