app.py

from flask import Flask, jsonify, request
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from keras.preprocessing.text import Tokenizer
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.model_selection import train_test_split
from sklearn.svm import LinearSVC
from sklearn.pipeline import Pipeline
from sklearn.model_selection import GridSearchCV
from sklearn.metrics import accuracy_score
from sklearn.linear_model import LogisticRegression
from tqdm.notebook import tqdm
from nltk.stem.snowball import SnowballStemmer
import warnings
warnings.filterwarnings('ignore')

app = Flask(__name__)

@app.route('/')
def home():
    return 'Welcome to the MBTI Prediction API'

@app.route('/predict', methods=['POST'])
def predict():
    # POST 요청에서 데이터 추출
    data = request.json

    # 데이터를 기반으로 알고리즘 실행
    result = run_algorithm(data)

    # 결과를 JSON 형식으로 반환
    return jsonify(result)

def run_algorithm(data):
    train = pd.read_csv('c:\\Users\\신유민\\Desktop\\MBTI 500.csv', encoding='utf-8')
    train.head()
    train.tail()
    train
    test = train.drop(['type'], axis=1)
    test.head()

    # 설명변수
    X = train['posts']
    # 예측변수
    Y = train['type']

    print(f"{len(train['type'].unique())}개")
    train['type'].value_counts()

    # plt.pie(train['type'])
    train.isnull().sum()
    train['posts'].nunique() == len(train['posts'])

    print("train data : ", train.shape)
    print("test data : ", test.shape)

    # E, I 빈도수 확인
    first = []
    for i in range(len(train)):
        first.append(train['type'][i][0])
    first = pd.DataFrame(first)
    first[0].value_counts()

    # N, S 빈도수 확인
    second = []
    for i in range(len(train)):
        second.append(train['type'][i][1])
    second = pd.DataFrame(second)
    second[0].value_counts()

    # T, F 빈도수 확인
    third = []
    for i in range(len(train)):
        third.append(train['type'][i][2])
    third = pd.DataFrame(third)
    third[0].value_counts()

    # P, J 빈도수 확인
    fourth = []
    for i in range(len(train)):
        fourth.append(train['type'][i][3])
    fourth = pd.DataFrame(fourth)
    fourth[0].value_counts()

    X = train['posts'] # data features
    Y = train['type'] # labels
    X_train, X_valid, y_train, y_valid = train_test_split(X, Y, test_size=0.2, random_state=1) # test size = 0.3도 해보기

    # 벡터화
    tfidf = TfidfVectorizer()

    # 훈련 데이터 벡터화
    X_train_tfidf = tfidf.fit_transform(X_train)

    clf = LinearSVC()
    # 정확도 기준 설정
    cv = GridSearchCV(clf, {'C': [0.35, 0.4, 0.45]}, scoring = "accuracy")
    cv.fit(X_train_tfidf, y_train)

    C = cv.best_params_['C']
    print("최적의 파라미터 C: ", C)

    text_clf = Pipeline([('tfidf', TfidfVectorizer()), ('clf', LinearSVC(C=C))])
    text_clf.fit(X_train, y_train)

    # valid 데이터의 mbti 예측
    pred = text_clf.predict(X_valid)

    # valid data에서의 정확도
    accuracy = accuracy_score(pred, y_valid)
    print(f'Accuracy: {accuracy}')

     # 결과 반환
    return pred

if __name__ == '__main__':
    app.run()