def _load_from_original_file(self):
URM_all = self._originalReader.URM_DICT['URM_train'] + \
self._originalReader.URM_DICT['URM_validation'] + \
self._originalReader.URM_DICT['URM_test']
n_users, n_items = URM_all.shape
loaded_URM_dict = {"URM_all": URM_all,
"URM_test_negative": self._originalReader.URM_DICT['URM_test_negative']}
loaded_ICM_dict = {"ICM_all": self._originalReader.ICM_DICT["ICM_all"]}
loaded_ICM_mapper_dict = {"ICM_all": { i:i for i in range(self._originalReader.ICM_DICT["ICM_all"].shape[1])}}
loaded_UCM_dict = {"UCM_all": self._originalReader.ICM_DICT["UCM_all"]}
loaded_UCM_mapper_dict = {"UCM_all": { i:i for i in range(self._originalReader.ICM_DICT["UCM_all"].shape[1])}}
user_original_ID_to_index = { i:i for i in range(n_users) }
item_original_ID_to_index = { i:i for i in range(n_items) }
loaded_dataset = Dataset(dataset_name = self._get_dataset_name(),
URM_dictionary = loaded_URM_dict,
ICM_dictionary = loaded_ICM_dict,
ICM_feature_mapper_dictionary = loaded_ICM_mapper_dict,
UCM_dictionary = loaded_UCM_dict,
UCM_feature_mapper_dictionary = loaded_UCM_mapper_dict,
user_original_ID_to_index= user_original_ID_to_index,
item_original_ID_to_index= item_original_ID_to_index,
is_implicit = self.IS_IMPLICIT,
)
return loaded_dataset
def _load_from_original_file(self):
# Load data from original
self._print("Loading original data")
folder_path = self.DATASET_SPLIT_ROOT_FOLDER + self.DATASET_SUBFOLDER
try:
compressed_file = gzip.open(folder_path + self.ZIP_NAME, )
except FileNotFoundError:
self._print("Unable to find data zip file. Downloading...")
download_from_URL(self.DATASET_URL, folder_path, self.ZIP_NAME)
compressed_file = gzip.open(folder_path + self.ZIP_NAME)
URM_path = folder_path + self.FILE_RATINGS_PATH
decompressed_file = open(URM_path, "w")
self._save_GZ_in_text_file(compressed_file, decompressed_file)
decompressed_file.close()
self._print("loading URM")
URM_all, item_original_ID_to_index, user_original_ID_to_index = load_CSV_into_SparseBuilder(
URM_path,
header=False,
separator="\t",
remove_duplicates=True,
custom_user_item_rating_columns=[0, 4, 2])
# URM_all contains the coordinates in textual format
URM_all.data = np.ones_like(URM_all.data)
loaded_URM_dict = {"URM_all": URM_all}
loaded_dataset = Dataset(
dataset_name=self._get_dataset_name(),
URM_dictionary=loaded_URM_dict,
ICM_dictionary=None,
ICM_feature_mapper_dictionary=None,
UCM_dictionary=None,
UCM_feature_mapper_dictionary=None,
user_original_ID_to_index=user_original_ID_to_index,
item_original_ID_to_index=item_original_ID_to_index,
is_implicit=self.IS_IMPLICIT,
)
self._print("cleaning temporary files")
os.remove(URM_path)
self._print("loading complete")
return loaded_dataset
def _load_from_original_file(self):
# Load data from original
zipFile_path = self.DATASET_SPLIT_ROOT_FOLDER + self.DATASET_SUBFOLDER
try:
dataFile = zipfile.ZipFile(zipFile_path +
"hetrec2011-movielens-2k-v2.zip")
except (FileNotFoundError, zipfile.BadZipFile):
self._print("Unable to fild data zip file. Downloading...")
