As stated in unify_data.py as a TODO: we should add tests for this function.
Currently, it seems to work based on my initial implementation and usage. But all teams were in the same situation right now, with no history yet. Only a few teams have had a roster difference between the three rankings. There will be other situations to test for sure, such as when teams revert to an old roster again, or when there is a new team with the same name as an old historic team but a totally different roster.
It would be good to create example dfs for all kinds of situations, and test them with different versions of teamname_mapping and rosters. (This is a new team, team is renamed, roster changed with 1 person, roster changed with 3 people, old name is reused but roster different, uppercase/lowercase of playernames, etc)
As stated in
unify_data.pyas a TODO: we should add tests for this function.Currently, it seems to work based on my initial implementation and usage. But all teams were in the same situation right now, with no history yet. Only a few teams have had a roster difference between the three rankings. There will be other situations to test for sure, such as when teams revert to an old roster again, or when there is a new team with the same name as an old historic team but a totally different roster.
It would be good to create example dfs for all kinds of situations, and test them with different versions of teamname_mapping and rosters. (This is a new team, team is renamed, roster changed with 1 person, roster changed with 3 people, old name is reused but roster different, uppercase/lowercase of playernames, etc)