def set_item_param(self, item, key, value):
"""
Sets item param, applying additional conversion if needed.
"""
try:
if key in config.parser.list_params:
value = utils.strip_split_list(value, config.parser.list_sep)
elif key in config.parser.file_list_params:
value = utils.strip_split_list(value, config.parser.list_sep)
filesize_value = []
for single_filename in value:
#filenames start from slash, trimming it
abspath = os.path.join(config.www.elibrary_dir, single_filename[1:])
if (os.path.isfile(abspath)):
filesize_value.append(os.path.getsize(abspath))
else:
logging.warn("File is not accessible: {0}".format(abspath))
filesize_value.append(0)
item.set(const.FILE_SIZE_PARAM, filesize_value)
elif key in config.parser.keyword_list_params:
value = utils.strip_split_list(value, config.parser.list_sep)
elif (key in config.parser.int_params) and value.isdecimal():
value = int(value)
elif key in config.parser.year_params:
(year_from, year_to, year_circa) = utils.parse_year(value)
item.set(key + config.parser.start_suffix, year_from)
item.set(key + config.parser.end_suffix, year_to)
item.set(key + config.parser.circa_suffix, year_circa)
elif key in config.parser.date_params:
value = datetime.datetime.strptime(value, config.parser.date_format)
except ValueError:
self.raise_error()
item.set(key, value)
def set_item_param(self, item, key, value):
"""
Sets item param, applying additional conversion if needed.
"""
if key in config.parser.latex_params:
utils.validate_latex(item, key, value)
value = utils.parse_latex(value)
try:
if key in config.parser.list_params:
value = utils.strip_split_list(value, config.parser.list_sep)
elif key in config.parser.file_list_params:
value = utils.strip_split_list(value, config.parser.list_sep)
filesize_value = []
for single_filename in value:
#filenames start from slash, trimming it
abspath = os.path.join(config.www.elibrary_dir, single_filename[1:])
if (os.path.isfile(abspath)):
filesize_value.append(os.path.getsize(abspath))
else:
logging.warn("File is not accessible: {0}".format(abspath))
filesize_value.append(0)
item.set(const.FILE_SIZE_PARAM, filesize_value)
elif key in config.parser.keyword_list_params:
value = utils.strip_split_list(value, config.parser.list_sep)
useful_keywords = (set(value) <= config.parser.useless_keywords)
item.set("useful_" + key, useful_keywords)
elif key in config.parser.int_params:
value = int(value)
elif key in config.parser.year_params:
(year_from, year_to, year_circa) = utils.parse_year(value)
item.set(key + config.parser.start_suffix, year_from)
item.set(key + config.parser.end_suffix, year_to)
item.set(key + config.parser.circa_suffix, year_circa)
elif key in config.parser.date_params:
value = datetime.datetime.strptime(value, config.parser.date_format)
except ValueError:
self.raise_error()
item.set(key, value)
def set_item_param(self, item: BibItem, key: str, value: str): """ Sets item param, applying additional conversion if needed. """ value = utils.parse_latex(value) try: if key in config.parser.list_params: value = utils.strip_split_list(value, config.parser.list_sep) elif key in config.parser.int_params: value = int(value) elif key in config.parser.year_params: (year_from, year_to, year_circa) = utils.parse_year(value) item.set(key + config.parser.start_suffix, year_from) item.set(key + config.parser.end_suffix, year_to) item.set(key + config.parser.circa_suffix, year_circa) elif key in config.parser.date_params: value = datetime.datetime.strptime(value, config.parser.date_format) except ValueError: self.raise_error() item.set(key, value)
def retrieve_metadata(mutagen_file, filename):
# Album artist
album_artist = mutagen_file.get("albumartist")
if album_artist == None:
album_artist = mutagen_file.get("TPE2")
if album_artist == None:
album_artist = mutagen_file.get('aART')
if album_artist == None:
album_artist = mutagen_file.get('ALBUM ARTIST')
album_artist = fix_index(album_artist)
# Year
date = mutagen_file.get("date", None)
if date == None:
date = mutagen_file.get("TDRC", None)
if date == None:
try:
date = mutagen_file.get("\xa9day", None)
except ValueError: # For Ogg Opus files
pass
if date == None:
date = mutagen_file.get('DATE') # TODO ensure discovery date exists
if date != None:
date = parse_year(str(fix_index(date)))
else:
print(f"No date loaded for: {filename}")
date = ""
# Codec
codec = type(mutagen_file).__name__
if codec == "MP4":
codec = "M4A"
if codec == "WAVE":
codec = "WAV"
if codec == "OggOpus":
codec = "OPUS"
# Album
album = mutagen_file.get("album")
if album == None:
album = mutagen_file.get("TALB")
if album == None:
album = mutagen_file.get("\xa9alb") # ©alb
if album == None:
album = mutagen_file.get("ALBUM") # ©alb
album = fix_index(album)
# Title
title = mutagen_file.get("title")
if title == None:
title = mutagen_file.get("TIT2")
if title == None:
title = mutagen_file.get("\xa9nam") # ©nam
if title == None:
title = mutagen_file.get("TITLE")
title = fix_index(title)
# Track Number
tracknum = mutagen_file.get("tracknumber")
if tracknum == None:
tracknum = mutagen_file.get("TRCK")
if tracknum == None:
tracknum = mutagen_file.get("trkn")
tracknum = fix_index(tracknum)
if tracknum != None:
if codec == "M4A":
tracknum = tracknum[0]
tracknum = str(tracknum).zfill(2)
else:
print(f"No track number loaded for: {filename}")
return album, album_artist, str(date), codec, title, tracknum
parser.add_argument('--graph_path', help='path to citation-cited csv')
args = parser.parse_args()
logging.StreamHandler(sys.stdout)
logging.basicConfig(format='%(asctime)s - %(message)s',
datefmt='%d-%b-%y %H:%M:%S',
level=logging.INFO)
if __name__ == '__main__':
paper_year_dict = {}
logging.info('Parsing Year from dataset')
for file in os.listdir(args.dataset):
if file.startswith(('P', 'RM')):
paper_year_dict = utils.parse_year(args.dataset + file,
paper_year_dict)
logging.info('Serialising Paper- Year Dictionary')
utils.dump_file(args.dumps, 'paper_year_dict', paper_year_dict)
global_citation_graph = ''
logging.info('Parsing Dataset')
global_citation_graph = utils.create_graph(args.graph_path,
paper_year_dict)
logging.info('Serialising Global Citation Graph')
utils.dump_file(args.dumps, 'global_citation_graph_full',
global_citation_graph)
logging.info('Removing Cycles')
global_citation_graph = utils.remove_cycles(global_citation_graph)
