refactor/build: create more realistic mock data and store in the shared module

[jaguarfi]

The generateMockData is currently a big file that contains only data for one election. Making it difficult to know if we are testing the right features. I suggest the following changes:

• Split the file into different modules, e.g., one for Election creation, another for candidate creation, another for Party creation and so on.
• Use the faker internationalization features according to our data.
• Add perhaps environment variables to define if we want to create presidential elections, parliament elections, municipality elections, city council elections and some other ones that may reflect use cases for the openVAA.
• Try to use more realistic Question categories for the questions

[kaljarv]

Update:

• The data should also be converted into the canonical format consumed by LocalDataProvider.
• The data should be stored in the shared module for use in backend, frontend and e2e tests.
• Mock data might be merged with default data, which is introduced in commit 3a2e1a83674b8664f2a999a52dbe8b3c33af9c3c

See also this discussion.