Change batch aggregation to select only time periods that are being aggregated

[eiffel777]

Currently, when aggregating in batches, the data loaded into the temporary table and aggregated is between the min and max time periods for the batch and it may be the case that not all time periods in that range need to be aggregated. This PR changes to only loading the time periods that are being aggregated into the batch. It takes the WHERE clause in the aggregation and rewrites it to be multiple clauses, one for each time period be aggregated.

For example, if the original WHERE clause is:

task.start_day_id <= :period_end_day_id AND task.end_day_id >= :period_start_day_id AND task.is_deleted = 0

and we are aggregating 2 periods with period_start_day_id/period_end_day_id values of 20200101/20200131 and 20200201/20200229, the WHERE clause will be rewritten to:

(task.start_day_id <= :period_end_day_id_0 AND task.end_day_id >= :period_start_day_id_0 AND task.is_deleted = 0) 
OR 
(task.start_day_id <= :period_end_day_id_1 AND task.end_day_id >= :period_start_day_id_1 AND task.is_deleted = 0)

This speeds up the batch temp table creation and aggregation when there is a large gap in the time periods being aggregated in a batch.

Tests performed

Tested in docker and on xdmod-dev. New unit tests are added too.

Checklist:

• The pull request description is suitable for a Changelog entry
• The milestone is set correctly on the pull request
• The appropriate labels have been added to the pull request