Fix database schema and remove Scrapy scheduler fix

[LuisDuarte1]

Per PR #99:

The root cause is that the relation course <--> course_unit is incorrect. It should be a many-to-many relationship instead of a 1-to-many relationship. To quickly fix this without redesigning the schema, we duplicate the course_unit but to a different course (ofc). This causes everything that depends on course units (metadata, schedule) to be duplicated hence the need to disable the deduplication on the scrappy scheduler side.

To disable the scheduler, this PR also bumps Scrappy to 2.7.0, where this feature was added.

This nasty fix should be removed ASAP when we solve the schema issue and use the meta no_filter on course_unit requests instead.

For example, this Course Unit (Mecânica dos Solos) is shared between L.EMG and L.EC (https://sigarra.up.pt/feup/pt/ucurr_geral.ficha_uc_view?pv_ocorrencia_id=519393)