download_from_URL(self.DATASET_URL, zipFile_path,
"hetrec2011-movielens-2k-v2.zip")
dataFile = zipfile.ZipFile(zipFile_path +
"hetrec2011-movielens-2k-v2.zip")
URM_path = dataFile.extract("user_ratedmovies.dat",
path=zipFile_path + "decompressed/")
URM_all, item_original_ID_to_index, user_original_ID_to_index = load_CSV_into_SparseBuilder(
URM_path,
separator="\t",
header=True,
custom_user_item_rating_columns=[0, 1, 2])
loaded_URM_dict = {"URM_all": URM_all}
loaded_dataset = Dataset(
dataset_name=self._get_dataset_name(),
URM_dictionary=loaded_URM_dict,
ICM_dictionary=None,
ICM_feature_mapper_dictionary=None,
UCM_dictionary=None,
UCM_feature_mapper_dictionary=None,
user_original_ID_to_index=user_original_ID_to_index,
item_original_ID_to_index=item_original_ID_to_index,
is_implicit=self.IS_IMPLICIT,
)
self._print("cleaning temporary files")
shutil.rmtree(zipFile_path + "decompressed", ignore_errors=True)
self._print("loading complete")
return loaded_dataset
def _load_from_original_file(self):
# Load data from original
self.zip_file_folder = self.DATASET_OFFLINE_ROOT_FOLDER + self.DATASET_SUBFOLDER
self.decompressed_zip_file_folder = self.DATASET_SPLIT_ROOT_FOLDER + self.DATASET_SUBFOLDER
try:
self.dataFile = zipfile.ZipFile(self.zip_file_folder +
"netflix-prize-data.zip")
except (FileNotFoundError, zipfile.BadZipFile):
self._print("Unable to find data zip file.")
self._print(
"Automatic download not available, please ensure the ZIP data file is in folder {}."
.format(self.zip_file_folder))
self._print("Data can be downloaded here: {}".format(
self.DATASET_URL))
# If directory does not exist, create
if not os.path.exists(self.zip_file_folder):
os.makedirs(self.zip_file_folder)
raise FileNotFoundError("Automatic download not available.")
URM_all, self.item_original_ID_to_index, self.user_original_ID_to_index = self._loadURM(
)
loaded_URM_dict = {"URM_all": URM_all}
loaded_dataset = Dataset(
dataset_name=self._get_dataset_name(),
URM_dictionary=loaded_URM_dict,
ICM_dictionary=None,
ICM_feature_mapper_dictionary=None,
UCM_dictionary=None,
UCM_feature_mapper_dictionary=None,
user_original_ID_to_index=self.user_original_ID_to_index,
item_original_ID_to_index=self.item_original_ID_to_index,
is_implicit=self.IS_IMPLICIT,
)
self._print("loading complete")
return loaded_dataset
def _load_from_original_file(self):
# Load data from original
self._print("Loading original data")
zipFile_path = self.DATASET_SPLIT_ROOT_FOLDER + self.DATASET_SUBFOLDER
try:
dataFile = zipfile.ZipFile(zipFile_path + "filmtrust.zip")
except (FileNotFoundError, zipfile.BadZipFile):
print("FilmTrust: Unable to fild data zip file. Downloading...")
download_from_URL(self.DATASET_URL, zipFile_path, "filmtrust.zip")
dataFile = zipfile.ZipFile(zipFile_path + "filmtrust.zip")
URM_path = dataFile.extract("ratings.txt",
path=zipFile_path + "decompressed/")
URM_all, item_original_ID_to_index, user_original_ID_to_index = load_CSV_into_SparseBuilder(
URM_path, separator=" ", header=False, remove_duplicates=True)
loaded_URM_dict = {"URM_all": URM_all}
loaded_dataset = Dataset(
dataset_name=self._get_dataset_name(),
URM_dictionary=loaded_URM_dict,
ICM_dictionary=None,
ICM_feature_mapper_dictionary=None,
UCM_dictionary=None,
UCM_feature_mapper_dictionary=None,
user_original_ID_to_index=user_original_ID_to_index,
item_original_ID_to_index=item_original_ID_to_index,
is_implicit=self.IS_IMPLICIT,
)
self._print("cleaning temporary files")
shutil.rmtree(zipFile_path + "decompressed", ignore_errors=True)
self._print("loading complete")
return loaded_dataset
def _load_from_original_file(self):
# Load data from original
zipFile_path = self.DATASET_SPLIT_ROOT_FOLDER + self.DATASET_SUBFOLDER
try:
dataFile = zipfile.ZipFile(zipFile_path + "ml-1m.zip")
except (FileNotFoundError, zipfile.BadZipFile):
self._print("Unable to fild data zip file. Downloading...")
