Custom Calibration Info Subclass

Whenever you want to implement your own version of a calibration or simply want to add some meta information to your calibration info objects, you need to create a new subclass of CalibrationInfo.

When creating a entirely new calibration, you should subclass from CalibrationInfo directly. If you just want to extend an existing object, subclass from the respective class.

In the following, we will see how to extend the FerrarisCalibrationInfo.

Here are the important things to keep in mind:

1. Each subclass needs to overwrite CAL_TYPE. Otherwise it is not recognised as a separate class when loading objects from file.

2. The new class needs to inherit from dataclass, if you add fields that should be serialized when using the export methods.

3. All new attributes need Type annotation and a default value (if in doubt use None)

4. Note, that all new attributes need to be json serializable by default, if you want ot use json export.

from dataclasses import dataclass
from typing import Optional

from example_data import EXAMPLE_PATH
from imucal import FerrarisCalibrationInfo, ferraris_regions_from_df


@dataclass
class ExtendedFerrarisCalibrationInfo(FerrarisCalibrationInfo):
    CAL_TYPE = "ExtendedFerraris"
    new_meta_info: Optional[str] = None

With that we have a new subclass of the ferraris calibration info. As it is marked as dataclass, the __init__ is created automatically and saving and loading from file is also taken care of. To use the new class, we need to provide it when initializing our FerrarisCalibration.

from imucal import FerrarisCalibration

cal = FerrarisCalibration(calibration_info_class=ExtendedFerrarisCalibrationInfo)

Now, we can calculate the calibration as normal (here we just use some dummy data). To provide a value for our new_meta_info field, we can pass it directly to the calculate method.

import pandas as pd

cal_data = ferraris_regions_from_df(pd.read_csv(EXAMPLE_PATH / "annotated_session.csv", header=0, index_col=[0, 1]))

cal_info = cal.compute(
    cal_data, sampling_rate_hz=204.8, from_acc_unit="a.u.", from_gyr_unit="a.u.", new_meta_info="my value"
)

cal_info.new_meta_info

'my value'

And of course we can simply export and reimport the new calibration info to json or hdf5 (we will use a tempfile here to not clutter the example folder).

import tempfile
from pathlib import Path
from imucal.management import load_calibration_info

with tempfile.TemporaryDirectory() as d:
    file_name = Path(d) / "my_cal_info.json"
    cal_info.to_json_file(file_name)

    # An load it again
    loaded_cal_info = load_calibration_info(file_name)

loaded_cal_info.new_meta_info

'my value'

Estimated memory usage: 13 MB