def columns(self):
"""Return columns in selected dataset"""
if self.is_db:
ds = get_table(self.db_url, self.table)
else:
ds = pd.read_csv(self.dataset.file)
return ds.columns.to_list()
def index(request):
if 'step' not in request.session:
request.session['step'] = 1
# Step 1: upload file directly
if request.method == 'POST' and 'uploaded_file' in request.FILES:
file = request.FILES['uploaded_file']
print("get file: {}".format(file.name))
print("type: {}".format(type(file)))
sep = request.POST['sep'] if request.POST['sep'] else ','
ds = get_dataframe(file, sep=sep)
if ds is not None:
print("dataframe:")
print(ds.head())
request.session['dataset_name'] = file.name
request.session['ds_data'] = ds.to_json()
request.session['sep'] = sep
request.session['is_db'] = False
if 'db_url' in request.session:
del request.session['db_url']
request.session['step'] = 2
return render(request, 'msp/index.html')
else:
print('dataframe {} isn\'t valid'.format(file.name))
return render(request, 'msp/index.html', {
'upload_error':
'dataframe {} isn\'t valid'.format(file.name)
})
# Step 1: get database connection
if request.method == 'POST' and 'db_connection' in request.POST:
db_connection = request.POST['db_connection']
print("get connection: {}".format(db_connection))
engine = get_connector(db_connection)
if engine is not None:
request.session['db_url'] = db_connection
request.session['is_db'] = True
tables = get_tables(db_connection)
return render(request, 'msp/index.html', {'tables': tables})
else:
return render(
request, 'msp/index.html',
{'upload_error': 'url {} isn\'t valid'.format(db_connection)})
# Step 1: get table in DBMS
if request.method == 'POST' and 'table' in request.POST:
table = request.POST['table']
print("get table: {}".format(table))
request.session['dataset_name'] = table
request.session['is_db'] = True
if 'ds_data' in request.session:
del request.session['ds_data']
del request.session['sep']
request.session['step'] = 2
return render(request, 'msp/index.html')
# Step 2: select modality
if request.method == 'POST' and 'mode' in request.POST:
mode = request.POST['mode']
print("select modality: {}".format(mode))
request.session['mode'] = mode
if mode == 'test':
request.session['step'] = 3
return render(request, 'msp/index.html')
elif mode == 'train':
# case 1: passing another table inside the DBMS
if request.session['is_db'] and request.POST['label_table']:
label_table = request.POST['label_table']
print('get label table: {}'.format(label_table))
if not check_table(request.session['db_url'], label_table):
return render(
request, 'msp/index.html', {
'label_error':
'Table {} doens\t exists'.format(label_table)
})
else:
pass
# case 2
elif request.POST['label_column']:
label_column = request.POST['label_column']
print('get label column: {}'.format(label_column))
# check column on table in DBMS
if request.session['is_db']:
print('check column in table {}'.format(
request.session['dataset_name']))
res = check_column(request.session['db_url'],
request.session['dataset_name'],
label_column)
if res:
labels = get_column(request.session['db_url'],
request.session['dataset_name'],
label_column)
print(labels)
request.session['labels'] = labels
request.session['is_label_column'] = label_column
request.session['step'] = 3
return render(request, 'msp/index.html')
# check column on dataframe
else:
print('check column in dataframe {}'.format(
request.session['dataset_name']))
ds = pd.read_json(request.session['ds_data'])
if label_column in ds.columns:
labels = ds[label_column].to_list()
print(ds[label_column])
request.session['labels'] = labels
request.session['is_label_column'] = label_column
request.session['step'] = 3
return render(request, 'msp/index.html')
return render(request, 'msp/index.html', {
'label_error':
'Column {} doesnt exists'.format(label_column)
})
# case 3: upload label file
elif 'label_dataframe' in request.FILES:
file = request.FILES['label_dataframe']
print("get label dataframe: {}".format(file.name))
print(type(file))
ds = get_dataframe(file, sep=',')
print(ds.head())
if ds is not None:
labels = ds.iloc[:, 0].to_list()
print(ds.iloc[:, 0])
request.session['labels'] = labels
request.session['is_label_column'] = False
request.session['step'] = 3
return render(request, 'msp/index.html')
return render(request, 'msp/index.html',
{'label_error': 'Label Dataframe error'})
else:
return render(
request, 'msp/index.html',
{'label_error': 'Mode {} doesnt exists'.format(mode)})
