Jekyll WordPress import post processing gulpfile for Hugo

gulpfile.js

// This is a throwaway / single-use gulpfile I used to process and tidy up the
// files resulting from using Jekyll's WordPress exporter. I wanted to move from
// WordPress to Hugo with a relatively large number of posts; this gulpfile let
// me quickly process all the content files and get on with building the site.

var
  gulp     = require('gulp'),
  replace  = require('gulp-replace'),
  rename   = require('gulp-rename'),
  entities = require('gulp-html-entities'),
  filter   = require('gulp-filter'),
  chalk    = require('chalk'),
  through  = require('through-gulp');

function fileName(file) {
  var pathParts = file.path.split('/');
  return pathParts[pathParts.length - 1];
}

gulp.task('posts', function() {
  return gulp.src('./_posts/*.markdown')

  // get rid of junk files and warn about short ones
  .pipe(filter(function(file) {
    if (file.contents.length < 2000) {
      var name = fileName(file);
      if (file.contents.length < 1500) {
        console.log(chalk.red(
          'Removed junk file ' + name +' (' + file.contents.length + ' chars)'
        ));
        return false;
      } else {
        console.log(chalk.yellow(
          name + ' is quite short (' + file.contents.length + ' chars)'
        ));
      }
    }
    return true;
  }))

  // organise into year-month folders
  .pipe(rename(function(path) {
    var year, month;
    year = path.basename.slice(0,4);
    month = path.basename.slice(5,7);
    path.dirname += '/' + year + '/' + month + '/';
    path.extname = '.md';
  }))

  // save slug in front matter
  .pipe(through.map(function(file) {
    var name = fileName(file),
      slug = name.slice(11, name.length - 3);
    file.contents = new Buffer(file.contents.toString().replace(
      '---\nlayout: post\n',
      '---\nlayout: post\nslug: ' + slug + '\n'
    ));
    return file;
  }))

  // tidy up front matter
  .pipe(replace(/(wordpress_id|wordpress_url|author_login|author_email|author_url): .*?\n/g, ''))
  .pipe(replace(/\npublished: .*\n/, '\n'))
  .pipe(replace(/\nstatus: .*\n/, '\n'))
  .pipe(replace(/author:\n(.*?\n){4}/, 'author: Hugh Grigg\n'))
  .pipe(replace(/.*?: !binary \|-\n  .*?\n/g, ''))
  .pipe(replace(/\n(title|excerpt):[\s!]+(.*)\n/g, '\n$1: >\n  $2\n'))
  .pipe(replace(/\ntitle: >\n  '(.*?)'/, '\ntitle: >\n  $1'))
  .pipe(replace(/\nexcerpt:/, '\ndescription:'))

  // decode wordpress' html entity encoding
  .pipe(entities('decode'))

  // fix urls
  .pipe(replace('http://eastasiastudent.net/wp-content/uploads/', '/img/'))
  .pipe(replace('http://eastasiastudent.net/', '/'))

  // apply some html -> markdown replacements
  .pipe(replace(/<br ?\/?>\n/g, '\n'))
  .pipe(replace(/<a href="(.*?)">(.*?)<\/a>/g, '[$2]($1)'))
  .pipe(replace(/<a href="(.*?)" title="(.*?)">(.*?)<\/a>/g, '[$3]($1 "$2")'))
  .pipe(replace(/<a title="(.*?)" href="(.*?)">(.*?)<\/a>/g, '[$3]($2 "$1")'))
  .pipe(replace(/<em>(.*?)<\/em>/g, '_$1_'))
  .pipe(replace(/<strong>(.*?)<\/strong>/g, '*$1*'))
  .pipe(replace(/<h([0-9])>(.*?)<\/h([0-9])>/g, '\n<h$1>$2</h$3>\n'))
  .pipe(replace(/<(ul|ol|p|h[0-9]|blockquote)/g, '\n<$1'))
  .pipe(replace(/<\/(ul|ol|p|h[0-9]|blockquote)>/g, '\n</$1>\n'))
  .pipe(replace(/<h1>([\s\S]*?)<\/h1>/g, '# $1'))
  .pipe(replace(/<h2>([\s\S]*?)<\/h2>/g, '## $1'))
  .pipe(replace(/<h3>([\s\S]*?)<\/h3>/g, '### $1'))
  .pipe(replace(/<h4>([\s\S]*?)<\/h4>/g, '#### $1'))
  .pipe(replace(/<h5>([\s\S]*?)<\/h5>/g, '##### $1'))
  .pipe(replace(/<h6>([\s\S]*?)<\/h6>/g, '###### $1'))
  .pipe(replace(/<p( style=".*?")?>([\s\S]*?)<\/p>/g, '\n$2\n'))
  .pipe(replace(/<li>(.*?)<\/li>/g, ' - $1'))
  .pipe(replace(/<\/?[ou]l( style=.*?)?>/g, ''))
  .pipe(replace(/\n\s?<\/p>\n?/g, ''))
  .pipe(replace(/\n\s+\n/g, '\n\n'))

  // apply some custom hugo tags
  .pipe(replace(/<span class="chinese"( lang="zh")?>(.*?)<\/span>/g, '{{< hanzi >}}$2{{< /hanzi >}}'))
  .pipe(replace(/<span class="hanzi"( lang="zh")?>(.*?)<\/span>/g, '{{< hanzi >}}$2{{< /hanzi >}}'))
  .pipe(replace(/<span class="reading">(.*?)<\/span>/g, '{{< reading >}}$1{{< /reading >}}'))
  .pipe(replace(/<span class="pinyin">(.*?)<\/span>/g, '{{< reading >}}$1{{< /reading >}}'))
  .pipe(replace(/<span class="gloss">(.*?)<\/span>/g, '{{< gloss >}}$1{{< /gloss >}}'))
  .pipe(replace(/<span class="note">(.*?)<\/span>/g, '{{< note >}}$1{{< /note >}}'))
  .pipe(replace(/<span style="font-size: smaller;">(.*?)<\/span>/g, '{{< note >}}$1{{< /note >}}'))
  .pipe(replace(/<p style="font-size: smaller;">(.*?)<\/p>/g, '\n{{< note >}}$1{{< /note >}}\n'))
  .pipe(replace(/<span style="color: #888888;">(.*?)<\/span>/g, '{{< note >}}$1{{< /note >}}'))
  .pipe(replace(/<p style="font-size: larger;">(.*?)<\/p>/g, '{{< larger >}}$1{{< /larger >}}'))

  // tidy up some leftovers
  .pipe(replace(/<p style="font-size: larger;">/, ''))
  .pipe(replace(/<span style="line-height: ?(.*?);?">(.*?)<\/span>/g, '$2'))
  .pipe(replace(/ ?wp-image-[0-9]+/g, ''))
  .pipe(replace(/ (width|height)="[0-9]+"/g, ''))
  .pipe(replace(/\n\n\n/g, '\n\n'))
  .pipe(replace(/\n\n$/, '\n'))

  // write
  .pipe(gulp.dest('./processed/'));
});

package.json

{
  "name": "post-process-jekyll-wordpress",
  "version": "0.0.0",
  "description": "",
  "main": "gulpfile.js",
  "author": "",
  "license": "BSD-2-Clause",
  "devDependencies": {
    "gulp-rename": "~1.2.2",
    "gulp-replace": "~0.5.3",
    "gulp": "~3.8.11",
    "gulp-html-entities": "0.0.3",
    "gulp-filter": "~2.0.2",
    "chalk": "~1.0.0",
    "through-gulp": "~0.3.8"
  }
}