def mlregressionop(action):
try:
if action == "create":
method = request.args['type']
model = createModel("Regression", method)
return jsonify(result="Success", model=model.getId())
elif action == "train":
modelId = request.args['id']
dataName = request.args['data']
label = request.args['target']
features = request.args['train'].split(",")
model = getModel(modelId)
datadf = datautil.load(dataName)
labelData = datautil.getColValues(datadf, label)
featureData = datautil.getColsValues(datadf, features)
data = dict()
data["train"] = featureData
data["target"] = labelData
model.train(data)
predit_data = model.predict(data['train'])
true_data = data['target']
rmse = np.sqrt(np.mean((predit_data - true_data)**2))
return jsonify(result="Success", model=modelId, rmse=rmse)
elif action == "predict":
modelId = request.args['id']
data = json.loads(request.args['data'])
model = getModel(modelId)
return jsonify(result="Success", predict=str(model.predict(data)))
elif action == "predictViz":
modelId = request.args['id']
scale = request.args['scale']
model = getModel(modelId)
return jsonify(result="Success",
predict=str(model.predictViz(int(scale))))
else:
return jsonify(result="Failed",
msg="Do not support this action {}".format(action))
except:
traceback.print_exc()
return jsonify(result="Failed", msg="Some Exception")
def mlaaop(action):
try:
if action == "create":
method = request.args['type']
model = createModel("association_analysis", method)
return jsonify(result="Success", model=model.getId())
elif action == "train":
modelId = request.args['id']
dataName = request.args['data']
label = request.args['label']
features = request.args['features'].split(",")
model = getModel(modelId)
datadf = datautil.load(dataName)
labelData = datautil.getColValues(datadf, label)
featureData = datautil.getColsValues(datadf, features)
data = dict()
data["train"] = featureData
data["target"] = labelData
model.train(data)
return jsonify(result="Success", model=modelId)
elif action == "predict":
'''
modelId = request.args['id']
data = json.loads(request.args['data'])
model = getModel(modelId)
return jsonify(result="Success", predict=str(model.predict(data)))
'''
print "pass"
pass
elif action == "predictViz":
modelId = request.args['id']
scale = request.args['scale']
model = getModel(modelId)
return jsonify(result="Success",
predict=str(model.predictViz(int(scale))))
else:
return jsonify(result="Failed",
msg="Do not support this action {}".format(action))
except:
traceback.print_exc()
return jsonify(result="Failed", msg="Some Exception")
def mlclsop(action):
try:
if action == "create":
method = request.args['type']
model = createModel("Classification", method)
return jsonify(result="Success", model=model.getId())
elif action == "train":
modelId = request.args['id']
dataName = request.args['data']
label = request.args['label']
features = request.args['features'].split(",")
model = getModel(modelId)
datadf = datautil.load(dataName)
labelData = datautil.getColValues(datadf, label)
featureData = datautil.getColsValues(datadf, features)
data = dict()
data["features"] = featureData
data["label"] = labelData
model.train(data)
return jsonify(result="Success", model=modelId)
elif action == "predict":
modelId = request.args['id']
data = json.loads(request.args['data'])
model = getModel(modelId)
return jsonify(result="Success", predict=str(model.predict(data)))
elif action == "predictViz":
modelId = request.args['id']
scale = request.args['scale']
model = getModel(modelId)
return jsonify(result="Success",
predict=str(model.predictViz(int(scale))))
else:
return jsonify(result="Failed",
msg="Do not support this action {}".format(action))
except:
traceback.print_exc()
return jsonify(result="Failed", msg="Some Exception")
def mlclsop(action):
try:
if action == "create":
method = request.args['type']
model = createModel("Classification", method)
return jsonify(result="Success", model=model.getId())