download_from_URL(self.DATASET_URL, zipFile_path, "ml-1m.zip")
dataFile = zipfile.ZipFile(zipFile_path + "ml-1m.zip")
ICM_genre_path = dataFile.extract("ml-1m/movies.dat",
path=zipFile_path + "decompressed/")
UCM_path = dataFile.extract("ml-1m/users.dat",
path=zipFile_path + "decompressed/")
URM_path = dataFile.extract("ml-1m/ratings.dat",
path=zipFile_path + "decompressed/")
self._print("loading genres")
ICM_genres, tokenToFeatureMapper_ICM_genres, item_original_ID_to_index = _loadICM_genres(
ICM_genre_path, header=True, separator='::', genresSeparator="|")
self._print("loading UCM")
UCM_all, tokenToFeatureMapper_UCM_all, user_original_ID_to_index = _loadUCM(
UCM_path, header=True, separator='::')
self._print("loading URM")
URM_all, item_original_ID_to_index, user_original_ID_to_index, URM_timestamp = _loadURM_preinitialized_item_id(
URM_path,
separator="::",
header=False,
if_new_user="******",
if_new_item="ignore",
item_original_ID_to_index=item_original_ID_to_index,
user_original_ID_to_index=user_original_ID_to_index)
loaded_URM_dict = {"URM_all": URM_all, "URM_timestamp": URM_timestamp}
loaded_ICM_dict = {"ICM_genres": ICM_genres}
loaded_ICM_mapper_dict = {
"ICM_genres": tokenToFeatureMapper_ICM_genres
}
loaded_UCM_dict = {"UCM_all": UCM_all}
loaded_UCM_mapper_dict = {"UCM_all": tokenToFeatureMapper_UCM_all}
loaded_dataset = Dataset(
dataset_name=self._get_dataset_name(),
URM_dictionary=loaded_URM_dict,
ICM_dictionary=loaded_ICM_dict,
ICM_feature_mapper_dictionary=loaded_ICM_mapper_dict,
UCM_dictionary=loaded_UCM_dict,
UCM_feature_mapper_dictionary=loaded_UCM_mapper_dict,
user_original_ID_to_index=user_original_ID_to_index,
item_original_ID_to_index=item_original_ID_to_index,
is_implicit=self.IS_IMPLICIT,
)
self._print("cleaning temporary files")
shutil.rmtree(zipFile_path + "decompressed", ignore_errors=True)
self._print("loading complete")
return loaded_dataset
def generate_Dataset(self, dataset_name, is_implicit):
assert (
not self.__Dataset_finalized
), "Dataset mappers have already been generated, adding new data is forbidden"
self.__Dataset_finalized = True
# Generate ID to index mappers
self._generate_global_mappers()
self._generate_ICM_UCM_mappers()
URM_DICT_sparse = {}
ICM_DICT_sparse = {}
UCM_DICT_sparse = {}
on_new_ID = "ignore"
for URM_name, URM_dataframe in self.URM_DICT.items():
URM_sparse_builder = IncrementalSparseMatrix_FilterIDs(
preinitialized_col_mapper=self.item_original_ID_to_index,
preinitialized_row_mapper=self.user_original_ID_to_index,
on_new_col=on_new_ID,
on_new_row=on_new_ID,
)
URM_sparse_builder.add_data_lists(
URM_dataframe["UserID"].values,
URM_dataframe["ItemID"].values,
URM_dataframe["Data"].values,
)
URM_DICT_sparse[URM_name] = URM_sparse_builder.get_SparseMatrix()
for ICM_name, ICM_dataframe in self.ICM_DICT.items():
feature_ID_to_index = self.ICM_mapper_DICT[ICM_name]
ICM_sparse_builder = IncrementalSparseMatrix_FilterIDs(
preinitialized_col_mapper=feature_ID_to_index,
