Your question is a bit specific if you “only” want to automatically import some posts/pages. There are other ways to do this then using a XML export file.
If you have text-only posts, then you should use LOAD DATA INFILE. At first you have to export your posts.
global $wpdb, $wp_filesystem;
$tables = array(
'posts' => array( 'posts', 'postmeta' ),
'comments' => array( 'comments', 'commentmeta' ),
'terms' => array( 'terms', 'term_taxonomy', 'term_relationships' ),
'users' => array( 'user', 'usermeta' ),
'links' => array( 'links' ),
'options' => array( 'options' ),
'other' => array(),
// for multiside
'multiside' => array( 'blogs', 'signups', 'site', 'sitemeta', 'sitecategories', 'registration_log', 'blog_versions' )
);
$exports = array( 'posts', 'comments', 'users' );
$exportdir = TEMPLATEPATH . '/export';
if ( ! is_dir( $exportdir ) ) {
$mkdir = wp_mkdir_p( $exportdir );
if ( false == $mkdir || ! is_dir( $exportdir ) )
throw new Exception( 'Cannot create export directory. Aborting.' );
}
// empty the export dir else MySQL throws errors
$files = glob( $exportdir . '/*' );
if ( ! empty( $files ) ) {
foreach( $files as $file )
unlink( $file );
}
foreach ( $exports as $export ) {
if ( ! isset( $tables[$export] ) )
continue;
if ( ! empty( $tables[$export] ) ) {
foreach ( $tables[$export] as $table ) {
$outfile = sprintf( '%s/%s_dump.sql', $exportdir, $table );
$sql = "SELECT * FROM {$wpdb->$table} INTO OUTFILE '%s'";
$res = $wpdb->query( $wpdb->prepare( $sql, $outfile ) );
if ( is_wp_error( $res ) )
echo "<p>Cannot export {$table} into {$outfile}</p>";
}
}
}
This will create a directory in your theme folder (be sure it is writeable!) and export the posts and comments (with it’s meta) into dump files. Use the array export to define what you want to export. I grouped the most things more or less logical (if you want to export the posts, than you should also export postsmeta and so on).
The benefit of this solution is, with the SELECT statement you can define particular stuff (e.g. only posts from a special category or only pages or only trashed posts).
Now you want to import this stuff in a new blog
global $wpdb;
$exportdir = TEMPLATEPATH . '/export';
$files = glob( $exportdir . '/*_dump.sql' );
foreach ( $files as $file ) {
preg_match( '#/([^/]+)_dump.sql$#is', $file, $match );
if ( ! isset( $match[1] ) )
continue;
$sql = "LOAD DATA LOCAL INFILE '%s' INTO TABLE {$wpdb->$match[1]};";
$res = $wpdb->query( $wpdb->prepare( $sql, $file ) );
if ( is_wp_error( $res ) )
echo "<p>Cannot import data from file {$file} into table {$wpdb->$match[1]}</p>";
}
This solution is good if the posts did not contain any attachments like images. Another problem is, no users and no categories will be imported. Be sure both are created befor the import starts (or include users and categories in your export). It is a very rough method to import things, it will override existing stuff!
If you want to export the attachments also, you have to do a bit more work.
(Sidenote: Please read the complete answer and the Very Last Words at the end! This topic is not for beginners and I do not write a warning at every risky line of code)
The WordPress Importer Plugin seems to be a good way to import the whole stuff and automatically import/download the attachments. So let’s have a look what this plugin will do.
At first the plugin ask for a XML file to upload. Then it parse the XML file and ask for an author mapping and if the attachments should be downloaded or not.
For an automatically run of the plugin we need to change some things. At first we have to skip the upload process. Thats quite easy because you can bundle the XML file with the theme and you know where the XML file is. Then we have to skip the questions that appears after uploading the XML file. We can predefine our own values and pass them to the import process.
Start with a copy of the plugin. Create a directory in your theme like autoimport and copy the files wordpress-importer.php and parsers.php to it. It is a good idea to rename the file wordpress-importer.php to something like autoimporter.php. In your theme function add a function call to trigger the automated impoprt
/**
* Auto import a XML file
*/
add_action( 'after_setup_theme', 'autoimport' );
function autoimport() {
// get the file
require_once TEMPLATEPATH . '/autoimport/autoimporter.php';
if ( ! class_exists( 'Auto_Importer' ) )
die( 'Auto_Importer not found' );
// call the function
$args = array(
'file' => TEMPLATEPATH . '/autoimport/import.xml',
'map_user_id' => 1
);
auto_import( $args );
}
At first we setup some arguments. The first thing is the complete path to the XML file. The second one is the ID of an existing user. We need this user for author mapping, this is the user where all posts will be mapped to when no new authors should be created.
Now we have to understand how the plugin works. Open your renamed plugin file and scroll down to the end. There is a function wordpress_importer_init() and an action call. Remove both, it’s not longer needed. Now go to the top of the file and remove the plugin header (the comment at the beginning of the file). After that, rename the class WP_Importer to something like Auto_Importer, do not forget to adjust the function_exists statement and the first method WP_Importer (this is the constructor in PHP4 style).
Later we will pass the XML file direct to the class constructor, modify the first method to this
var $xmlfile="";
var $map_user_id = 0;
function Auto_Importer( $args ) {
if ( file_exists( $args['file'] ) ) {
// for windows systems
$file = str_replace( '\\', "https://wordpress.stackexchange.com/", $args['file'] );
$this->xmlfile = $file;
}
if ( isset( $args['map_user_id'] ) )
$this->map_user_id = $args['map_user_id'];
}
Now we have a to remove and modify some methods inside the class. The first method is the dispatch() method. This method tells you how the class works. It do three steps. At first upload the XML file, then process it and at last import the data.
Case zero is the first step, it is the greeting. This is the part that you see if you call the import at the first time. It will ask for a file to upload. Case two handles the upload and display a form for the import options. Case three finally do the import. In other words: the first two steps only ask for data we can provide ourself. We only need step 3 (case 2) and have to provide the data asked in step one and two.
In step two you see a function call to wp_import_handle_upload(). This function setup some informations about the xml file. We cannot use this function anymore because we haven’t uploaded a file. So we have to copy and modify the function. Create a new method within the class
function import_handle_upload() {
$url = get_template_directory_uri() . str_replace( TEMPLATEPATH, '', $this->xmlfile );
$type="application/xml"; // we know the mime type of our file
$file = $this->xmlfile;
$filename = basename( $this->xmlfile );
// Construct the object array
$object = array( 'post_title' => $filename,
'post_content' => $url,
'post_mime_type' => $type,
'guid' => $url,
'context' => 'import',
'post_status' => 'private'
);
// Save the data
$id = wp_insert_attachment( $object, $file );
// schedule a cleanup for one day from now in case of failed import or missing wp_import_cleanup() call
wp_schedule_single_event( time() + DAY_IN_SECONDS, 'importer_scheduled_cleanup', array( $id ) );
return array( 'file' => $file, 'id' => $id );
}
And replace the function call $file = wp_import_handle_upload(); in the method handle_upload() with our new method $file = $this->import_handle_upload();
We replaced now the upload process with our own file (that should already exists). Go on and remove more unneeded methods. The methods gereet(), header() and footer() are not longer needed (header and footer only print some text) and can be removed from the class. In the dispatch() method remove the calls to this methods ($this->header() and $this->footer()).
The first step is done, now we have to care about the second step, the import options. The import options ask if it should be allowed to download the attachments and mapping the authors.
The first part is easy. Set to true if the attachments should be downloaded or false if not. The author mapping is a bit more complicated. If it is allowed to create new users (the authors from the import file), create them. If not, assign the postss to an existing user. This is been done in the method get_author_mapping(). We have to replace the $_POST data with existing data. Here we need a simple solution, so we simply map all new authors to an existing one if it is not allowed to create new users. Or simply create all new users. In the second case, be sure all new users are dummy users. If not, everytime you import them, they get an email with login and password to the new blog!! I do not explain every line of code, here is the complete rewritten method
function get_author_mapping( $map_users_id ) {
if ( empty( $this->authors ) )
return;
$create_users = $this->allow_create_users();
foreach ( (array) $this->authors as $i => $data ) {
$old_login = $data['author_login'];
// Multisite adds strtolower to sanitize_user. Need to sanitize here to stop breakage in process_posts.
