[server] Add support for managing data sources and data sets

[jjmerchante]

This commit includes DataSource, DataSet and DataSourceType models and the initial migrations file as described in https://github.com/chaoss/grimoirelab-bestiary/issues/39

Every DataSet object must have a related project, a related data source, a category and can include some filters.

Every DataSource object must have a DataSourceType associated and an uri for the backend.

DataSourceType database must be filled with data with the supported data sources.

This commit also includes some methods at database level to:

• Find data sources
• Add data sources
• Delete data sources
• Find data sets
• Add data sets
• Date data sets
• Update data sets

This PR includes tests for each data base method

[jjmerchante]

I think we should use datasource instead of data source and the same with dataset. I know they aren't valid English words but in this case we are talking more about the objects we are storing. Not sure if this makes sense to you.

I had this discussion some time ago with Jesus and Manrique in a different context and we decided to go for data_source. Anyway it is something internal to the code and I don't see any problem going your way.

In commentaries I keep data source and data set, and I also keep class names as DataSource and DataSet. I rename only the attributes of the classes and any word written as data_source to datasource (and the same for data set)

[sduenas]

I think we should use datasource instead of data source and the same with dataset. I know they aren't valid English words but in this case we are talking more about the objects we are storing. Not sure if this makes sense to you.

I had this discussion some time ago with Jesus and Manrique in a different context and we decided to go for data_source. Anyway it is something internal to the code and I don't see any problem going your way.

In commentaries I keep data source and data set, and I also keep class names as DataSource and DataSet. I rename only the attributes of the classes and any word written as data_source to datasource (and the same for data set)

Yes, in the comments is maybe better to reference as data source and data set but in methods, parameters, etc better to have datasource and dataset because we consider it as a whole.

[jjmerchante]

Changes in the last push:

• Tests for models
• Changed all data_source and data_set to datasource and dataset in variables and function names
• Included some missing last_modified and operations tests
• Save the project and ecosystem whenever a project or a dataset is added or updated (I forgot to include tests related with this, I include them later)