preinitialized_row_mapper=self.item_original_ID_to_index,
on_new_col=on_new_ID,
on_new_row=on_new_ID,
)
ICM_sparse_builder.add_data_lists(
ICM_dataframe["ItemID"].values,
ICM_dataframe["FeatureID"].values,
ICM_dataframe["Data"].values,
)
ICM_DICT_sparse[ICM_name] = ICM_sparse_builder.get_SparseMatrix()
for UCM_name, UCM_dataframe in self.UCM_DICT.items():
feature_ID_to_index = self.UCM_mapper_DICT[UCM_name]
UCM_sparse_builder = IncrementalSparseMatrix_FilterIDs(
preinitialized_col_mapper=feature_ID_to_index,
preinitialized_row_mapper=self.user_original_ID_to_index,
on_new_col=on_new_ID,
on_new_row=on_new_ID,
)
UCM_sparse_builder.add_data_lists(
UCM_dataframe["UserID"].values,
UCM_dataframe["FeatureID"].values,
UCM_dataframe["Data"].values,
)
UCM_DICT_sparse[UCM_name] = UCM_sparse_builder.get_SparseMatrix()
loaded_dataset = Dataset(
dataset_name=dataset_name,
URM_dictionary=URM_DICT_sparse,
ICM_dictionary=ICM_DICT_sparse,
ICM_feature_mapper_dictionary=self.ICM_mapper_DICT,
UCM_dictionary=UCM_DICT_sparse,
UCM_feature_mapper_dictionary=self.UCM_mapper_DICT,
user_original_ID_to_index=self.user_original_ID_to_index,
item_original_ID_to_index=self.item_original_ID_to_index,
is_implicit=is_implicit,
)
return loaded_dataset
def load_data(self, save_folder_path=None):
"""
:param save_folder_path: path in which to save the loaded dataset
None use default "dataset_name/original/"
False do not save
:return:
"""
# Use default e.g., "dataset_name/original/"
if save_folder_path is None:
save_folder_path = self.DATASET_SPLIT_ROOT_FOLDER + self._get_dataset_name_root(
) + self._get_dataset_name_data_subfolder()
# If save_folder_path contains any path try to load a previously built split from it
if save_folder_path is not False and not self.reload_from_original_data:
try:
loaded_dataset = Dataset()
loaded_dataset.load_data(save_folder_path)
self._print("Verifying data consistency...")
loaded_dataset.verify_data_consistency()
self._print("Verifying data consistency... Passed!")
loaded_dataset.print_statistics()
return loaded_dataset
except FileNotFoundError:
self._print(
"Preloaded data not found, reading from original files...")
except Exception:
self._print(
"Reading split from {} caused the following exception...".
format(save_folder_path))
traceback.print_exc()
raise Exception("{}: Exception while reading split".format(
self._get_dataset_name()))
self._print("Loading original data")
loaded_dataset = self._load_from_original_file()
self._print("Verifying data consistency...")
loaded_dataset.verify_data_consistency()
self._print("Verifying data consistency... Passed!")
if save_folder_path not in [False]:
# If directory does not exist, create
if not os.path.exists(save_folder_path):
self._print("Creating folder '{}'".format(save_folder_path))
os.makedirs(save_folder_path)
else:
self._print("Found already existing folder '{}'".format(
save_folder_path))
loaded_dataset.save_data(save_folder_path)
self._print("Saving complete!")