# Step 3: add transformations
if request.method == 'POST' and 'transform_type' in request.POST:
transform_type = request.POST['transform_type']
transform_column = request.POST['transform_column']
error = False
if request.session['is_db']:
print('check column {} in table {}'.format(
transform_column, request.session['dataset_name']))
res = check_column(request.session['db_url'],
request.session['dataset_name'],
transform_column)
if not res:
error = True
else:
print('check column {} in dataframe'.format(transform_column))
ds = pd.read_json(request.session['ds_data'])
if transform_column not in ds.columns:
error = True
if error:
return render(
request, 'msp/index.html', {
'transform_error':
'Impossible apply transform {} {}'.format(
transform_type, transform_column)
})
else:
if 'transforms' not in request.session:
print('create new transform variable')
request.session['transforms'] = []
print(request.session['transforms'])
l = request.session['transforms']
l.append({
'transform_type': transform_type,
'transform_column': transform_column
})
request.session['transforms'] = l
print(request.session['transforms'])
# Step 3: remove transformations
if request.method == 'POST' and 'delete_transforms' in request.POST:
print('delete all transforms')
if 'transforms' in request.session:
del request.session['transforms']
# Step 4: execute
if request.method == 'POST' and 'model_type' in request.POST:
print('run model')
model_type = request.POST['model_type']
model_file = None
if request.session['mode'] == 'test':
if 'uploaded_model' not in request.FILES:
return render(request, 'msp/index.html',
{'model_error': 'No pre trained model uploaded'})
model_file = request.FILES['uploaded_model']
# request.session['model_file'] = model_file
run_db = False
if request.session['mode'] == 'test' and request.session['is_db']:
run_db = request.POST['run_db']
request.session['model_type'] = model_type
request.session['run_db'] = run_db
request.session['step'] = 4
manager = MLManager()
manager.select_model('GradientBoostingRegressor')
if request.session['mode'] == 'train':
if request.session['is_db']:
ds = get_table(request.session['db_url'],
request.session['dataset_name'])
else:
ds = pd.read_json(request.session['ds_data'])
print(ds.head(10))
print(request.session['is_label_column'])
if request.session['is_label_column']:
del ds[request.session['is_label_column']]
model = manager.fit(ds, request.session['labels'])
save_model(model, "data/model_{}.joblib".format(model_type))
print('save_model')
elif request.session['mode'] == 'test':
if request.session['is_db']:
print(
'runjdfisbgsijkbgjsdlgosdbgoisdhgilsdhgodspngfposifdòlhgaofdis'
)
columns = list(
get_table(request.session['db_url'],
request.session['dataset_name']).columns)
print(columns)
query = manager.generate_query(model_file,
request.session['dataset_name'],
columns)
print(query)
ds = execute_query(request.session['db_url'], query)
print(ds)
else:
ds = pd.read_json(request.session['ds_data'])
y_pred = manager.predict(ds, model_file)
y_pred = pd.Series(y_pred, )
y_pred.to_csv("data/{}_prediction.csv".format(model_type),
index=False)
print('save_result')
return render(
request,
'msp/index.html',
)
def main():
print('Workflow Training')
t_start = datetime.now()
scenario = get_train_scenario()
check_train_consistency(scenario)
# Dataset
print('Get Dataset')
ds = get_dataset(scenario)
features = ds.columns.to_list()
# Label
print('Get Label')
if scenario['labels_type'] == 'file':
# Get labels from file
labels = get_dataframe(scenario['labels'])
if labels is None:
raise ValueError('Impossible read label from selected file')
# Get first column from file
labels = labels.iloc[:, 0].to_list()
elif scenario['labels_type'] == 'table':
# Get labels from table
labels = get_table(scenario['db_url'], scenario['labels'])
if ds is None:
raise ValueError(
'Impossible read labels from DBMS, please check DBMS input')
# Get first column from table
labels = labels.iloc[:, 0].to_list()
elif scenario['labels_type'] == 'column':
# Get labels from column
labels = ds[scenario['labels']].to_list()
# Remove label column from dataset
features.remove(scenario['labels'])
else:
raise ValueError('Selected the wrong label type')
# Get train and test from dataset
x_train, y_train, x_test, y_test = get_train_test(ds[features], labels,
scenario['validation'])
# ML Manager
print('Create ML Manager')
manager = MLManager()
# Set ML model
manager.select_model(scenario['model'])
# Set ML transforms
manager.set_transforms(scenario['transforms'])
# Training
print('Training...')
