def index(request):
LR_tfidf = pickle.load(
open(settings.BASE_DIR / "Machine_Learning/Models/LR_tfidf.sav", "rb"))
SVC_tfidf = pickle.load(
open(settings.BASE_DIR / "Machine_Learning/Models/SVC_tfidf.sav",
"rb"))
NB_count = pickle.load(
open(settings.BASE_DIR / "Machine_Learning/Models/NB_count.sav", "rb"))
LR_count = pickle.load(
open(settings.BASE_DIR / "Machine_Learning/Models/LR_count.sav", "rb"))
SVC_count = pickle.load(
open(settings.BASE_DIR / "Machine_Learning/Models/SVC_count.sav",
"rb"))
if request.method == 'GET':
# print(request.GET.get("text"))
try:
if request.GET.get("text"):
data = {'text': request.GET.get("text")}
else:
data = json.loads(request.body.decode("utf-8"))
except (TypeError, json.JSONDecodeError):
return JsonResponse(
{
"message":
"Bad Request, check if you are not sending Empty or invalid Data"
},
status=400)
try:
x = clean_and_extract([str(data['text'])])
x_te = x[0]
x_te1 = x[1]
x_c = x[2]
prob_LR_tfidf = LR_tfidf.predict_proba(x_te)[:, 1][0]
prob_LR_count = LR_count.predict_proba(x_te1)[:, 1][0]
prob_SVC_tfidf = SVC_tfidf._predict_proba_lr(x_te)[:, 1][0]
prob_SVC_count = SVC_count._predict_proba_lr(x_te1)[:, 1][0]
prob_NB_count = NB_count.predict_proba(x_c)[:, 1][0]
Type = []
probabilities = [
prob_LR_tfidf, prob_LR_count, prob_SVC_tfidf, prob_SVC_count,
prob_NB_count
]
threshold = [0.18, 0.17, 0.4, 0.4, 0.12]
for i in range(len(probabilities)):
if probabilities[i] >= threshold[i]:
Type.append("Insincere")
else:
Type.append("Sincere")
response_body = {
"text": str(data['text']),
"models": {
"LR_TFIDF": {
"type": Type[0],
"probability": probabilities[0]
},
"LR_COUNT": {
"type": Type[1],
"probability": probabilities[1]
},
"SVC_TFIDF": {
"type": Type[2],
"probability": probabilities[2]
},
"SVC_COUNT": {
"type": Type[3],
"probability": probabilities[3]
},
"NB": {
"type": Type[4],
"probability": probabilities[4]
}
}
}
response = JsonResponse(response_body, status=200)
return response
except Exception as exception:
return JsonResponse(
{
"message":
"Sorry something went wrong, counld not process your request. Try again",
"error": str(exception)
},
status=500)
if request.method == "POST":
try:
data = json.loads(request.body.decode("utf-8"))
except Exception as e:
return JsonResponse({
"message": "Bad Data",
"error": str(e)
},
status=422)
if data == []:
return JsonResponse(
{
"message": "Empty data",
"error": "No Data sent for processing"
},
status=400)
if type(data) != type([]):
return JsonResponse(
{
"message": "Bad Data",
"error": "Data is not in JSON array/list format"
},
status=422)
try:
questions = []
try:
for ele in data:
questions.append(str(ele["text"]))
except KeyError as ke:
return JsonResponse(
{
"message": """"text" attribute not found. Try again""",
"error": str(ke)
},
status=422)
x = clean_and_extract(questions)
x_te = x[0]
x_te1 = x[1]
x_c = x[2]
prob_LR_tfidf = LR_tfidf.predict_proba(x_te)[:, 1]
prob_SVC_tfidf = SVC_tfidf._predict_proba_lr(x_te)[:, 1]
prob_NB_count = NB_count.predict_proba(x_c)[:, 1]
prob_LR_count = LR_count.predict_proba(x_te1)[:, 1]
prob_SVC_count = SVC_count._predict_proba_lr(x_te1)[:, 1]
probabilities = [
prob_LR_tfidf, prob_LR_count, prob_SVC_tfidf, prob_SVC_count,
prob_NB_count
]
Type = [[], [], [], [], []]
response = []
threshold = [0.18, 0.17, 0.4, 0.4, 0.12]
print(probabilities)
for i in range(len(probabilities)):
for j in range(len(probabilities[i])):
if probabilities[i][j] >= threshold[i]:
Type[i].append("Insincere")
else:
Type[i].append("Sincere")
for i in range(len(questions)):
response_body = {
"text": questions[i],
"models": {
"LR_TFIDF": {
"type": Type[0][i],
"probability": probabilities[0][i]
},
"LR_COUNT": {
"type": Type[1][i],
"probability": probabilities[1][i]
},
"SVC_TFIDF": {
"type": Type[2][i],
"probability": probabilities[2][i]
},
"SVC_COUNT": {
"type": Type[3][i],
"probability": probabilities[3][i]
},
"NB": {
"type": Type[4][i],
"probability": probabilities[4][i]
}
}
}
response.append(response_body)
response = JsonResponse(response, safe=False)
return response
except Exception as exception:
return JsonResponse(
{
"message":
"Sorry something went wrong, counld not process your request. Try again",
"error": str(exception)
},
status=500)