loaded_dataset.print_statistics()
return loaded_dataset
def _load_from_original_file_all_amazon_datasets(self,
URM_path,
metadata_path=None,
reviews_path=None):
# Load data from original
self._print("loading URM")
URM_all, URM_timestamp, self.item_original_ID_to_index, self.user_original_ID_to_index = load_CSV_into_SparseBuilder(
URM_path, separator=",", header=False, timestamp=True)
loaded_URM_dict = {"URM_all": URM_all, "URM_timestamp": URM_timestamp}
loaded_ICM_dict = {}
loaded_ICM_mapper_dict = {}
if metadata_path is not None:
self._print("loading metadata")
ICM_metadata, tokenToFeatureMapper_ICM_metadata, _ = self._loadMetadata(
metadata_path, if_new_item="ignore")
ICM_metadata, _, tokenToFeatureMapper_ICM_metadata = remove_features(
ICM_metadata,
min_occurrence=5,
max_percentage_occurrence=0.30,
reconcile_mapper=tokenToFeatureMapper_ICM_metadata)
loaded_ICM_dict["ICM_metadata"] = ICM_metadata
loaded_ICM_mapper_dict[
"ICM_metadata"] = tokenToFeatureMapper_ICM_metadata
if reviews_path is not None:
self._print("loading reviews")
ICM_reviews, tokenToFeatureMapper_ICM_reviews, _ = self._loadReviews(
reviews_path, if_new_item="ignore")
ICM_reviews, _, tokenToFeatureMapper_ICM_reviews = remove_features(
ICM_reviews,
min_occurrence=5,
max_percentage_occurrence=0.30,
reconcile_mapper=tokenToFeatureMapper_ICM_reviews)
loaded_ICM_dict["ICM_reviews"] = ICM_reviews
loaded_ICM_mapper_dict[
"ICM_reviews"] = tokenToFeatureMapper_ICM_reviews
loaded_dataset = Dataset(
dataset_name=self._get_dataset_name(),
URM_dictionary=loaded_URM_dict,
ICM_dictionary=loaded_ICM_dict,
ICM_feature_mapper_dictionary=loaded_ICM_mapper_dict,
UCM_dictionary=None,
UCM_feature_mapper_dictionary=None,
user_original_ID_to_index=self.user_original_ID_to_index,
item_original_ID_to_index=self.item_original_ID_to_index,
is_implicit=self.IS_IMPLICIT,
)
# Clean temp files
self._print("cleaning temporary files")
if metadata_path is not None:
os.remove(metadata_path)
if reviews_path is not None:
os.remove(reviews_path)
self._print("loading complete")
return loaded_dataset
def _load_from_original_file(self):
# Load data from original
self._print("Loading original data")
zipFile_path = self.DATASET_SPLIT_ROOT_FOLDER + self.DATASET_SUBFOLDER
try:
dataFile = zipfile.ZipFile(
zipFile_path + "neural_factorization_machine-master.zip")
except (FileNotFoundError, zipfile.BadZipFile):
self._print("Unable to fild data zip file. Downloading...")
download_from_URL(self.DATASET_URL, zipFile_path,
"neural_factorization_machine-master.zip")
dataFile = zipfile.ZipFile(
zipFile_path + "neural_factorization_machine-master.zip")
inner_path_in_zip = "neural_factorization_machine-master/data/frappe/"
URM_train_path = dataFile.extract(inner_path_in_zip +
"frappe.train.libfm",
path=zipFile_path + "decompressed/")
URM_test_path = dataFile.extract(inner_path_in_zip +
"frappe.test.libfm",
path=zipFile_path + "decompressed/")
URM_validation_path = dataFile.extract(
inner_path_in_zip + "frappe.validation.libfm",
path=zipFile_path + "decompressed/")
tmp_URM_train, item_original_ID_to_index, user_original_ID_to_index = self._loadURM(
URM_train_path,
item_original_ID_to_index=None,
user_original_ID_to_index=None)
tmp_URM_test, item_original_ID_to_index, user_original_ID_to_index = self._loadURM(
URM_test_path,
item_original_ID_to_index=item_original_ID_to_index,
user_original_ID_to_index=user_original_ID_to_index)
tmp_URM_validation, item_original_ID_to_index, user_original_ID_to_index = self._loadURM(
