[DRAFT] Specific build files

[Gautzilla]

🐳 What's new?

This PR allows to pick specific files that are to be included in the public dataset build, with an option to either move or to copy the files to the dataset folder.

🐳 How does it work?

Let's say we have the following file structure:

cool
├── stuff
│   ├── 2007-11-05_00-00-00.wav
│   ├── 2007-11-05_00-01-00.wav
│   └── 2007-11-05_00-02-00.wav
└── things
    ├── 2007-11-05_00-03-00.wav
    ├── 2007-11-05_00-04-00.wav
    └── 2007-11-05_00-05-00.wav

And we want to build a dataset from specific files (let's say those recorded at an odd hour for some reason 🥸):

from pathlib import Path
from osekit.public_api.dataset import Dataset

from osekit import setup_logging
setup_logging()

# Pick the files you want to include to the dataset
files = (
    r"cool\stuff\2007-11-05_00-01-00.wav",
    r"cool\things\2007-11-05_00-03-00.wav",
    r"cool\things\2007-11-05_00-05-00.wav"
)

# Set the DESTINATION folder of the dataset
dataset = Dataset(
    folder = Path(r"cool\odd_hours"),
    strptime_format="%Y-%m-%d_%H-%M-%S",
)

dataset.build_from_files(
    files=files,
    move_files=False,
)

This'll lead to the following structure (note that if move_files=True, the files used for the build would have been moved from their original location rather than copied):

cool
├── stuff
│   ├── 2007-11-5_00-00-00.wav
│   ├── 2007-11-5_00-01-00.wav
│   └── 2007-11-5_00-02-00.wav
├── things
│   ├── 2007-11-5_00-03-00.wav
│   ├── 2007-11-5_00-04-00.wav
│   └── 2007-11-5_00-05-00.wav
└── odd_hours
    ├── data
    │   └── audio
    │       └── original
    │           ├── 2007-11-5_00-01-00.wav
    │           ├── 2007-11-5_00-02-00.wav
    │           └── 2007-11-5_00-03-00.wav
    ├── log
    │   └── logs.log
    └── dataset.json

🐬 Related issue

Merging this PR will close #302