@hyperaudio/transcript-parser/es/index.js

Hyperaudio HTML transcript to JSON converter

import htmlparser from 'htmlparser2';

export default (function (html) {
  var offsets = arguments.length > 1 && arguments[1] !== undefined ? arguments[1] : false;
  var digits = arguments.length > 2 && arguments[2] !== undefined ? arguments[2] : 2;
  return new Promise(function (resolve, reject) {
    var json = {
      words: [],
      paragraphs: [],
      speakers: []
    };

    if (offsets) json.transcript = '';

    var currentWord = {};
    var currentParagraph = void 0;

    var parser = new htmlparser.Parser({
      onopentag: function onopentag(name, attrs) {
        var word = {};

        if (name === 'p') {
          currentParagraph = {};
          if (offsets) currentParagraph.startOffset = json.transcript.length;
          json.paragraphs.push(currentParagraph);

          // if (attrs[`data-tc`]) {
          //   const [hh, mm, ss] = attrs[`data-tc`].split(`:`);
          //   currentParagraph.start = parseInt(hh) * 3600 + parseInt(mm) * 60 + parseInt(ss);
          // }
        }

        if (attrs['data-m']) {
          word.start = parseInt(attrs['data-m']) / 1e3;
          if (attrs['data-d']) word.end = word.start + parseInt(attrs['data-d']) / 1e3;
        }

        if (attrs['data-t']) {
          var _attrs$$split = attrs['data-t'].split(','),
              start = _attrs$$split[0],
              duration = _attrs$$split[1];

          word.start = parseFloat(start);
          if (!isNaN(parseFloat(duration))) word.end = word.start + parseFloat(duration);
        }

        if (attrs['class'] && attrs['class'] === 'speaker') {
          word.speaker = true;
        }

        if (word.start) word.start = parseFloat(word.start.toFixed(digits));
        if (word.end) word.end = parseFloat(word.end.toFixed(digits));

        if (currentParagraph) {
          if (!currentParagraph.start && word.start) {
            currentParagraph.start = word.start;
          }
          if (word.end) currentParagraph.end = word.end;
        }

        json.words.push(word);
        currentWord = word;
      },

      ontext: function ontext(text) {
        var word = currentWord;

        if (word.speaker) {
          currentParagraph.speaker = text.trim().replace(/[:|\[|\]]/g, '');
          if (!json.speakers.includes(currentParagraph.speaker)) json.speakers.push(currentParagraph.speaker);
        }

        word.text = text.trim();

        if (offsets) {
          word.startOffset = json.transcript.length;
          json.transcript += text;
          word.endOffset = json.transcript.length;
        }

        if (json.words.length === 0 || word !== json.words[json.words.length - 1]) json.words.push(word);
      },

      onclosetag: function onclosetag(name) {
        if (name === 'p') {
          if (offsets) currentParagraph.endOffset = json.transcript.length;
          currentParagraph = null;
        }
        currentWord = {};
      },

      onend: function onend() {
        json.words = json.words.filter(function (word) {
          return !!word.text;
        });
        resolve(json);
      },

      onerror: function onerror(error) {
        return reject(error);
      }
    }, {
      decodeEntities: true,
      lowerCaseTags: true,
      lowerCaseAttributeNames: true,
      recognizeSelfClosing: true
    });

    parser.write(html);
    parser.end();
  });
});