model = manager.fit(x_train, y_train)
# Finish training
t_end = datetime.now()
print('Execution Time: ', t_end - t_start)
# Compute evaluation
if y_test and scenario['metric']:
y_pred = model.predict(x_test)
res_evaluation = manager.evaluate(scenario['metric'], y_test, y_pred)
print('Evaluation: ', res_evaluation)
print('Save Result Model: train.joblib')
trained_pipeline_name = '../data/result/train.joblib'
save_model(model, trained_pipeline_name)
print(':)')
def post(self, request):
# Params
scenario = create_simulation_scenario(request.data)
try:
check_simulation_consistency(scenario)
except ValueError as e:
return Response({
'detail': str(e),
},
status=status.HTTP_400_BAD_REQUEST)
# Get Dataset
t_load_start = datetime.now()
ds = get_table(scenario['db_url'], scenario['table'])
features = ds.columns.to_list()
if scenario['labels_type'] == 'column':
# Remove Label column if exists
features.remove(scenario['labels'])
t_load_end = datetime.now()
# ML Manager
manager = MLManager()
t_ml = []
t_db = []
# Testing Phase
for i in range(scenario['batch_number']):
ds_batch = get_batch(ds, i, scenario['batch_size'])
if ds_batch.empty:
break
# Execute predict using MLManager and ML Library
t_start = datetime.now()
_ = manager.predict(ds[features], scenario['pipeline'].file)
t_end = datetime.now()
t_ml.append(t_end - t_start)
# Create Batch for DBMS
ds_batch.to_sql('batch',
con=scenario["db_url"],
if_exists="replace",
index=False)
# Generate query using MLManager
dbms = DBMSUtils.get_dbms_from_str_connection(scenario['db_url'])
queries, query = manager.generate_query(scenario['pipeline'].file,
scenario['table'],
features, dbms,
scenario['optimizer'])
# Execute query
t_start = datetime.now()
# Execute query
_ = execute_multi_queries(scenario["db_url"], queries)
t_end = datetime.now()
t_db.append(t_end - t_start)
# Finish Simulation
return Response(
{
'detail': 'Successfully predicted result',
'ml_results': {
'execution_time':
(np.mean(t_ml) + (t_load_end - t_load_start))
},
'dbms_results': {
'execution_time': np.mean(t_db)
},
},
status=status.HTTP_200_OK)
def get_dataset(scenario):
if scenario.is_db:
ds = get_table(scenario.db_url, scenario.table)
else:
ds = get_dataframe(scenario.dataset.file)
return ds
def post(self, request):
scenario = create_test_scenario(request.data)
inference_time = 0
try:
check_test_consistency(scenario)
except ValueError as e:
return Response({
'detail': str(e),
},
status=status.HTTP_400_BAD_REQUEST)
# Load pipeline model from pipeline file
pipeline = load_model(scenario.pipeline.file)
if not pipeline:
return Response(
{'detail': 'The selected file isn\'t a pipeline objects'},
status=status.HTTP_400_BAD_REQUEST)
# Extract pipeline information from loaded model
pipeline = MLManager.extract_pipeline_components(pipeline)
if not pipeline:
return Response(
{'detail': 'The selected file isn\'t a pipeline objects'},
status=status.HTTP_400_BAD_REQUEST)
# Model params
scenario.model = pipeline.get('model')
scenario.transforms = json.dumps(pipeline.get('transforms', []))
# Dataset
if scenario.run_db:
# Get features from table
features = get_columns(scenario.db_url, scenario.table)
else:
data_extractor_start = datetime.now()
# Get Dataset
ds = get_dataset(scenario)
features = ds.columns.to_list()
data_extractor_end = datetime.now()
data_extractor_time = (data_extractor_end -
data_extractor_start).total_seconds()
inference_time += data_extractor_time
if scenario.labels_type == 'column':
# Remove Label column if exists
features.remove(scenario.labels)
# ML Manager
manager = MLManager()
# Testing Phase
query = None
if scenario.run_db:
inference_start = datetime.now()
# Generate query using MLManager
dbms = DBMSUtils.get_dbms_from_str_connection(scenario.db_url)
queries, query = manager.generate_query(scenario.pipeline.file,
scenario.table, features,
dbms, scenario.optimizer)
# Execute query
y_pred = execute_multi_queries(scenario.db_url, queries)
y_pred = pd.Series(y_pred.iloc[:, 0], name='Label')
inference_end = datetime.now()
inference_time += (inference_end - inference_start).total_seconds()
else:
inference_start = datetime.now()
# Execute predict using MLManager and ML Library
y_pred = manager.predict(ds[features], scenario.pipeline.file)
y_pred = pd.Series(y_pred, name='Label')
inference_end = datetime.now()
inference_time += (inference_end - inference_start).total_seconds()
# Label
labels = []
# Compute evaluation
if scenario.labels_type:
if scenario.labels_type == 'file':
# Get labels from file
labels_document = get_document_object(scenario.labels)
labels = get_dataframe(labels_document.file)
if labels is None:
return Response(
{
'detail':
'The selected labels file {} isn\'t valid'.format(
labels_document.filename)
},
status=status.HTTP_400_BAD_REQUEST)
# Get first column from file
labels = labels.iloc[:, 0].to_list()
elif scenario.labels_type == 'table':
# Get labels from table
labels = get_table(scenario.db_url, scenario.labels)
if not labels:
return Response(
{
'detail':
'The selected table {} isn\'t valid for label'.
