Major simplification of event_* tables

[hugotiburtino]

Currently, the database structure regarding event is absurdly complex.

For example, to get the the parent of a checkout revision event we have to go through the event_parameter table and then to the event_parameter_uuid, besides that we have to see in the event_parameter_string the corresponding reason.

A much simpler way of dealing with it would be to have nullable columns in the event_log table. Or we could have just one table like event_data and have all info there. Anyway, we should figure out a simpler solution, of course with the corresponding db migration.

[kulla]

The following file contains all current descriptions of events:

https://github.com/serlo/api.serlo.org/blob/af1639f0a6e6e2ebca63083016cfa85ba103dfec/packages/server/src/schema/events/event.ts#L439-L533

There are at most two extra uuid_paramaters and at most one string parameter needed. A first easy simplification is to merge all extra event_* tables into one with the parameters:

• uuid_parameter: Int or null
• uuid_paramater2: Int or null
• string_parameter: String or null

In case we need more sematic we can add uuid_parameter_type or uuid_parameter_type2... => However the semantics is set by https://github.com/serlo/api.serlo.org/blob/af1639f0a6e6e2ebca63083016cfa85ba103dfec/packages/server/src/schema/events/event.ts#L439-L533 in the conversion of the events

[kulla]

A next step might be to store JSON in the event table. Also the following renamings can be done:

• Rename event_log into event
• Rename event into event_type