def test_merge_db(self): merge_from = "/Users/superCat/Desktop/PycharmProjectPortable/Seeds/CategorySeedDB2.db" merge_to = "/Users/superCat/Desktop/PycharmProjectPortable/Seeds/CategorySeedDB.db" db_from = CategorySeedSiteDB(merge_from) db_to = CategorySeedSiteDB(merge_to) from_cat = db_from.get_sub_category_tables_name() for item in from_cat: results = db_from.get_from_table(item, 0, 10000000, reverse_read=False, random_read=False) print("adding:", item, "result:", len(results)) db_to.save_to_table(item, results)
def test_db_migration(self): old_seed_db_addr = "/Users/superCat/Desktop/PycharmProjectPortable/Seeds/CategorySeedDB.db" new_seed_db_addr = "/Users/superCat/Desktop/PycharmProjectPortable/Seeds/NewCategorySeedDB.db" old_db = CategorySeedSiteDB(old_seed_db_addr) new_db = CategorySeedSiteDB(new_seed_db_addr) old_tables = [x[0] for x in old_db.cur.execute("SELECT name FROM sqlite_master WHERE type='table';")] print('table length:', len(old_tables)) for table in old_tables: print('doing table:', table) data = old_db.get_from_table(table, 0, 10000000, reverse_read=False, random_read=False) print('data len for table:', len(data)) new_db.save_to_table(table, data) old_db.close() new_db.close()
def test_db_migration(self): old_seed_db_addr = "/Users/superCat/Desktop/PycharmProjectPortable/Seeds/CategorySeedDB.db" new_seed_db_addr = "/Users/superCat/Desktop/PycharmProjectPortable/Seeds/NewCategorySeedDB.db" old_db = CategorySeedSiteDB(old_seed_db_addr) new_db = CategorySeedSiteDB(new_seed_db_addr) old_tables = [ x[0] for x in old_db.cur.execute( "SELECT name FROM sqlite_master WHERE type='table';") ] print('table length:', len(old_tables)) for table in old_tables: print('doing table:', table) data = old_db.get_from_table(table, 0, 10000000, reverse_read=False, random_read=False) print('data len for table:', len(data)) new_db.save_to_table(table, data) old_db.close() new_db.close()
def testImportSeeds0(self): seed_db_addr = "/Users/superCat/Desktop/PycharmProjectPortable/Seeds/NewCategorySeedDB.db" path = "/Users/superCat/Desktop/PycharmProjectPortable/Seeds/Gambling3.csv" db = CategorySeedSiteDB(seed_db_addr) with open(path, mode='rt') as csv_file: # lines = len(csv_file.readlines()) rd = csv.reader(csv_file, delimiter=',') header = next(rd) # skip header counter = 0 temp = [] while True: try: row = next(rd) if len(row) == 0: break if len(row) == 6: domain, backlink, tf, cf, topic, topical_tf = row print("current loc:", counter, "data:", row) # if len(topic) > 0: # decoded_topic = basic_manager.decode_sub_category(topic, False) data = MajesticBacklinkDataStruct( ref_domain=domain, src_cf=int(cf), src_tf=int(tf), src_topical_tf=int(topical_tf)) temp.append(data) except StopIteration: print('stop iteration') break except Exception as ex: print("exception:", str(ex), "row:", str(counter)) if len(str(ex)) == 0: break finally: counter += 1 db.save_to_table('Games/Gambling', temp) db.close()