Comparing output for `current_period` and `past` values of `PREFER_DATES_FROM` setting

I observed one interesting behavior where for some cases the current_period and past values of PREFER_DATES_FROM return the same result but it's different for others. For example, suppose today is Thursday 23th, then

text = "Thursday"
print(dateparser.parse(text, settings={'PREFER_DATES_FROM': 'current_period'}))
>>> 2021-12-23 00:00:00
print(dateparser.parse(text, settings={'PREFER_DATES_FROM': 'past'}))
>>> 2021-12-16 00:00:00

This is the expected behavior, where for current_period we get today's date (at the time of writing) i.e. 23rd, and for past its last Thursday 16th.

Now for another case,

text = "Friday"
print(dateparser.parse(text, settings={'PREFER_DATES_FROM': 'current_period'}))
>>> 2021-12-17 00:00:00
print(dateparser.parse(text, settings={'PREFER_DATES_FROM': 'past'}))
>>> 2021-12-17 00:00:00

Here, the output is the same for both cases. But shouldn't the current_period return the nearest Friday or at least the Friday of the ongoing week which will be 24th and not last Friday that is 17th?

Start by reproducing the Thursday and Friday examples through dateparser.parse with the PREFER_DATES_FROM setting. Trace the relative-weekday parsing entry point to determine why current_period and past converge, then check existing tests for related preference behavior. Done means the intended semantics are established and the differing cases are covered by regression tests.