Refactor SCORM database schema to improve performance

The current structure of the scorm_scoes_track table:

scorm_scoes_track:
id | userid | scormid | scoid | attempt | element | value | timemodified

is problematic for reporting purposes because there is no id defining a single attempt so you need to group by userid, scormid, scoid and attempt to extract that info.

Also the indexes on this table can become huge because it is not normalised properly.

We propose splitting into separate tables:

scorm_scoes_track:
id | attemptid | elementid | value | timemodified

scorm_scoes_attempt:
id | userid | scormid | scoid | attempt

scorm_scoes_element:
id | element