Revise cohort structure inside community

[davidorme]

Description

This PR:

• replaces the community.cohort_data: dict[str, NDArray] dictionary of arrays with a new dataclass Cohorts.
• updates the test suite to add some simple testing of the Cohorts constructor
• updates the rest of the test suite and the code to the new structure.

The main reasons are:

• more consistent attribute access within model rather than a mix of dataclass dot notation and dictionary keys.
• ability to extend the functionality of the Cohorts dataclass with shared functionality (see Shared base classes for demography classes #338).

Fixes #336

Type of change

• New feature (non-breaking change which adds functionality)
• Optimization (back-end change that speeds up the code)
• Bug fix (non-breaking change which fixes an issue)

Key checklist

• Make sure you've run the pre-commit checks: $ pre-commit run -a
• All tests pass: $ poetry run pytest

Further checks

• Code is commented, particularly in hard-to-understand areas
• Tests added that prove fix is effective or that feature works

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 94.89%. Comparing base (1f315ba) to head (c43bedf).
Report is 298 commits behind head on develop.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #340      +/-   ##
===========================================
- Coverage    95.29%   94.89%   -0.40%     
===========================================
  Files           28       34       +6     
  Lines         1720     2549     +829     
===========================================
+ Hits          1639     2419     +780     
- Misses          81      130      +49     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[MarionBWeinzierl]

The changes look ok to me (with the explanations provided), and this seems a cleaner way to implement it.