@sugarcube/plugin-tika/_dist/utils.js

Parse files and metadata using Tika.

"use strict";

Object.defineProperty(exports, "__esModule", {
  value: true
});
exports.entity = exports.safeExtract = exports.extract = void 0;

var _fp = require("lodash/fp");

var _moment = _interopRequireDefault(require("moment"));

var _utils = require("@sugarcube/utils");

var _contentTypes = _interopRequireDefault(require("./content-types"));

function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; }

const contentTypes = {
  "image/jpeg": _contentTypes.default.imageJpeg,
  "application/pdf": _contentTypes.default.applicationPdf
};
const extract = _utils.extract; // safeExtract is useful for fetching links, that might otherwise throw silly
// errors. E.g. LinkedIn return error 999 and extract throws up on that.

exports.extract = extract;

const safeExtract = location => extract(location).catch(() => ["", {}]);

exports.safeExtract = safeExtract;
const entity = (0, _fp.curry)((unit, text, meta) => {
  const created = (0, _moment.default)(meta.date).utc().toDate();
  const idFields = (0, _fp.reduce)((memo, f) => {
    if ((0, _fp.has)(f, meta)) {
      return (0, _fp.concat)(memo, f);
    }

    return memo;
  }, unit._sc_id_fields || [], ["Content-Type"]);
  const contentFields = (0, _fp.reduce)((memo, f) => {
    if ((0, _fp.has)(f, meta)) {
      return (0, _fp.concat)(memo, f);
    }

    return memo;
  }, unit._sc_content_fields || [], ["title"]);
  const contentType = (0, _fp.has)(meta["Content-Type"], contentTypes) ? contentTypes[meta["Content-Type"]] : _fp.identity;
  return contentType((0, _fp.mergeAll)([unit, meta, {
    text
  }, {
    _sc_pubdates: {
      tika: created
    },
    _sc_id_fields: idFields,
    _sc_content_fields: contentFields
  }]));
});
exports.entity = entity;