format(scenario.labels)
},
status=status.HTTP_400_BAD_REQUEST)
# Get first column from table
labels = labels.iloc[:, 0].to_list()
elif scenario.labels_type == 'column' and not scenario.run_db:
# Get labels from column
labels = ds[scenario.labels].to_list()
elif scenario.labels_type == 'column' and scenario.run_db:
# Get labels from table
labels = get_column(scenario.db_url, scenario.table,
scenario.labels)
else:
return Response({'detail': 'Select the wrong labels type'},
status=status.HTTP_400_BAD_REQUEST)
# Compute evaluation
res_evaluation = None
if labels and scenario.metric:
res_evaluation = manager.evaluate(scenario.metric, labels, y_pred)
# Create predictions file
test_result_name = "test_{}_{}.csv".format(scenario.model,
datetime.now())
test_result_name = test_result_name.replace(' ', '_')
test_result_name = test_result_name.replace(':', '_')
y_pred.to_csv(test_result_name, index=False, header=True)
# Save predictions in Document model
f = open(test_result_name, 'rb')
document = Document(file=File(f), filename=test_result_name)
document.save()
f.close()
# Remove temporally predictions file
os.remove(test_result_name)
# Save Scenario model
scenario.save()
# Save ResultScenario
result_scenario = ResultScenario()
result_scenario.scenario = scenario
result_scenario.execution_time = inference_time
result_scenario.throughput = result_scenario.execution_time / len(
y_pred)
result_scenario.score = res_evaluation
result_scenario.file_result = document.filename
result_scenario.query = query
result_scenario.save()
return Response(
{
'detail': 'Successfully predicted result',
'filename': test_result_name,
'scenario_id': scenario.id
},
status=status.HTTP_201_CREATED)
def post(self, request):
t_start = datetime.now()
scenario: Scenario = create_train_scenario(request.data)
try:
check_train_consistency(scenario)
except ValueError as e:
return Response({
'detail': str(e),
},
status=status.HTTP_400_BAD_REQUEST)
# Dataset
print('Get Dataset')
ds = get_dataset(scenario)
if ds is None:
return Response(
{
'detail': 'Impossible read the selected dataset',
},
status=status.HTTP_400_BAD_REQUEST)
features = ds.columns.to_list()
# Label
print('Get Label')
if scenario.labels_type == 'file':
# Get labels from file
labels_document = get_document_object(scenario.labels)
labels = get_dataframe(labels_document.file)
if labels is None:
return Response(
{
'detail':
'The selected labels file {} isn\'t valid'.format(
labels_document.filename)
},
status=status.HTTP_400_BAD_REQUEST)
# Get first column from file
labels = labels.iloc[:, 0].to_list()
elif scenario.labels_type == 'table':
# Get labels from table
labels = get_table(scenario.db_url, scenario.labels)
if not labels:
return Response(
{
'detail':
'The selected table {} isn\'t valid for label'.format(
scenario.labels)
},
status=status.HTTP_400_BAD_REQUEST)
# Get first column from table
labels = labels.iloc[:, 0].to_list()
elif scenario.labels_type == 'column':
# Get labels from column
if scenario.labels not in ds.columns:
return Response(
{
'detail':
'The selected column {} isn\'t valid for label'.format(
scenario.labels)
},
status=status.HTTP_400_BAD_REQUEST)
labels = ds[scenario.labels].to_list()
# Remove label column from dataset
features.remove(scenario.labels)
else:
return Response({'detail': 'Select the wrong labels type'},
status=status.HTTP_400_BAD_REQUEST)
# Get train and test from dataset
x_train, y_train, x_test, y_test = get_train_test(
ds[features], labels, scenario.validation)
# ML Manager
manager = MLManager()
# Set ML model
manager.select_model(scenario.model)
# Set ML transforms
manager.set_transforms(json.loads(scenario.transforms))
# Training
model = manager.fit(x_train, y_train)
