def __init__(self, data_path, **kwargs):
"""
options:
loglevel
"""
loglevel = logging.ERROR if 'loglevel' not in kwargs else kwargs['loglevel']
logging.basicConfig(format='[%(levelname)s][%(name)s] %(message)s')
self.logger = logging.getLogger(__name__+'.'+self.__class__.__name__)
self.logger.setLevel(loglevel)
# only use 40 emotions
self.emotions = filename.emotions['LJ40K']
self.X = OrderedDict()
for emotion in self.emotions:
fpath = FeatureList.get_full_data_path(emotion, data_path)
self.logger.info("load features from %s", fpath)
Xy = utils.load_pkl_file(fpath)
self.X[emotion] = np.zeros((len(Xy), Xy[0]['X'].shape[1]), dtype="float32")
for i in range(len(Xy)):
# make sure only one feature vector in each doc
assert Xy[i]['X'].shape[0] == 1
self.X[emotion][i] = Xy[i]['X']
def __init__(self, file_path, **kwargs):
loglevel = logging.ERROR if 'loglevel' not in kwargs else kwargs['loglevel']
logging.basicConfig(format='[%(levelname)s][%(name)s] %(message)s')
self.logger = logging.getLogger(__name__+'.'+self.__class__.__name__)
self.logger.setLevel(loglevel)
self.logger.info('load feature file %s' % (file_path))
Xy = utils.load_pkl_file(file_path)
n_doc = len(Xy)
self.X = [Xy[idoc]['X'] for idoc in range(n_doc)]
else:
loglevel = logging.ERROR
logging.basicConfig(format='[%(levelname)s][%(name)s] %(message)s', level=loglevel)
logger = logging.getLogger(__name__)
# pre-checking
# if not os.path.exists(args.output_folder):
# logger.info('create output folder %s' % (args.output_folder))
# os.makedirs(args.output_folder)
# load feature list
feature_list = preprocessing.FeatureList(args.feature_list_file)
emotions = filename.emotions['LJ40K']
# create fused dataset
idxs = utils.load_pkl_file(args.index_file) if args.index_file is not None else None
fused_dataset = preprocessing.FusedDataset(idxs, loglevel=loglevel)
for feature_name, data_path in feature_list:
dataset = preprocessing.Dataset(data_path, loglevel=loglevel)
fused_dataset.add_feature(feature_name, dataset)
# load models
if args.linear:
from model import linearsvm
else:
from model import svm
# main loop
output_dict = OrderedDict()
if args.debug:
loglevel = logging.DEBUG
elif args.verbose:
loglevel = logging.INFO
else:
loglevel = logging.ERROR
logging.basicConfig(format='[%(levelname)s][%(name)s] %(message)s', level=loglevel)
logger = logging.getLogger(__name__)
# pre-checking
if not os.path.exists(args.output_folder):
logger.info('create output folder %s' % (args.output_folder))
os.makedirs(args.output_folder)
all_idxs = utils.load_pkl_file(args.index_file) if args.index_file is not None else None
emotions = filename.emotions['LJ40K']
for emotion_id in args.emotion_ids:
emotion_name = emotions[emotion_id]
fname = filename.get_filename_by_emotion(emotion_name, args.input_folder)
fpath = os.path.join(args.input_folder, fname)
logger.info("load features from %s", fpath)
Xy = utils.load_pkl_file(fpath)
idxs = all_idxs['train'][emotion_name][emotion_name] if all_idxs is not None else range(len(Xy))
X = np.zeros((len(idxs), Xy[0]['X'].shape[1]), dtype="float32")
logger.info('X.shape = (%u, %u)' % (X.shape[0], X.shape[1]))
loglevel = logging.INFO
else:
loglevel = logging.ERROR
logging.basicConfig(format='[%(levelname)s][%(name)s] %(message)s', level=loglevel)
logger = logging.getLogger(__name__)
# pre-checking
if not os.path.exists(args.output_folder):
logger.info('create output folder %s' % (args.output_folder))
os.makedirs(args.output_folder)
# load features
feature_list = preprocessing.FeatureList(args.feature_list_file)
# load the index file
idxs = utils.load_pkl_file(args.index_file)
# create fused dataset
fused_dataset = preprocessing.FusedDataset(idxs, loglevel=loglevel)
for feature_name, data_path in feature_list:
dataset = preprocessing.Dataset(data_path, loglevel=loglevel)
fused_dataset.add_feature(feature_name, dataset)
# read parameter file
if args.parameter_file != None:
param_dict = utils.read_parameter_file(args.parameter_file)
# main loop
best_res = {}
emotions = filename.emotions['LJ40K']
# load models
if args.linear:
from model import linearsvm
else:
from model import svm
learners = {}
scalers = {}
for emotion in emotions:
learners[emotion] = linearsvm.LinearSVM(loglevel=loglevel) if args.linear else svm.SVM(loglevel=loglevel)
if args.scaler_folder != None:
fpath = os.path.join(args.scaler_folder, filename.get_filename_by_emotion(emotion, args.scaler_folder))
scalers[emotion] = utils.load_pkl_file(fpath)
fpath = os.path.join(args.model_folder, filename.get_filename_by_emotion(emotion, args.model_folder))
logger.info('loading model for emotion %s' % (emotion))
learners[emotion].load_model(fpath)
# main loop
for emotion_id in args.emotion_ids:
emotion_name = emotions[emotion_id]
logger.info('predicting model for emotion "%s"' % emotion_name)
# create output dir
emotion_dir = os.path.join(args.output_folder, emotion_name)
logger.info('create output folder %s' % (emotion_dir))
def load_results(self, result_files=[]):
self.results = {}
for f in result_files:
res = utils.load_pkl_file(f)
self.results.update(res)
def load(self, filename):
self = utils.load_pkl_file(filename)
def load_results(self, result_files=[]):
self.results = {}
for f in result_files:
res = utils.load_pkl_file(f)
self.results.update(res)
def load(self, filename):
self = utils.load_pkl_file(filename)