import csv
from pathlib import Path
import logging
import joblib
import os
from typing import Optional, Union
import numpy as np
from sklearn.datasets import _base
from sklearn.utils import check_random_state, Bunch
ARCHIVE = _base.RemoteFileMetadata(
filename='musicseer-results-2002-04-29.txt',
url='https://www.labrosa.org/projects/musicsim/musicseer.org/results/musicseer-results-2002-04-29.txt',
checksum=('4506fc16a8f9a30a92207720bc599602b514de20e94494f67eb7a1e348401e2d'))
logger = logging.getLogger(__name__)
[docs]def fetch_musician_similarity(data_home: Optional[os.PathLike] = None, download_if_missing: bool = True,
shuffle: bool = True, random_state: Optional[np.random.RandomState] = None,
return_triplets: bool = False) -> Union[Bunch, np.ndarray]:
""" Load the MusicSeer musician similarity dataset (triplets).
=================== =====================
Triplets 131.970
Objects (Artists) 448
Dimensionality unknown
=================== =====================
See :ref:`musician_similarity_dataset` for a detailed description.
Args:
data_home : optional, default: None
Specify another download and cache folder for the datasets. By default
all scikit-learn data is stored in '~/scikit_learn_data' subfolders.
download_if_missing : optional, default=True
shuffle: default = True
Shuffle the order of triplet constraints.
random_state: optional, default = None
Initialization for shuffle random generator
return_triplets : boolean, default=False.
If True, returns numpy array instead of a Bunch object.
Returns:
dataset : :class:`~sklearn.utils.Bunch`
Dictionary-like object, with the following attributes.
data : ndarray, shape (n_triplets, 3)
Each row corresponding a triplet constraint.
The columns represent the target, choosen and other musician index.
judgement_id : np.ndarray, shape (n_triplets, )
Id of survey Query.
survey_or_game : np.ndarray, shape (n_triplets,)
Letter 'S' or 'G' indicating if comparison origins from survey or game.
user : np.ndarray, shape (n_triplets, )
Array of the user ids, answering the triplet question
artist_name : np.ndarray, shape (413,)
Names of artists, corresponding to the triplet indices.
artist_id : np.ndarray, shape (413,)
Ids of artists, corresponding to the triplet indices.
DESCR : string
Description of the dataset.
triplets : numpy array (n_triplets, 3)
Only present when `return_triplets=True`.
Raises:
IOError: If the data is not locally available, but download_if_missing=False
"""
data_home = Path(_base.get_data_home(data_home=data_home))
if not data_home.exists():
data_home.mkdir()
filepath = Path(_base._pkl_filepath(data_home, 'musician_similarity.pkz'))
if not filepath.exists():
if not download_if_missing:
raise IOError("Data not found and `download_if_missing` is False")
logger.info('Downloading musician similarity from {} to {}'.format(ARCHIVE.url, data_home))
archive_path = _base._fetch_remote(ARCHIVE, dirname=data_home)
with open(archive_path, 'r') as f:
cols = ('judgement', 'survey', 'user', 'target', 'chosen', 'other')
musicians_data = np.array(list(csv.reader(f, delimiter=' '))[1:-2]).T
musicians_data = {k: v for k, v in zip(cols, musicians_data)}
joblib.dump(musicians_data, filepath, compress=6)
os.remove(archive_path)
else:
musicians_data = joblib.load(filepath)
if shuffle:
random_state = check_random_state(random_state)
ix = random_state.permutation(len(musicians_data['target']))
musicians_data = {k: v[ix] for k, v in musicians_data.items()}
module_path = Path(__file__).parent
artists = np.genfromtxt(module_path.joinpath('data', 'musician_names.txt'), delimiter=' ',
dtype={'names': ('name', 'id'), 'formats': ('U29', '<i8')})
triplet_filter = musicians_data['other'] != '' # remove bi-tuples.
triplet_ids = np.c_[musicians_data['target'], musicians_data['chosen'], musicians_data['other']]
triplet_ids = triplet_ids[triplet_filter].astype(int)
all_ids, triplets = np.unique(triplet_ids, return_inverse=True)
triplets = triplets.reshape(triplet_ids.shape)
if return_triplets:
return triplets
ids_to_names = {id: n for id, n in zip(artists['id'], artists['name'])}
for ix, id in enumerate(sorted(np.setdiff1d(all_ids, artists['id']))):
ids_to_names[int(id)] = f"unknown_{ix}"
with module_path.joinpath('descr', 'musician_similarity.rst').open() as rst_file:
fdescr = rst_file.read()
return Bunch(data=triplets,
judgement_id=musicians_data['judgement'][triplet_filter],
survey_or_game=musicians_data['survey'][triplet_filter],
user=musicians_data['user'][triplet_filter],
artist_name=np.asarray([ids_to_names[id] for id in all_ids]),
artist_id=all_ids,
DESCR=fdescr)