def test_sql_queries(function: Callable):
"""
Проверка SQL запроса на корректность
:param function: функция, возвращаюшая SQL запрос
"""
filename = f'./results/{function.__name__}'
orig_values = utils.read_pickled(filename)
query = function()
if type(query) != str:
pytest.fail(
f'Ваша функция {function.__name__} должна возвращать SQL запрос в виде строки'
)
if 'select' not in query.lower():
pytest.fail(
f'Запрос функции {function.__name__} должна быть SELECT запросом, но вместо этого там "{query}"'
)
try:
results, columns = utils.execute_query(query)
except Exception as e:
pytest.fail(
f'Непредвиденная ошибка во время выполнения запроса функции {function.__name__} '
f'! Получена ошибка: "{e}"')
return
try:
utils.compare_results(results, orig_values)
except ValueError as e:
pytest.fail(f'Неверный ответ для функции {function.__name__} ! {e}')
except Exception as e:
pytest.fail(
f'Непредвиденная ошибка во время проверки результата функции {function.__name__}'
)
def get_accuracy(datasets):
accuracy = {}
for dname in datasets:
ground_truth = []
predicted = []
data = datasets[dname]
for image_path, label in tqdm(data["gen"], total=data["len"]):
image = get_image_file_object(image_path)
job = computervision_client.read_in_stream(image=image, raw=True)
operation_id = job.headers['Operation-Location'].split('/')[-1]
image_analysis = computervision_client.get_read_result(
operation_id)
while image_analysis.status in ['notStarted', 'running']:
time.sleep(1)
image_analysis = computervision_client.get_read_result(
operation_id=operation_id)
response = image_analysis.as_dict()
result = reformat_azure_response(response)
ground_truth.append(label)
predicted.append(result)
postprocessing = datasets[dname]["postprocessing"]
predicted = postprocessing(predicted)
accuracy[dname] = compare_results(ground_truth, predicted)
return accuracy
def get_accuracy(datasets):
accuracy = {}
for dname in datasets:
ground_truth = []
predicted = []
data = datasets[dname]
for image_path, label in tqdm(data["gen"], total=data["len"]):
with open(image_path, 'br') as image_file:
content = image_file.read()
response = client.detect_text(Image={'Bytes': content})
result = reformat_aws_response(response)
ground_truth.append(label)
predicted.append(result)
postprocessing = datasets[dname]["postprocessing"]
predicted = postprocessing(predicted)
accuracy[dname] = compare_results(ground_truth, predicted)
return accuracy
import utils
# get entropy score
utils.get_entropy_from_file('recommend.txt', False)
# Compare recommend files
utils.compare_results('recommend.txt', 'recommend_2.txt', False)
# Check recs duplication
utils.check_recs_duplication('recommend.txt', False)
]],
tablefmt="psql",
))
pd.DataFrame(
results_dict[data_i][learner_name][encoder_name]
["grid_results"]).to_csv(
"./results_regression/grid_results/MAEvsMSE{}_{}.csv".
format(f"{learner_name}_{encoder_name}_grid_results",
data_i[5:10]))
if encoder_name != "te":
pvalue = compare_results(
results_dict[data_i][learner_name]["te"]["grid_results"],
results_dict[data_i][learner_name][encoder_name]
["grid_results"],
)
print(
f"p-value of wilcoxon test with {encoder_name}: {pvalue}")
# Add Results
resultados.append([
data_i.split("/")[1],
# Scores
results_dict[data_i]["lm"]["te"]["train_mae"],
results_dict[data_i]["lm"]["te"]["test_mae"],
results_dict[data_i]["lm"]["te"]["train_mse"],
results_dict[data_i]["lm"]["te"]["test_mse"],
results_dict[data_i]["lm"]["pe"]["train_mae"],
results_dict[data_i]["lm"]["pe"]["test_mae"],
");")[0].split(",")[0:2]
# car desc is just text of the html
car_desc = soup_descs[car_idx].get_text(' ').strip()
# calculate distance between user coords and station coords from infos
distance = utils.get_distance(user_coords, [
station_stored_infos['Longitude'],
station_stored_infos['Latitude']
])
slot["cars"].append({
"Name": station_name,
"distance": distance,
"description": car_desc
})
new_slots[id_slot] = slot
flag_new, new_slots = utils.compare_results(new_slots, old_slots)
# save the new json file containing most recent slots locally or in Gdrive
if "gdrive_results" in os.environ.keys():
gdrive_conn.save_byte_file(new_slots, results_fileid)
else:
json.dump(new_slots, json_file, indent=4, separators=(',', ': '))
# send mail if something new happened
if flag_new:
pprint.pprint(new_slots)
# if we pass a list (stringyfied), convert it to list again
to_email = os.environ["communauto_mailto"]
if '[' in to_email:
to_email = ast.literal_eval(to_email)
utils.send_mail(new_slots, to_email)