Since the existing parser (version 1) only deals with 2 tables (ptevent and pteventclass), the code must be re-written / extended to support for multiple tables. This means that each of the 8 modules need to be changed accordingly. These require the following stages / milestones:
1. The current system (version 2) can now enable the user to type in the query question from the interface module.
2. The syntactic parser then splits the question up into each of the above part , and generates separate objects (from patient_class, answer_class and condition_class) accordingly.
3. These objects are then passed to the SQL generator that is responsible to convert the question into the corresponding SQL for the underlying databases. Note that there are 2 sub-tasks here: the realtime database and the archive database.
4. The results are then passed to the the Response Generator that is responsible to display the query results.
At the moment, stage 2 is completed. Stage 1 is under reviewed (see below). Stage 3 is in progress.