fill_gaps function gives uninformative Exception in case of frequency error

[AlexMilanor]

What happened + What you expected to happen

Thank you a lot for your work! This lib is helping me a lot.

When running the fill_gaps function with a mismatch between time_col and freq(for example, time_col is int but freq is string), it returns a Pandas AttributeError about datetimelike values. It would be better to receive a ValueError about the mismatch between time_col and freq.

Versions / Dependencies

Click to expand

utilsforecast - 0.2.11

Reproducible example

import pandas as pd
from utilsforecast.preprocessing import fill_gaps

df = pd.DataFrame({
        'uid': [1, 1, 1, 1, 1, 2, 2, 2, 2],
         'dt': [1, 2, 3, 4, 5, 1, 3, 5, 7],
        'val': [5, 4, 6, 2, 2, 5, 6, 4, 4]
        })

fill_gaps(
    df=df,
    id_col='uid',
    time_col='dt',
    freq='MS',
    start='per_serie',
    end='per_serie'
)

The output is an AttributeError when calling id_time_grid.

File ~/miniconda3/envs/env_time_series/lib/python3.12/site-packages/pandas/core/indexes/accessors.py:643, in CombinedDatetimelikeProperties.__new__(cls, data)
    640 elif isinstance(data.dtype, PeriodDtype):
    641     return PeriodProperties(data, orig)
--> 643 raise AttributeError("Can only use .dt accessor with datetimelike values")

AttributeError: Can only use .dt accessor with datetimelike values

Issue Severity

Low: It annoys or frustrates me.

[AlexMilanor]

It seems like it would be simple to implement using validate_freq from the validation module, so I could open a PR if it is ok.

[jmoralez]

Hey. Yes, adding validate_freq at the top should provide a better error message, feel free to open a PR.