# Finish training
t_end = datetime.now()
# Compute evaluation
res_evaluation = None
if y_test and scenario.metric:
y_pred = model.predict(x_test)
res_evaluation = manager.evaluate(scenario.metric, y_test, y_pred)
# Create joblib file
trained_pipeline_name = "train_{}_{}.joblib".format(
scenario.model, datetime.now())
trained_pipeline_name = trained_pipeline_name.replace(' ', '_')
trained_pipeline_name = trained_pipeline_name.replace(':', '_')
save_model(model, trained_pipeline_name)
# Save trained pipeline in Document model
f = open(trained_pipeline_name, 'rb')
document = Document(file=File(f), filename=trained_pipeline_name)
document.save()
f.close()
# Remove joblib file
os.remove(trained_pipeline_name)
# Save Scenario model
scenario.save()
# Save ResultScenario
result_scenario = ResultScenario()
result_scenario.scenario = scenario
result_scenario.execution_time = (t_end - t_start).total_seconds()
result_scenario.throughput = result_scenario.execution_time / len(
x_train)
result_scenario.score = res_evaluation
result_scenario.file_result = document.filename
result_scenario.save()
return Response(
{
'detail': 'Pipeline trained correctly',
'filename': trained_pipeline_name,
'scenario_id': scenario.id
},
status=status.HTTP_201_CREATED)
def main():
print('Workflow Testing')
t_start = datetime.now()
scenario = get_test_scenario()
check_test_consistency(scenario)
# Dataset
if scenario['run_db']:
# Get features from table
print('Extract Features')
features = get_columns(scenario['db_url'], scenario['table'])
else:
# Get Dataset
print('Get Dataset and Features')
ds = get_dataset(scenario)
features = ds.columns.to_list()
if scenario['labels_type'] == 'column':
# Remove Label column if exists
features.remove(scenario['labels'])
# ML Manager
print('Create ML Manager')
manager = MLManager()
# Testing Phase
print('Testing...')
if scenario['run_db']:
# Generate query using MLManager
print('Query Generation...')
query = manager.generate_query(scenario['pipeline'], scenario['table'],
features)
# Execute query
print('Query Execution...')
y_pred = execute_query(scenario['db_url'], query)
y_pred = pd.Series(y_pred.iloc[:, 0], name='Label')
else:
# Execute predict using MLManager and ML Library
y_pred = manager.predict(ds[features], scenario['pipeline'])
y_pred = pd.Series(y_pred.flatten(), name='Label')
# Finish testing
t_end = datetime.now()
print('Execution Time: ', t_end - t_start)
print('Save prediction: test.csv')
# test_result_name = '../data/result/test_run_db{}.csv'.format(scenario['run_db'])
test_result_name = '../data/result/test.csv'
y_pred.to_csv(test_result_name, index=False, header=True)
# Compute evaluation
if scenario['labels_type']:
# Label
print('Get Label')
if scenario['labels_type'] == 'file':
# Get labels from file
labels = get_dataframe(scenario['labels'])
if labels is None:
raise ValueError('Impossible read label from selected file')
# Get first column from file
labels = labels.iloc[:, 0].to_list()
elif scenario['labels_type'] == 'table':
# Get labels from table
labels = get_table(scenario['db_url'], scenario['labels'])
if ds is None:
raise ValueError(
'Impossible read labels from DBMS, please check DBMS input'
)
# Get first column from table
labels = labels.iloc[:, 0].to_list()
elif scenario['labels_type'] == 'column' and not scenario['is_db']:
# Get labels from column
labels = ds[scenario['labels']].to_list()
elif scenario['labels_type'] == 'column' and scenario['is_db']:
# Get labels from table
labels = get_column(scenario['db_url'], scenario['table'],
scenario['labels'])
else:
raise ValueError('Select the wrong label type')
res_evaluation = manager.evaluate(scenario['metric'], labels, y_pred)
print('Evaluation: ', res_evaluation)
print(':)')