lorenzo.volpi
/
QuaPy
forked from moreo/QuaPy
1
0
Fork 0
Code Pull Requests Activity

format fix

This commit is contained in:
Alejandro Moreo Fernandez 2021-11-09 15:44:57 +01:00
parent 238a30520c
commit 611d080ca6
5 changed files with 104 additions and 75 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
LeQua2022/baselines_T1Amodsel.py
View File

@ -1,12 +1,5 @@
import pickle
import numpy as np
from sklearn.linear_model import LogisticRegression
from tqdm import tqdm
import pandas as pd
import quapy as qp
from quapy.data import LabelledCollection
from quapy.method.aggregative import *
import quapy.functional as F
from data import *
@ -50,8 +43,10 @@ def gen_samples():
return gen_load_samples_T1(T1A_devvectors_path, nF, ground_truth_path=T1A_devprevalence_path, return_id=False)
for quantifier in [CC]: #, ACC, PCC, PACC, EMQ, HDy]:
#classifier = CalibratedClassifierCV(LogisticRegression(), n_jobs=-1)
for quantifier in [EMQ]: # [CC, ACC, PCC, PACC, EMQ, HDy]:
if quantifier == EMQ:
classifier = CalibratedClassifierCV(LogisticRegression(), n_jobs=-1)
else:
classifier = LogisticRegression()
model = quantifier(classifier)
print(f'{model.__class__.__name__}: Model selection')

5
LeQua2022/evaluation.py → LeQua2022/evaluate.py
View File

@ -2,7 +2,6 @@ import argparse
import quapy as qp
from data import ResultSubmission, evaluate_submission
import constants
import os
"""
LeQua2022 Official evaluation script
@ -20,9 +19,7 @@ def main(args):
print(f'MRAE: {mrae:.4f}')
if args.output is not None:
outdir = os.path.dirname(args.output)
if outdir:
os.makedirs(outdir, exist_ok=True)
qp.util.create_parent_dir(args.output)
with open(args.output, 'wt') as foo:
foo.write(f'MAE: {mae:.4f}\n')
foo.write(f'MRAE: {mrae:.4f}\n')

13
LeQua2022/predict.py
View File

@ -1,11 +1,11 @@
import argparse
import quapy as qp
from data import ResultSubmission, evaluate_submission
from data import ResultSubmission
import constants
import os
import pickle
from tqdm import tqdm
from data import gen_load_samples_T1, load_category_map
from data import gen_load_samples_T1
from glob import glob
import constants
@ -22,21 +22,16 @@ def main(args):
f'dev samples ({constants.DEV_SAMPLES}) nor with the expected number of '
f'test samples ({constants.TEST_SAMPLES}).')
# _, categories = load_category_map(args.catmap)
# load pickled model
model = pickle.load(open(args.model, 'rb'))
# predictions
predictions = ResultSubmission()
for sampleid, sample in tqdm(gen_load_samples_T1(args.samples, args.nf),
desc='predicting', total=nsamples):
for sampleid, sample in tqdm(gen_load_samples_T1(args.samples, args.nf), desc='predicting', total=nsamples):
predictions.add(sampleid, model.quantify(sample))
# saving
basedir = os.path.basename(args.output)
if basedir:
os.makedirs(basedir, exist_ok=True)
qp.util.create_parent_dir(args.output)
predictions.dump(args.output)