$santized_old_login = sanitize_user( $old_login, true );
$old_id = isset( $this->authors[$old_login]['author_id'] ) ? intval($this->authors[$old_login]['author_id']) : false;
if ( ! $create_users ) {
$user = get_userdata( intval($map_users_id) );
if ( isset( $user->ID ) ) {
if ( $old_id )
$this->processed_authors[$old_id] = $user->ID;
$this->author_mapping[$santized_old_login] = $user->ID;
}
} else if ( $create_users ) {
if ( ! empty($this->authors[$i]) ) {
$user_id = wp_create_user( $this->authors[$i]['author_login'], wp_generate_password() );
} else if ( $this->version != '1.0' ) {
$user_data = array(
'user_login' => $old_login,
'user_pass' => wp_generate_password(),
'user_email' => isset( $this->authors[$old_login]['author_email'] ) ? $this->authors[$old_login]['author_email'] : '',
'display_name' => $this->authors[$old_login]['author_display_name'],
'first_name' => isset( $this->authors[$old_login]['author_first_name'] ) ? $this->authors[$old_login]['author_first_name'] : '',
'last_name' => isset( $this->authors[$old_login]['author_last_name'] ) ? $this->authors[$old_login]['author_last_name'] : '',
);
$user_id = wp_insert_user( $user_data );
}
if ( ! is_wp_error( $user_id ) ) {
if ( $old_id )
$this->processed_authors[$old_id] = $user_id;
$this->author_mapping[$santized_old_login] = $user_id;
} else {
printf( __( 'Failed to create new user for %s. Their posts will be attributed to the current user.', 'wordpress-importer' ), esc_html($this->authors[$old_login]['author_display_name']) );
if ( defined('IMPORT_DEBUG') && IMPORT_DEBUG )
echo ' ' . $user_id->get_error_message();
echo '<br />';
}
}
// failsafe: if the user_id was invalid, default to the current user
if ( ! isset( $this->author_mapping[$santized_old_login] ) ) {
if ( $old_id )
$this->processed_authors[$old_id] = (int) get_current_user_id();
$this->author_mapping[$santized_old_login] = (int) get_current_user_id();
}
}
}
There is some work left to do. Adding a function auto_import() first
function auto_import( $args ) {
$defaults = array( 'file' => '', 'map_user_id' => 0);
$args = wp_parse_args( $args, $defaults );
$autoimport = new Auto_Importer( $args );
$autoimport->do_import();
}
Place this function after the class. This function miss some error handling and checking (e.g. for an empty file argument).
If you now run the class, you got a lot of error messages. The first one is, that the class is missing. This is because there is a if statement at the beginning.
if ( ! defined( 'WP_LOAD_IMPORTERS' ) )
return;
We have to remove it, otherwise the file would not be parsed completely. Than there are some functions that are not loaded at this point. We have to include some files.
$required = array(
'post_exists' => ABSPATH . 'wp-admin/includes/post.php',
'wp_generate_attachment_metadata' => ABSPATH . 'wp-admin/includes/image.php',
'comment_exists' => ABSPATH . 'wp-admin/includes/comment.php'
);
foreach ( $required as $func => $req_file ) {
if ( ! function_exists( $func ) )
require_once $req_file;
}
Basically thats all. I test this on a local installation with the test data XML from WordPress. It work for me but it is not a perfect solution for production!
And some last words on setting up some options. There are two options that can be modified by a filter:
add_filter( 'import_allow_create_users', function() { return false; } );
add_filter( 'import_allow_fetch_attachments', '__return_false' );
I think I do not have to explain it. Put this filters in your functions.php and setup true or false (first one is PHP5.3 style, second is WP style).
Very Last Words
I put alltogether in this gist. Use it at your own risk! I’m not responsible for anything!. Please have a look at the files in the gist, I did not explain every little step here.
Thinks I haven’t done: Set a value e.g. in the (theme) options after importing. Else the import starts every time the theme will be activated.
Maybe I will work on it in the future, clean up some things and run more tests on it.