def prepare_db(database, pm):
# Sets the selected database folder
# in the path manager for referencing via full path.
pm.set_current_selected_folder(database)
# Processes the file path of the database into
# a pre processed database ready to be used as a learning/training set.
db = process_data.process_database_file(pm)
output_file.write('CURRENT DATASET: ' + database + '\n')
debug_file.write('CURRENT DATASET: ' + database + '\n')
output_file.write('DATA TYPE: ' + db.get_dataset_type() + '\n')
debug_file.write('DATA TYPE: ' + db.get_dataset_type() + '\n')
# Sanity checks.
normal_data, irregular_data = process_data.identify_missing_data(db)
corrected_data = process_data.extrapolate_data(normal_data, irregular_data,
db.get_missing_symbol())
# repaired_db is the total database once the missing values have been filled in.
if len(corrected_data) > 0:
repaired_db = normal_data + corrected_data
else:
repaired_db = normal_data
db.set_data(repaired_db)
# Convert the discrete data to type float.
db.convert_discrete_to_float()
# TODO: make it append the database name to the debug file aswell, so we can get every dataset when running for all of them.
debug_file.write('\n\nFULL DATASET: \n')
for row in db.get_data():
debug_file.write(str(row) + '\n')
return db
def prepare_db(database, pm):
# Sets the selected database folder
# in the path manager for referencing via full path.
pm.set_current_selected_folder(database)
# Processes the file path of the database into
# a pre processed database ready to be used as a learning/training set.
db = process_data.process_database_file(pm)
save_state = verify_save_folder(pm, db)
if save_state is not False:
# This is where we use the loaded save state object specified
pass
# Sanity checks.
normal_data, irregular_data = process_data.identify_missing_data(db)
corrected_data = process_data.extrapolate_data(normal_data, irregular_data,
db.get_missing_symbol())
# repaired_db is the total database once the missing values have been filled in.
if len(corrected_data) > 0:
repaired_db = normal_data + corrected_data
else:
repaired_db = normal_data
db.set_data(repaired_db)
# Convert the discrete data to type float.
db.convert_discrete_to_float()
return db
for row in database:
print(row)
# Initializes path manager with default directory as databases.
path_manager = pm()
# Loads in a list of database folders for the user to select as the current database.
selected_database = select_database(
path_manager.find_folders(path_manager.get_databases_dir()))
# Sets the selected database folder in the path manager for referencing via full path.
path_manager.set_current_selected_folder(selected_database)
# Processes the file path of the database into a pre processed database ready to be used as a learning/training set.
db = process_data.process_database_file(path_manager)
# Sanity checks.
normal_data, irregular_data = process_data.identify_missing_data(db)
corrected_data = process_data.extrapolate_data(normal_data, irregular_data,
db.get_missing_symbol())
# repaired_db is the total database once the missing values have been filled in.
if len(corrected_data) > 0:
repaired_db = normal_data + corrected_data
else:
repaired_db = normal_data
db.set_data(repaired_db)