Rename and reorganise files

epiforecasts / scoringutils

Utilities for Scoring and Assessing Predictions

https://epiforecasts.io/scoringutils/

Other

48 stars 20 forks source link

Rename and reorganise files #746

Open nikosbosse opened 5 months ago

nikosbosse commented 5 months ago

[ ] convencience-functions.R should be renamed to transform_forecasts.R
[ ] set_forecast_unit() should be moved to a file get_forecast_unit.R (maybe forecast_unit.R?) together with get_forecast_unit().
[ ] correlations.R should be get_correlations.R
[ ] different get_-functions should be moved out of get_-functions.R
[ ] metrics-range.R --> metrics-interval_range.R
[ ] customise_metric() should maybe get its own file. Alternatively, the corresponding tests should be moved.
[ ] the current function validate_scores should really be assert_scores()

seabbs commented 5 months ago

Is this all done now?

nikosbosse commented 5 months ago

We did another set of renaming/restructuring files previously, this one is still open.

sbfnk commented 4 months ago

get_-functions.R also contains some assert_... and test_... functions which I think should be moved to one of the check-....R fiiles.

jamesmbaazam commented 4 months ago

@nikosbosse You could perhaps do this as a final step before release since it's not a functional change?

nikosbosse commented 1 month ago

Following one of Sam's suggestions, we should think about moving all the files related to one forecast class to one file (i.e. everything related to converting into a forecast and assertions and tests and checks etc.)

seabbs commented 1 month ago

I really like this from the view of thinking about implementing new classes as the process to do so is then copy a single file that you know contains everything you need to create and then work through doing so. I can see the argument that it might be hard to get an impression of the overall package structure though

jamesmbaazam commented 1 month ago

Following one of Sam's suggestions, we should think about moving all the files related to one forecast class to one file (i.e. everything related to converting into a forecast and assertions and tests and checks etc.)

Agreed. It makes it easier to find everything about the class in one place. See dist_spec.R for example.

jamesmbaazam commented 1 month ago

I really like this from the view of thinking about implementing new classes as the process to do so is then copy a single file that you know contains everything you need to create and then work through doing so. I can see the argument that it might be hard to get an impression of the overall package structure though

That's how I learnt to develop classes actually.