How to assign frequency attribute to a time series?

I have a pandas series, y, whose index column is:

Index(['2020-01-03', '2020-01-07', '2020-01-10', '2020-01-14', '2020-01-17',
       ...
       '2023-07-25', '2023-07-28', '2023-08-01', '2023-08-04', '2023-08-08'],
      dtype='object', name='Date', length=376)

As you can see pandas cannot infer the frequency of the series (there is no freq attribute to the y.index object besides dtype, name, and length)!

Question:

How can I force a frequency attribute such as freq='C', weekmask='1001000' to the series so that when I do y.index.freq I get the frequency of the series back?

What I have tried:

Here is where I am creating the series to begin with:

def ensureSeriesHasAttributes(y: object, series_name: str) -> pd.Series:
    if not isinstance(y, pd.Series):
        y = pd.Series(y, name=series_name)
    else:
        y.name = series_name
    return y

When I change the above function to the following, I get AssertionError:

def ensureSeriesHasAttributes(y: object, series_name: str) -> pd.Series:
    if not isinstance(y, pd.Series):
        y = pd.Series(y, name=series_name, freq='C', weekmask='1001000')
    else:
        y.name = series_name
        y.index.freq = 'C'
        y.index.weekmask = '1001000'
    return y

Solution

How can I force a frequency attribute such as freq='C', weekmask='1001000' to the series?

You can try:

# Custom frequency with a valid weekmask related to your index
cfreq = pd.offsets.CustomBusinessDay(weekmask='0100100')

# Convert Index to DatetimeIndex and apply the custom frequency
y = y.set_axis(pd.to_datetime(y.index)).asfreq(cfreq)

Output:

>>> y
2020-01-03    0
2020-01-07    1
2020-01-10    2
2020-01-14    3
2020-01-17    4
Freq: C, Name: TS, dtype: int64

>>> y.index
DatetimeIndex(['2020-01-03', '2020-01-07', '2020-01-10', '2020-01-14',
               '2020-01-17'],
              dtype='datetime64[ns]', freq='C')

>>> y.index.freq
<CustomBusinessDay>

>>> y.index.freq.weekmask
'0100100'

Minimal Working Example:

import pandas as pd

idx = ['2020-01-03', '2020-01-07', '2020-01-10', '2020-01-14', '2020-01-17']
y = pd.Series(data=range(len(idx)), index=idx, name='TS')

>>> y
2020-01-03    0
2020-01-07    1
2020-01-10    2
2020-01-14    3
2020-01-17    4
Name: TS, dtype: int64

>>> y.index
Index(['2020-01-03', '2020-01-07', '2020-01-10', '2020-01-14', '2020-01-17'],
      dtype='object')

There are some errors and ambiguities in your code especially when y is not a Series.