def create_connection_mysql():
""" create a connection to the mysql database
specified by host, user, password and
database name defined in config.ini
:param
:return: Connection object
"""
mydb = create_default_connection_mysql()
mydb.database = cfg.getInstance().database
return mydb
def create_default_connection_mysql():
"""
Create a default connection to the mysql database specified by host, user
and password defined in config.ini
:param
:return mydb: connection object
"""
conf = cfg.getInstance()
mydb = mysql.connector.connect(host=conf.host,
user=conf.user,
password=conf.password)
return mydb
def create_engine_mysql():
"""
Create a pool and dialect together connection to provide a source of database
and behavior
:param:
:return: Connection engine
"""
# in case of connection error, change the host as in the next commented code
conf = cfg.getInstance()
host = conf.host
user = conf.user
passwd = conf.password
db_name = conf.database
return create_engine("mysql+pymysql://{user}:{pw}@{host}/{db}".format(
user=user, pw=passwd, host=host, db=db_name),
pool_pre_ping=True)
def check_database():
"""
check the existence of a database with the name defined in config.ini
if none, a new is created as well as a table of similarity
:param: none
:return:
"""
global mydb, mycursor
try:
check = False
mydb = create_default_connection_mysql()
mycursor = mydb.cursor()
db_name = cfg.getInstance().database
mycursor.execute("SHOW DATABASES")
for x in mycursor:
# x = x[0].decode("unicode-escape") # decode was giving an error (note: decode is needed when using mysql docker)
if x[0].decode("unicode-escape") == db_name:
# x[0].encode().decode('utf-8') == db_name:
check = True
if check != False:
print("Database already exists")
else:
print(
"The database is not available. Calculate similarity based " +
"on SemanticSimDBCreator project")
except Error as e:
print("Error while connecting to MySQL", e)
finally:
if mydb.is_connected():
mycursor.close()
mydb.close()
def check_database():
"""
Check the existence of a database with the name defined in config.ini
if none, a new is created as well as a table of similarity
:param
:return none
"""
global mydb
try:
check = False
mydb = create_default_connection_mysql()
mycursor = mydb.cursor()
db_name = cfg.getInstance().database
mycursor.execute("SHOW DATABASES")
for x in mycursor:
if x[0].decode("unicode-escape") == db_name:
# or,
# if x[0].encode().decode( 'utf-8' ) == db_name:
check = True
if not check:
print("Will create database")
mycursor.execute("CREATE DATABASE " + db_name)
create_table()
else:
print("Database already exists")
except Error as e:
print("Error while connecting to MySQL", e)
finally:
if mydb.is_connected():
mycursor.close()
mydb.close()
pd.set_option('display.max_columns', None)
if __name__ == '__main__':
start_time = datetime.now()
# ---------------------------------------------------------------------------------------- #
# connect db
check_database()
# ---------------------------------------------------------------------------------------- #
# read dataset; select the right columns; select items only from ontology
# you must
df_dataset = upload_dataset( cfg.getInstance().dataset,\
cfg.getInstance().item_prefix )
# ---------------------------------------------------------------------------------------- #
# check ontology database
if os.path.isfile(cfg.getInstance().path_to_ontology):
print("Database ontology file already exists")
else:
print("Database from owl does not exit. Creating...")
ssmpy.create_semantic_base(
cfg.getInstance().path_to_owl,
cfg.getInstance().path_to_ontology,
"http://purl.obolibrary.org/obo/",
"http://www.w3.org/2000/01/rdf-schema#subClassOf", "")