[Python] How do I define a week start frequency in Pandas?

I am trying to come up with a frequency in Pandas that represents the start of a calendar week (configurable by week start). For example, for all dates from 2025-01-06 (Monday) to 2025-01-13 (Sunday), I want them to be categorized under Monday's date of 2025-01-06.

However, because of how Pandas DateOffset rolls dates forward to the next period, not all the dates in that week are categorized under 2025-01-06.

You can see what I'm describing in this sample example:

>>> import pandas as pd
>>>
>>> range_dex = pd.date_range(start="2025-01-06", end="2025-01-30")
>>> sample_series = pd.Series(1, index=range_dex)
>>> sample_series
2025-01-06 1
2025-01-07 1
2025-01-08 1
...
2025-01-28 1
2025-01-29 1
2025-01-30 1
Freq: D, dtype: int64
>>>
>>> week_start = 0 # 0 for Monday, ..., 6 for Sunday
>>> freq_pd = pd.offsets.Week(n=1, weekday=week_start)
>>>
>>> sample_series.groupby(pd.Grouper(freq=freq_pd)).size()
2025-01-06 1
2025-01-13 7
2025-01-20 7
2025-01-27 7
2025-02-03 3
Freq: W-MON, dtype: int64


The result I am hoping to achieve instead is:

>>> sample_series.groupby(pd.Grouper(freq=freq_pd)).size()
2025-01-06 7
2025-01-13 7
2025-01-20 7
2025-01-27 4
Freq: W-MON, dtype: int64


I've also tried pd.offsets.Week(n=0, weekday=week_start) and pd.offsets.Week(n=-1, weekday=week_start), which both error, and I can't do the typical math I could do on a Series with something like series + pd.offsets.Week(n=1, weekday=week_start) - pd.offsets.Week(n=1, weekday=week_start). I also have a use case for .asfreq so really want this specified in terms of a frequency, if possible. Thanks!

For reference, this is the calendar for January 2025:

Sun​

Mon​

Tue​

Wed​

Thu​

Fri​

Sat
29​

30​

31​

1​

2​

3​

4
5​

6​

7​

8​

9​

10​

11
12​

13​

14​

15​

16​

17​

18
19​

20​

21​

22​

23​

24​

25
26​

27​

28​

29​

30​

31​

1​

Continue reading...