get_mps not compatible with Pandas copy_on_write mode

[matthew-brett]

Thanks for this package!

I nearly always use Pandas copy_on_write mode - which will be the default and only mode in upcoming Pandas 3.0 : https://pandas.pydata.org/pandas-docs/stable//user_guide/copy_on_write.html.

I enable it with:

pd.set_option('mode.copy_on_write', True)

However, it looks like this causes some internal problems for pdpy. When I do pdpy.fetch_mps with a date range, before enabling that option, all goes well:

import pdpy

# No error, a data frame results.
pdpy.fetch_mps(from_date='2000-10-11',to_date='2024-02-19')

However, if I enable copy_on_write, I get an error:

pd.set_option('mode.copy_on_write', True)
pdpy.fetch_mps(from_date='2000-10-11',to_date='2024-02-19')

This results in:

File ~/Library/Python/3.11/lib/python/site-packages/pdpy/mps.py:241, in fetch_commons_memberships(from_date, to_date, on_date)
    238         election = general_elections[j, 2]
    240         if date > dissolution and date <= election:
--> 241             end_dates[i] = dissolution
    242             continue
    244 commons_memberships['seat_incumbency_end_date'] = end_dates

ValueError: assignment destination is read-only

Is there something going in internally that does not work with this option?