Correct reshape timebin

[MaelleTtrt]

Correction of function reshape_timebin + associated tests

[mathieudpnt]

did some ruff/syntax changes but overall lokks good to me !
changed a test so it's a bit more complete
let me know

[mathieudpnt]

Return timezone(s) from APLOSE DataFrame.

[mathieudpnt]

"""Check if provided timestamp_audio list is correctly formated.

    Parameters
    ----------
    df: DataFrame
        APLOSE results Dataframe.
    timestamp_audio: list[Timestamp]
        list of start timestamps of corresponding audio file for each detection.

"""

[mathieudpnt]

what do you think ?

[MaelleTtrt]

I prefer your following suggestion

"""
timestamp_audio: list[Timestamp]
        A list of timestamps. Each timestamp is the start datetime of the
        corresponding audio file for each detection in df.
"""

[mathieudpnt]

"""
timestamp_audio: list[Timestamp]
        A list of timestamps. Each timestamp is the start datetime of the
        corresponding audio file for each detection in df.
"""

[mathieudpnt]

  if isinstance(get_timezone(df), list):
      df["start_datetime"] = [
          to_datetime(elem, utc=True) for elem in df["start_datetime"]
      ]
      df["end_datetime"] = [
          to_datetime(elem, utc=True) for elem in df["end_datetime"]
      ]

[mathieudpnt]

return (
        concat(results)
        .sort_values(by=["start_datetime", "end_datetime", "annotator", "annotation"])
        .reset_index(drop=True)
    )

[mathieudpnt]

ruff recommandation:

with sample_yaml.open(encoding="utf-8") as f:
    data_yaml = yaml.safe_load(f)

[mathieudpnt]

ruff

sample_key = next(iter(data_yaml.keys()))

[mathieudpnt]

@pytest.mark.parametrize("timestamp", [None, []])
def test_check_timestamp_empty(sample_df: DataFrame, timestamp: list | None) -> None:
    with pytest.raises(ValueError, match="`timestamp_wav` is empty"):
        check_timestamp(sample_df, timestamp)

this will test for None and empty list. I had to change change_timestamp because it would not treat timestamp_audio = [] as an empty list

def check_timestamp(df: DataFrame, timestamp_audio: list[Timestamp]) -> None:
    """Check if provided `timestamp_audio` list is correctly formated.

    Parameters
    ----------
    df: DataFrame
        APLOSE results Dataframe.
    timestamp_audio: list[Timestamp]
        list of start timestamps of corresponding audio file for each detection.

    """
    if timestamp_audio in [None, []]:
        msg = "`timestamp_wav` is empty"
        raise ValueError(msg)
    if len(timestamp_audio) != len(df):
        msg = "`timestamp_wav` is not the same length as `df`"
        raise ValueError(msg)

PS: must add from __future__ import annotations to test_filtering_utils

[MaelleTtrt]

I don't understand, you want me to replace the test test_check_timestamp_none by your test_check_timestamp_empty?

[mathieudpnt]

ruff

def test_check_timestamp_wrong_length(sample_df: DataFrame) -> None:
    len_sample_df = len(sample_df) + 1
    timestamps = [
        Timestamp("2025-01-01") + Timedelta(days=i)
        for i in range(len_sample_df)
    ]  # shorter list
    with pytest.raises(ValueError, match="is not the same length as `df`"):
        check_timestamp(sample_df, timestamps)

[mathieudpnt]

missing blank line