@nuclia/protobuf

JavaScript utility to handle Nuclia protobuffer messages.

Functions

• NucliaProtobufConverter is a function accepting a buffer and returning a Promise that resolves to a parsed protobuffer message.

• extractParagraphTexts is a function adding to each Paragraph object a text attribute containing the actual text of the paragraph (by default, Paragraph objects only contain the position of their first and last characters).

Installation

npm install @nuclia/protobuf

or

yarn add @nuclia/protobuf

Usage

When using a webhook with the Nuclia Understanding API, the message sent to the webhook is a protobuffer message. It can be decoded using the NucliaProtobufConverter.

import { NucliaProtobufConverter } from '@nuclia/protobuf';

router.post('/handle', (request, response) => {
  NucliaProtobufConverter(request.body).then((message) => {
    console.log(message);
  });
});

When calling Nuclia Understanding API /pull endpoint, you get a JSON message containing a payload entry with base64 encoded protobuffer message. Processing it could be done as follows:

import { NucliaProtobufConverter } from '@nuclia/protobuf';
import fetch from 'node-fetch';

fetch('https://europe-1.nuclia.cloud/api/v1/processing/pull', {
  headers: {
    'x-stf-nuakey': `Bearer ${NUA_KEY}`,
    'content-type': 'application/json',
  },
  method: 'GET',
})
  .then((res) => res.json())
  .then((data) => {
    if (data.status === 'ok' && data.payload) {
      return NucliaProtobufConverter(Buffer.from(data.payload, 'base64'));
    } else {
      return data;
    }
  })
  .then((res) => console.log(res));

To get paragraphs text, you can use the extractParagraphTexts function:

import { NucliaProtobufConverter, extractParagraphTexts } from '@nuclia/protobuf';

router.post('/handle', (request, response) => {
  NucliaProtobufConverter(request.body).then((message) => {
    console.log(extractParagraphTexts(message).fieldMetadata[0].metadata.metadata.paragraphs);
  });
});