kristjanjansen/aw_migrate

AutomatWeb to Drupal migration for Tartu University

Module for migrating Automatweb CMS data to Drupal

Installation:

1. Make sure you have Drush installed http://drupal.org/project/drush
2. Create empty database for Drupal
3. Download and install Drupal 7 either manually or using Drush
4. Create another database for Automatweb, and import existing data there
5. Add following snippet to sites/default/settings.php:

$databases['aw']['default'] = array (
  'driver' => 'mysql',
  'database' => 'your_automatweb_database_name',
  'username' => 'your_automatweb_database_username',
  'password' => 'your_automatweb_database_password',
);

...and run following commands inside your Drupal site root directory:

git clone https://github.com/kristjanjansen/aw_migrate.git sites/all/modules/aw_migrate
drush en aw_migrate -y
drush vset aw_domain http://your_automatweb_site_url_without_trailing_slash
drush mi --all

For big sites migrating everything in one go could be very time-consuming, it might be a good idea to run separate and/or partial migrations. Run drush mi --help to find out more.

aw_migrate's Issues

break aliasProcessor class to separate functions

Current monster switch construct is horrible.

Remove duplicate user names

Drupal requires both uid and username be unique. In some cases there could be duplicates in AW user names. Track and eliminate those (rename dupes as username_x?)

variable to control image and file copying

Running drush mi --all causes massive direct image/file copying from the source site. Add variable to control it (off by default)

Remove depencecy for migrate_extras

It seems we do need that shallow media support the module provides. At the moment.

Fix messy file and image saving

https://github.com/kristjanjansen/aw_migrate/blob/master/aw_migrate.inc#L138 serves two functions: define source URL for copy from and target directory structure to copy to. This creates messy directory tree and wrong filename.

separate _aw_migrate_clean_text to markup cleanup and alias detection functions

Also keep in mind #2

behaviour when threre's no aw document title

Strip a chunk from aw document content and strip tags. Something like

truncate_utf8(trim(decode_entities(strip_tags(check_markup($current_row->comment))), 29, TRUE);

Migrate menu structure

Very preliminary work is done but commented out https://github.com/kristjanjansen/aw_migrate/blob/master/aw_migrate.inc#L207

References:
http://drupal.org/node/592716
http://stackoverflow.com/questions/2829378/is-there-a-way-other-than-sql-to-get-the-mlid-for-a-given-nid-in-drupal

Remove table key hardcoding

Everythere where's

Database::getConnection('default', 'aw');

there should be something like

Database::getConnection('default', variable_get('aw_db_key', 'aw'));

Figure out what to do with ACL

Make quote replacement more robust

https://github.com/kristjanjansen/aw_migrate/blob/master/aw_migrate_utils.inc#L95 replaces « and » with " ". Make it robust to replace only [space]«[no_space]string etc.

Figure out what classes to migrate

https://github.com/kristjanjansen/aw_migrate/blob/master/ut_classes.txt

AW_VIDEO seems interesting - but is there any real content? There's also load of CRM stuff - whatever happens to that?

Language migration

This has two steps:

1. Migrate languages: aw.languages table to look up (figure out the statuses!), http://api.drupal.org/api/drupal/includes--locale.inc/function/locale_add_language/7. As there's no language target in Migrate APIs where lang migration should live? There''re some talk about MigrationBase descendant class but no docs http://drupal.org/node/1012818
2. Assign Drupal language ids / langcodes to content. As there's likely no direct mapping with aw lang id's, figure out how to make mapping possible.