feat: Improve Period parsing

[Herdi2]

Description

This feature refers to pandas-dev#48000 in the main Pandas repository. It is a comprehensive analysis of which kinds of Period can and cannot be created from a string alone. There are eight categories of dates that cannot be created from string representations alone, these are:

Features to implement

To be implemented

1 Business years not going from Jan - Dec
There needs to be a way to specify these unambiguously.
2 Quarters starting with a different month than January/April/July/October
There needs to be a way to specify these unambiguously.
4 Business days
In the original issue this does not seem to be a use case for these, but this refers to getting all business days in a date range.

Implemented

3 Weeks
There needs to be a way to represent a week using strings. Two ways are originally proposed, either follow the ISO 8601 string standard such as 2022-W31 or represent them using the output of printing a week period str(pd.Period("2017-01-25", freq="W")) => "2017-01-23/2017-01-29".
5 Hours
A problem lifted is that hours are interpreted as minutes due to their string representation, e.g. "2010-11-12 13:00". A proposed solution is to create them in a less ambiguous way such as "2010-11-12 13h". Note that this is shown when running p.freq which gives <Minutes> instead of <Hours> as output.
6 Weeks in the 60s, 70s, 80s or 90s of any century
These lead to DateParseError when trying to recreate them from their string representation.

p = pd.Period( freq='W', year=2272 )
pd.Period(str(p))

The example above does not work, since 2272 is interpreted as 22:72 which is invalid. Therefore, it throws a DateParseError.
7 Weeks from the 24th century onwards
These suffer from the same problem as 6.
8 ISO 8601 ordinal dates
Dates such as "1981-095" should be supported.
9 Support Multi-Year Spans in pandas.Period Parsing
Dates such as "2019-2021" → Represents a multi-year period (Jan 1, 2019 - Dec 31, 2021)
Business Dates such as "2020Q2-2023Q1" → Represents a quarter-based multi-year period

Test cases

Many test cases are given in the original issue pandas-dev#48000, and need to be implemented for each of these features. Other than the ones given in the original issue, other test cases should also be implemented.

Definition of Done

• Each feature is either implemented or given a valid reason for not being implemented

• Each feature has more than one unit test attached to it