66
quapy/model_selection.py
View File

@ -11,27 +11,14 @@ import inspect
class GridSearchQ(BaseQuantifier):
"""Grid Search optimization targeting a quantification-oriented metric.
def __init__(self,
model: BaseQuantifier,
param_grid: dict,
sample_size: Union[int, None],
protocol='app',
n_prevpoints: int = None,
n_repetitions: int = 1,
eval_budget: int = None,
error: Union[Callable, str] = qp.error.mae,
refit=True,
val_split=0.4,
n_jobs=1,
random_seed=42,
timeout=-1,
verbose=False):
"""
Optimizes the hyperparameters of a quantification method, based on an evaluation method and on an evaluation
protocol for quantification.
:param model: the quantifier to optimize
:param param_grid: a dictionary with keys the parameter names and values the list of values to explore for
:type model: BaseQuantifier
:param param_grid: a dictionary with keys the parameter names and values the list of values to explore
:param sample_size: the size of the samples to extract from the validation set (ignored if protocl='gen')
:param protocol: either 'app' for the artificial prevalence protocol, 'npp' for the natural prevalence
protocol, or 'gen' for using a custom sampling generator function
@ -64,6 +51,23 @@ class GridSearchQ(BaseQuantifier):
being ignored, a TimeoutError exception is raised. If -1 (default) then no time bound is set.
:param verbose: set to True to get information through the stdout
"""
def __init__(self,
model: BaseQuantifier,
param_grid: dict,
sample_size: Union[int, None],
protocol='app',
n_prevpoints: int = None,
n_repetitions: int = 1,
eval_budget: int = None,
error: Union[Callable, str] = qp.error.mae,
refit=True,
val_split=0.4,
n_jobs=1,
random_seed=42,
timeout=-1,
verbose=False):
self.model = model
self.param_grid = param_grid
self.sample_size = sample_size
@ -90,7 +94,7 @@ class GridSearchQ(BaseQuantifier):
if self.n_prevpoints != 1:
print('[warning] n_prevpoints has been set and will be ignored for the selected protocol')
def sout(self, msg):
def _sout(self, msg):
if self.verbose:
print(f'[{self.__class__.__name__}]: {msg}')
@ -145,7 +149,8 @@ class GridSearchQ(BaseQuantifier):
raise ValueError('unknown protocol')
def fit(self, training: LabelledCollection, val_split: Union[LabelledCollection, float, Callable] = None):
"""
""" Learning routine. Fits methods with all combinations of hyperparameters and selects the one minimizing
the error metric.
:param training: the training set on which to optimize the hyperparameters
:param val_split: either a LabelledCollection on which to test the performance of the different settings, or
a float in [0,1] indicating the proportion of labelled data to extract from the training set
@ -164,12 +169,12 @@ class GridSearchQ(BaseQuantifier):
if self.timeout > 0:
def handler(signum, frame):
self.sout('timeout reached')
self._sout('timeout reached')
raise TimeoutError()
signal.signal(signal.SIGALRM, handler)
self.sout(f'starting optimization with n_jobs={n_jobs}')
self._sout(f'starting optimization with n_jobs={n_jobs}')
self.param_scores_ = {}
self.best_score_ = None
some_timeouts = False
@ -185,7 +190,7 @@ class GridSearchQ(BaseQuantifier):
model.fit(training)
true_prevalences, estim_prevalences = self.__generate_predictions(model, val_split)
score = self.error(true_prevalences, estim_prevalences)
self.sout(f'checking hyperparams={params} got {self.error.__name__} score {score:.5f}')
self._sout(f'checking hyperparams={params} got {self.error.__name__} score {score:.5f}')
if self.best_score_ is None or score < self.best_score_:
self.best_score_ = score
self.best_params_ = params
@ -201,15 +206,19 @@ class GridSearchQ(BaseQuantifier):
if self.best_score_ is None and some_timeouts:
raise TimeoutError('all jobs took more than the timeout time to end')
self.sout(f'optimization finished: best params {self.best_params_} (score={self.best_score_:.5f})')
self._sout(f'optimization finished: best params {self.best_params_} (score={self.best_score_:.5f})')
if self.refit:
self.sout(f'refitting on the whole development set')
self._sout(f'refitting on the whole development set')
self.best_model_.fit(training + val_split)
return self
def quantify(self, instances):
"""Estimate class prevalence values
:param instances: sample contanining the instances
"""
assert hasattr(self, 'best_model_'), 'quantify called before fit'
return self.best_model().quantify(instances)
@ -218,9 +227,18 @@ class GridSearchQ(BaseQuantifier):
return self.best_model().classes_
def set_params(self, **parameters):
"""Sets the hyper-parameters to explore.
:param parameters: a dictionary with keys the parameter names and values the list of values to explore
"""
self.param_grid = parameters
def get_params(self, deep=True):
"""Returns the dictionary of hyper-parameters to explore (`param_grid`)
:param deep: Unused
:return: the dictionary `param_grid`
"""
return self.param_grid
def best_model(self):

36
quapy/util.py
View File

@ -11,13 +11,12 @@ import numpy as np
from joblib import Parallel, delayed
def get_parallel_slices(n_tasks, n_jobs=-1):
def _get_parallel_slices(n_tasks, n_jobs=-1):
if n_jobs == -1:
n_jobs = multiprocessing.cpu_count()
batch = int(n_tasks / n_jobs)
remainder = n_tasks % n_jobs
return [slice(job * batch, (job + 1) * batch + (remainder if job == n_jobs - 1 else 0)) for job in
range(n_jobs)]
return [slice(job * batch, (job + 1) * batch + (remainder if job == n_jobs - 1 else 0)) for job in range(n_jobs)]
def map_parallel(func, args, n_jobs):
@ -26,7 +25,7 @@ def map_parallel(func, args, n_jobs):
func is applied in two parallel processes to args[0:50] and to args[50:99]
"""
args = np.asarray(args)
slices = get_parallel_slices(len(args), n_jobs)
slices = _get_parallel_slices(len(args), n_jobs)
results = Parallel(n_jobs=n_jobs)(
delayed(func)(args[slice_i]) for slice_i in slices
)
@ -49,9 +48,14 @@ def parallel(func, args, n_jobs):
)
@contextlib.contextmanager
def temp_seed(seed):
"""
Can be used in a "with" context to set a temporal seed without modifying the outer numpy's current state. E.g.:
with temp_seed(random_seed):
# do any computation depending on np.random functionality
:param seed: the seed to set within the "with" context
"""
state = np.random.get_state()
np.random.seed(seed)
try:
@ -88,10 +92,30 @@ def get_quapy_home():
def create_parent_dir(path):
os.makedirs(Path(path).parent, exist_ok=True)
parentdir = Path(path).parent
if parentdir:
os.makedirs(parentdir, exist_ok=True)
def save_text_file(path, text):
create_parent_dir(path)
with open(text, 'wt') as fout:
fout.write(text)
def pickled_resource(pickle_path:str, generation_func:callable, *args):
"""
Allows for fast reuse of resources that are generated only once by calling generation_func(*args). The next times
this function is invoked, it loads the pickled resource. Example:
def some_array(n):
return np.random.rand(n)
pickled_resource('./my_array.pkl', some_array, 10) # the resource does not exist: it is created by some_array(10)
pickled_resource('./my_array.pkl', some_array, 10) # the resource exists: it is loaded from './my_array.pkl'
:param pickle_path: the path where to save (first time) and load (next times) the resource
:param generation_func: the function that generates the resource, in case it does not exist in pickle_path
:param args: any arg that generation_func uses for generating the resources
:return: the resource
"""
if pickle_path is None:
return generation_func(*args)
else: