Wordpress Exporter via Node

parse_articles.js

// General exporter of WP exports into JSON objects. From that, we can build models to save to the DB.
// In usage at shell...
// cat file_of_wp_export.xml | node parse_articles.js >> json_output_file.json

/* Module Includes */
var xml2js = require('xml2js'), // Parser that reads in XML and converts it to JSON
	YUI = require('yui3').YUI,  // The YUI library
	_ = require('underscore');  // Utility belt of JS functions

var data = '', file = '', arg = '';

function postMeta(arg, key) {
	var result = '';
	_.each(arg, function (data) {
		if (data['wp:meta_key'] == key) {
			result = data['wp:meta_value'];
		}
	});

	return result;
}

process.stdin.resume();
process.stdin.setEncoding('utf8');

process.stdin.on('data', function (chunk) {
	file += chunk.toString();
});

process.stdin.on('end', function () {
	var parser   = new xml2js.Parser();

	parser.parseString(file, function (err, content) {
		YUI().use("dataschema-json", function(Y) {
			var

			// Articles Layout
			article_schema = {
				resultListLocator: "channel.item",
				resultFields: [
					{key:"title"},
					{key:"wp_post_id", locator:"wp:post_id"},
					{key:"post_type", locator:"wp:post_type"},
					{key:"created", locator:"pubDate"},
					{key:"sub_title", locator:"wp:postmeta", parser: function (arg) { return postMeta(arg, '_aioseop_description')} },
					{key:"blurb", locator:"excerpt:encoded"},
					{key:"body", locator:"content:encoded"},
					{key:"date_active", locator:"wp:post_date"},
					{key:"slug", locator:"wp:post_name"},
					{key:"author", locator:"dc:creator"},
					{key:"flag_active", locator:"wp:status", parser: function (arg) {
						switch (arg) {
							case 'publish':
								return 1;
							case 'draft':
							case 'pending':
							case 'future':
							case 'private':
							default:
								return 0;
						}
					}},

					{key:"meta_keywords", locator:"wp:postmeta", parser: function (arg) { return postMeta(arg, '_aioseop_keywords')} },
					{key:"meta_description", locator:"wp:postmeta", parser: function (arg) { return postMeta(arg, '_aioseop_description')} },

					// If trying to write the JSON version of articles out, Comments appears as [Object].
					// Having to do a custom parser to setup the data as needed for importing.
					{key:"Comments", locator:"wp:comment", parser: function (arg) {
						var comments = [];
						_.each(arg, function (data) {
							var comment = {};
							if (data !== undefined) {
								comment.wp_comment_id = data['wp:comment_id'];
								comment.name          = data['wp:comment_author'];
								comment.email         = data['wp:comment_author_email'];
								comment.comment       = data['wp:comment_content'];
								comment.status        = data['wp:comment_approved'];
								comment.created       = data['wp:comment_date'];
								comments.push(comment);
							}
						});

						return JSON.stringify(comments);
					}},

					// Same as comments here
					{key:"Channels", locator: "category", parser: function (arg) {
						var categories = [];
						_.each(arg, function (data) {
							var category = {};
							if (data.hasOwnProperty('@')) {
								if (data['@'].domain == 'category') {
									if (data['@'].hasOwnProperty('nicename')) {
										category.url_name     = data['@'].nicename;
										category.channel_name = data['#'];
									} else {
										category.url_name     = data['#'];
									}
									categories.push(category);
								}
							}
						});

						return JSON.stringify(categories);
					}},

					// Same as comments here
					{key:"Tags", locator: "category", parser: function (arg) {
						var tags = [];
						_.each(arg, function (data) {
							var tag = {};
							if (data.hasOwnProperty('@')) {
								// This may change based on WP version. If WP < 3, use this. If not, domain should be "tag".
								if (data['@'].domain == 'post_tag') {
									if (data['@'].hasOwnProperty('nicename')) {
										tag.slug = data['@'].nicename;
										tag.name = data['#'];
									} else {
										tag.slug = data['#'];
									}
									tags.push(tag);
								}
							}
						});

						return JSON.stringify(tags);
					}}
				]
			},

			// Pictures are in the same items as articles, just defined by having an attachment URL
			attachment_schema = {
				resultListLocator: "channel.item",
				resultFields: [
					{key: "attachment_parent", locator:"wp:post_parent"},
					{key: "attachment", locator:"wp:attachment_url"}
				]
			},

			// Lives at the top of the WP XML file
			category_schema = {
				resultListLocator: "channel.wp:category",
				resultFields: [
					{key:"url_name",        locator:"wp:category_nicename"},
					{key:"channel_name",    locator:"wp:cat_name"},

					// When mapping these off in shell, you'll have to do recursive runs to find all parents until no parents are left.
					// Note that parent name is based off of wp:cat_name
					{key:"parent_category", locator:"wp:category_parent"}
				]
			},

			// Lives at the top of the WP XML file
			tag_schema = {
				resultListLocator: "channel.wp:tag",
				resultFields: [
					{key:"name", locator:"wp:tag_name"},
					{key:"slug", locator:"wp:tag_slug"}
				]
			},

			// Lives within all the articles, no clear definition
			// In WP > 3, all authors are defined at the top of the file.
			// However, dc:creator is still used within the post definition, which uses the nicename of an author instead of an ID or set of IDs. Strange.
			author_schema = {
				resultListLocator: "channel.item",
				resultFields: [
					{key: "name", locator:"dc:creator"},
					{key: "wp_post_id", locator:"wp:post_id"}
				]
			},

			old_url_schema = {
				resultListLocator: "channel.item",
				resultFields: [
					{key:"title"},
					{key:"wp_post_id", locator:"wp:post_id"},
					{key:"slug", locator:"wp:post_name"},
					{key:"old_url", locator:"wp:old_url"}
				]
			};


			// Depending on what you want to output, change the schema here. You could also merge some of schemas together if you want just one big JSON file with everything.
			// If think it's better to keep schemas modular in case you need to import something before something else.
			data_out = Y.DataSchema.JSON.apply(article_schema, content);
			console.log(JSON.stringify(data_out));
		});
	});
});