URM_validation_path,
item_original_ID_to_index=item_original_ID_to_index,
user_original_ID_to_index=user_original_ID_to_index)
shape = (len(user_original_ID_to_index),
len(item_original_ID_to_index))
tmp_URM_train = reshapeSparse(tmp_URM_train, shape)
tmp_URM_test = reshapeSparse(tmp_URM_test, shape)
tmp_URM_validation = reshapeSparse(tmp_URM_validation, shape)
URM_occurrence = tmp_URM_train + tmp_URM_test + tmp_URM_validation
URM_all = URM_occurrence.copy()
URM_all.data = np.ones_like(URM_all.data)
loaded_URM_dict = {
"URM_all": URM_all,
"URM_occurrence": URM_occurrence
}
loaded_dataset = Dataset(
dataset_name=self._get_dataset_name(),
URM_dictionary=loaded_URM_dict,
ICM_dictionary=None,
ICM_feature_mapper_dictionary=None,
UCM_dictionary=None,
UCM_feature_mapper_dictionary=None,
user_original_ID_to_index=user_original_ID_to_index,
item_original_ID_to_index=item_original_ID_to_index,
is_implicit=self.IS_IMPLICIT,
)
self._print("cleaning temporary files")
shutil.rmtree(zipFile_path + "decompressed", ignore_errors=True)
self._print("loading complete")
return loaded_dataset
def _load_from_original_file(self):
# Load data from original
zipFile_path = self.DATASET_SPLIT_ROOT_FOLDER + self.DATASET_SUBFOLDER
try:
dataFile = zipfile.ZipFile(zipFile_path + "ml-20m.zip")
except (FileNotFoundError, zipfile.BadZipFile):
print("Movielens20MReader: Unable to fild data zip file. Downloading...")
download_from_URL(self.DATASET_URL, zipFile_path, "ml-20m.zip")
dataFile = zipfile.ZipFile(zipFile_path + "ml-20m.zip")
genres_path = dataFile.extract("ml-20m/movies.csv", path=zipFile_path + "decompressed/")
tags_path = dataFile.extract("ml-20m/tags.csv", path=zipFile_path + "decompressed/")
URM_path = dataFile.extract("ml-20m/ratings.csv", path=zipFile_path + "decompressed/")
self._print("loading genres")
ICM_genres, tokenToFeatureMapper_ICM_genres, item_original_ID_to_index = _loadICM_genres(genres_path, header=True, separator=',', genresSeparator="|")
self._print("loading tags")
ICM_tags, tokenToFeatureMapper_ICM_tags, _ = _loadICM_tags(tags_path, header=True, separator=',', if_new_item = "ignore",
item_original_ID_to_index = item_original_ID_to_index)
self._print("loading URM")
URM_all, item_original_ID_to_index, user_original_ID_to_index, URM_timestamp = _loadURM_preinitialized_item_id(URM_path, separator=",",
header = True, if_new_user = "******", if_new_item = "ignore",
item_original_ID_to_index = item_original_ID_to_index)
ICM_all, tokenToFeatureMapper_ICM_all = merge_ICM(ICM_genres, ICM_tags,
tokenToFeatureMapper_ICM_genres,
tokenToFeatureMapper_ICM_tags)
loaded_URM_dict = {"URM_all": URM_all,
"URM_timestamp": URM_timestamp}
loaded_ICM_dict = {"ICM_genres": ICM_genres,
"ICM_tags": ICM_tags,
"ICM_all": ICM_all}
loaded_ICM_mapper_dict = {"ICM_genres": tokenToFeatureMapper_ICM_genres,
"ICM_tags": tokenToFeatureMapper_ICM_tags,
"ICM_all": tokenToFeatureMapper_ICM_all}
loaded_dataset = Dataset(dataset_name = self._get_dataset_name(),
URM_dictionary = loaded_URM_dict,
ICM_dictionary = loaded_ICM_dict,
ICM_feature_mapper_dictionary = loaded_ICM_mapper_dict,
UCM_dictionary = None,
UCM_feature_mapper_dictionary = None,
user_original_ID_to_index= user_original_ID_to_index,
item_original_ID_to_index= item_original_ID_to_index,
is_implicit = self.IS_IMPLICIT,
)
self._print("cleaning temporary files")
shutil.rmtree(zipFile_path + "decompressed", ignore_errors=True)
self._print("saving URM and ICM")
return loaded_dataset