DataAplose simplification

[mathieudpnt]

I simplified the process, now user only has to declare a dict of arguments and use DataAplose from_dict method. The rest of the process should remain the same.

Example:

from post_processing.dataclass.data_aplose import DataAplose

config = [
    {
        "detection_file": ...,
        "filename_format": ...,
    },
    {
        "detection_file": ...,
        "filename_format": ...,
    },
]

data = DataAplose.from_dict(config)

config can be either a dict or a list of dict.

A dict must contains the following arguments:

• detection_file: Path -> path to an APLOSE-formatted result csv file
• filename_format: str -> string to parse datetimes from filename column from the detection file

Others non-mandatory arguments can be given:

• timebin_new: Timedelta | None = None
• start_datetime: Timestamp | None = None
• end_datetime: Timestamp | None = None
• annotator: str | list[str] | None = None
• annotation: str | list[str] | None = None
• type: str | None = None
• recording_file: Path | None = None
• user_selection: str = "all"
• min_frequency: float | None = None
• max_frequency: float | None = None
• confidence: float | None = None

[mathieudpnt]

I also updated the utils script names, got rid of "_utils"

example:
glider_utils -> glider

from post_processing.utils import glider

[mathieudpnt]

I also (finally) renamed the package to disclose 🪩

[mathieudpnt]

and finally i update RecordingPeriod class, now to display the effort, user has to provide in the configuration dict recording_file which corresponds to the Path of a recording_period CSV file:

config ={
        "detection_file": Path(...),
        "filename_format": "...",
        "recording_file": Path(...),
}

data= DataAplose.from_dict(config)

fig, ax = plt.subplots(1, 1)
ax = data.set_ax(...)
data.plot(
    ...
    effort=True,
)
plt.tight_layout()
plt.show()

[MaelleTtrt]

I have an issue when building a DataAplose from a configuration containing multiple concatenated files. Calling data.plot(..., effort=True) raises an error.
From debugging, the error seems to come from RecordingPeriod.from_config, which expects config.recording_file. However, in my case config is a list of configs, and recording_file is only defined in the first element of that list.

I’m not sure what the intended behavior should be here:

• Should recording_file be provided in every config item when concatenating datasets?
• Or should plot(effort=True) handle the case where self.config is a list and retrieve the recording information accordingly? Because maybe there could be multiple recording-file that should be concatenated. This is not my case as I already concatenated the recording_files into one csv.

[MaelleTtrt]

The start_datetime and end_datetime specified in the config are different from the one present in the DataAplose built from this config. The one present in DataAplose are the start_datetime and end_datetime of the detections contained in data.df.
I think start_datetime and end_datetime of DataAplose should be the one specified by the user in the config.

[mathieudpnt]

@MaelleTtrt I have fixed DataAplose start_datetime and end_datetime, if specified in a config dict, the Timestamps will stick to these values, else it will take the min/max Timestamps of the DataFrame.

Also, i have created a concat method to DataAploseConfig to handle a recording_file even if a list of dict is provided for the conf. Let me know if this works for you

[mathieudpnt]

While we're at it, i changed every mention of annotation to label 👀