doctoc/lib/get-html-headers.js

'use strict';

var htmlparser = require('htmlparser2')
  , md         = require('@textlint/markdown-to-ast');

function addLinenos(lines, headers) {
  var current = 0, line;

  return headers.map(function (x) {
    for (var lineno = current; lineno < lines.length; lineno++) {
      line = lines[lineno];
      if (new RegExp(x.text[0]).test(line)) {
        current = lineno;
        x.line = lineno;
        x.name = x.text.join('');
        return x
      }
    }

    // in case we didn't find a matching line, which is odd,
    // we'll have to assume it's right on the next line
    x.line = ++current;
    x.name = x.text.join('');
    return x
  })
}

function rankify(headers, max) {
  return headers
    .map(function (x) {
      x.rank = parseInt(x.tag.slice(1), 10);
      return x;
    })
    .filter(function (x) {
      return x.rank <= max;
    })
}

var go = module.exports = function (lines, maxHeaderLevel) {
  var source = md.parse(lines.join('\n'))
    .children
    .filter(function(node) {
      return node.type === md.Syntax.HtmlBlock || node.type === md.Syntax.Html;
    })
    .map(function (node) {
      return node.raw;
    })
    .join('\n');

  //var headers = [], grabbing = null, text = [];
  var headers = [], grabbing = [], text = [];

  var parser = new htmlparser.Parser({
    onopentag: function (name, attr) {
      // Short circuit if we're already inside a pre
      if (grabbing[grabbing.length - 1] === 'pre') return;

      if (name === 'pre' || (/h\d/).test(name)) {
        grabbing.push(name);
      }
    },
    ontext: function (text_) {
      // Explicitly skip pre tags, and implicitly skip all others
      if (grabbing.length === 0 ||
          grabbing[grabbing.length - 1] === 'pre') return;

      text.push(text_);
    },
    onclosetag: function (name) {
      if (grabbing.length === 0) return;
      if (grabbing[grabbing.length - 1] === name) {
        var tag = grabbing.pop();
        headers.push({ text: text, tag: tag });
        text = [];
      }
    }
  },
  { decodeEntities: true })

  parser.write(source);
  parser.end();

  headers = addLinenos(lines, headers)
  // consider anything past h4 to small to warrant a link, may be made configurable in the future
  headers = rankify(headers, maxHeaderLevel);
  return headers;
}