elif action == "train":
ahora = datetime.now().strftime(
'%d%m%Y-%H%M%S') # Obtiene fecha y hora actual
print("Fecha y Hora:", ahora) # Muestra fecha y hora
f = open(ahora + '.txt', 'w')
modelId = request.args['id']
dataName = request.args['data']
label = request.args['label']
features = request.args['features'].split(",")
model = getModel(modelId)
# Guardo la cabecera del archivo, con datos genericos de lo que se eligio en la pantalla
f.write('%s : %s \n' % ('Modelo de entrenamiento', str(model)))
f.write('%s :%s \n' % ('Dataset seleccionado', str(dataName)))
f.write('%s :%s \n' % ('Dato a predecir Y', str(label)))
f.write('%s :%s \n' % ('Caracteristicas X', str(features)))
datadf = datautil.load(dataName)
labelData = datautil.getColValues(datadf, label)
featureData = datautil.getColsValues(datadf, features)
data = dict()
data["features"] = featureData
data["label"] = labelData
data["data"] = datadf
##model.train(data)
## este es un llamado nuevo que retornaria los resultados de las diferentes metricas que se aplicaron al conjunto de datos}}
modelPredic = model.train(data)
f.write('\n')
for elemento in modelPredic:
f.write('%s \n' % elemento)
f.close()
return jsonify(result="Success",
model=modelId,
metric=str(modelPredic))
##return jsonify(result="Success", model=modelId) ## asi estaba el llamado anterior
elif action == "predict":
modelId = request.args['id']
data = json.loads(request.args['data'])
model = getModel(modelId)
return jsonify(result="Success", predict=str(model.predict(data)))
elif action == "predictViz":
modelId = request.args['id']
scale = request.args['scale']
model = getModel(modelId)
return jsonify(result="Success",
predict=str(model.predictViz(int(scale))))
else:
return jsonify(result="Failed",
msg="No se soporta esta accion {}".format(action))
except:
traceback.print_exc()
return jsonify(result="Failed", msg="Some Exception")
def mlregressionop(action):
try:
if action == "create":
method = request.args['type']
model = createModel("Regression", method)
return jsonify(result="Success", model=model.getId())
elif action == "train":
ahora = datetime.now().strftime(
'%d%m%Y-%H%M%S') # Obtiene fecha y hora actual
print("Fecha y Hora:", ahora) # Muestra fecha y hora
f = open(ahora + '.txt', 'w')
modelId = request.args['id']
dataName = request.args['data']
label = request.args['target']
features = request.args['train'].split(",")
model = getModel(modelId)
# Guardo la cabecera del archivo, con datos genericos de lo que se eligio en la pantalla
f.write('%s : %s \n' % ('Modelo de entrenamiento', str(model)))
f.write('%s :%s \n' % ('Dataset seleccionado', str(dataName)))
f.write('%s :%s \n' % ('Dato a predecir Y', str(label)))
f.write('%s :%s \n' % ('Caracteristicas X', str(features)))
datadf = datautil.load(dataName)
labelData = datautil.getColValues(datadf, label)
featureData = datautil.getColsValues(datadf, features)
data = dict()
data["train"] = featureData
data["target"] = labelData
data["data"] = datadf
##model.train(data)
## en metric mandar los datos de las predicciones de los algoritmos, hacer el que model.train devuelva los datos concatenados o en un objeto
modelPredic = model.train(data)
f.write('\n')
for elemento in modelPredic:
f.write('%s \n' % elemento)
f.close()
return jsonify(result="Success",
model=modelId,
metric=str(modelPredic))
elif action == "predict":
modelId = request.args['id']
data = json.loads(request.args['data'])
model = getModel(modelId)
return jsonify(result="Success", predict=str(model.predict(data)))
elif action == "predictViz":
modelId = request.args['id']
scale = request.args['scale']
model = getModel(modelId)
return jsonify(result="Success",
predict=str(model.predictViz(int(scale))))
else:
return jsonify(result="Failed",
msg="No se soporta esta accion{}".format(action))
except:
traceback.print_exc()
return jsonify(result="Failed", msg="Some Exception")