pandas-stubs

(Work-in-progress, contributions are welcome) Stubs for common pandas classes for static type checking (e.g. with mypy).

Files in this repo are initially generated by stubgen (part of mypy) running on pandas source code. Manual edits are made on top of that to make type-checking work. This is work in progress.

How to use this:

Instructions for using with mypy:

1. Clone the repo to some folder, e.g. ~/pandas-stubs
2. Add the above folder path into either MYPYPATH environment variable, or as part of the mypy_path variable in a config file. See https://mypy.readthedocs.io/en/latest/config_file.html#import-discovery for details about the mypy config file.
3. Run mypy to typecheck.

Code examples:

# mypy should have no problem checking this annotated function:
def filter_col(df: pd.DataFrame, col: str, val: str) -> pd.Series:
    s = df[col]
    s = s.loc[s == val]
    return s


# reveal_type should know the types of all these:
df = pd.read_csv(csv_file_path)  # --> pandas.DataFrame
s = df['col1'] * 100  # --> pandas.Series

ts = pd.Timestamp('2020-01-01')  # --> pandas.Timestamp
delta = pd.Timestamp.now() - ts  # --> pandas.Timedelta
ts2 = ts + pd.Timedelta(10, unit='D')  # --> pandas.Timestamp

dates = pd.date_range('2020-01-01', '2022-01-01', freq='D')  # --> pandas.DatetimeIndex
dates[0]  # --> pandas.Timestamp

TODO

• make into a pip package ...