ContentExtract/module/predictor.py

import os
from keras import models
import numpy as np
from module.Utils import *

current_path = os.path.dirname(__file__)

class ListpageContentPredictor():
    
    def __init__(self,file=""):
        if file=="":
            self.model_file = current_path+"/listpage/content/model/ep005-acc0.970-loss0.047-val_acc0.944-val_loss0.077.h5"
        else:
            self.model_file = current_path+"/listpage/content/model/"+file
        self.model = None
        self.getModel()
        self.graph = tf.get_default_graph()
        
    def getModel(self):
        if self.model is None:
            self.model = models.load_model(self.model_file, custom_objects={"acc":acc,"precision":precision,"recall":recall,"f1_score":f1_score,"my_loss":my_loss})
            self.model.load_weights(self.model_file)
            
        return self.model
    
    def predict(self,x):
        with self.graph.as_default():
            pre= self.getModel().predict(x)
        max_index = np.argmax(pre,1)[0][1]
        return max_index

class DetailContentPredictor():
    
    def __init__(self,file=""):
        if file=="":
            self.model_file = current_path+"/detail/content/model/ep011-loss0.160-val_acc0.900-val_loss0.156-f10.4536.h5"
        else:
            self.model_file = current_path+"/detail/content/model/"+file
        self.model = None
        self.getModel()
        self.graph = tf.get_default_graph()
        
        
        
    def getModel(self):
        if self.model is None:
            self.model = models.load_model(self.model_file, custom_objects={"acc":acc,"precision":precision,"recall":recall,"f1_score":f1_score,"my_loss":my_loss})
            self.model.load_weights(self.model_file)
        return self.model
    
    def predict(self,x):
        with self.graph.as_default():
            pre= self.getModel().predict(x)
        max_index = np.argmax(pre,1)[0][1]
        return max_index
    
class DetailTitlePredictor():
    
    def __init__(self,file=""):
        if file=="":
            self.model_file = current_path+"/detail/title/model/ep009-acc0.995-loss0.006-val_acc0.986-val_loss0.018.h5"
        else:
            self.model_file = current_path+"/detail/title/model/"+file
        self.model = None
        self.getModel()
        self.graph = tf.get_default_graph()
        
        
    def getModel(self):
        if self.model is None:
            self.model = models.load_model(self.model_file, custom_objects={"acc":acc,"precision":precision,"recall":recall,"f1_score":f1_score,"my_loss":my_loss})
            self.model.load_weights(self.model_file)
        return self.model
    
    def predict(self,x):
        with self.graph.as_default():
            pre= self.getModel().predict(x)
        max_index = np.argmax(pre,1)[0][1]
        return max_index