Add flask API to the GitLab

FlaskAPI/review_prediction/Procfile

+web: gunicorn --bind 0.0.0.0:$PORT wsgi:app
\ No newline at end of file

FlaskAPI/review_prediction/README.md

+# flask-api
+* **Emotion Detection Flask API**
+    * Open the project emotion-recognition (Code -> flask-api -> emotion-recognition)
+    * Install the required libraries using `pip`.
+        * Flask - `pip install -U Flask`
+        * numpy - `pip install numpy`
+        * tenserflow - `pip install --user --upgrade tensorflow`
+        * Pillow - `pip install Pillow==2.2.1`
+        * Flask Cors - `pip install Flask-Cors`
+        * gunicorn - `pip install gunicorn`
+    * Download the pretrained model from : https://drive.google.com/file/d/1trPbUpfoVFMxssZxMPorqMbIekbnxhI4/view?usp=sharing
+    * Place the model inside the app folder. (Code -> flask-api -> emotion-recognition -> app)
+    * After successfully installing the dependencies run the API with `python wsgi.py`

FlaskAPI/review_prediction/app/main.py

+# Importing the libraries
+
+import time 
+import numpy as np
+from tensorflow import keras
+from flask import Flask, request, jsonify
+from flask_cors import CORS
+from keras.preprocessing.sequence import pad_sequences
+from keras.preprocessing.text import Tokenizer
+import pickle
+
+# Load Tensorflow and Keras Model
+
+print('Loading Model ...')
+model = keras.models.load_model('app/model.h5')
+print('Model Loaded')
+
+# Enabling Cors
+app = Flask(__name__)
+
+
+SEQUENCE_LENGTH = 300
+POSITIVE = "POSITIVE"
+NEGATIVE = "NEGATIVE"
+NEUTRAL = "NEUTRAL"
+SENTIMENT_THRESHOLDS = (0.6, 0.8)
+#tokenizer = Tokenizer()
+
+def conbinenegativeWords(comment):
+    notWord = 'not'
+    conbine = 'false'
+    tempWord = []
+    FinalText = ''
+    x = '0'
+    isSet ='false'
+    tokens = []
+    for eachWord in comment.split():
+        if eachWord in notWord:
+            if x in '0':
+                eachWord = 'ABC'
+                x = '1'
+                isSet = 'true'
+                tokens.append(eachWord)
+        else:
+            if x in '1':
+                tempWord.append(eachWord)
+                eachWord = ''
+                x = '0'
+                tokens.append(eachWord)
+                
+            else: 
+                tokens.append(eachWord)
+    if isSet in 'true':
+        
+        FinalText = " ".join(tokens)
+        
+        for tem in tempWord:
+            
+            FinalText = FinalText.replace('ABC','not'+tem)
+            
+        
+    else:
+        FinalText = comment
+    return FinalText     
+    
+    
+
+
+
+
+# loading
+with open('app/tokenizer.pickle', 'rb') as handle:
+    tokenizer = pickle.load(handle)
+
+def decode_sentiment(score, include_neutral=True):
+    if include_neutral:        
+        label = NEUTRAL
+        if score <= SENTIMENT_THRESHOLDS[0]:
+            label = NEGATIVE
+        elif score >= SENTIMENT_THRESHOLDS[1]:
+            label = POSITIVE
+
+        return label
+    else:
+        return NEGATIVE if score < 0.5 else POSITIVE
+
+#return label, score and Time
+def predict(text, include_neutral=True):
+    start_at = time.time()
+    # Tokenize text
+    x_test = pad_sequences(tokenizer.texts_to_sequences([text]), maxlen=SEQUENCE_LENGTH)
+    # Predict
+    score = model.predict([x_test])[0]
+    # Decode sentiment
+    label = decode_sentiment(score, include_neutral=include_neutral)
+
+    return {"label": label, "score": float(score),
+       "elapsed_time": time.time()-start_at}
+       
+#Return Only score      
+def predict_OnlyAccuracy(text, include_neutral=True):
+    start_at = time.time()
+    # Tokenize text
+    x_test = pad_sequences(tokenizer.texts_to_sequences([text]), maxlen=SEQUENCE_LENGTH)
+    # Predict
+    score = model.predict([x_test])[0]
+    # Decode sentiment
+    label = decode_sentiment(score, include_neutral=include_neutral)
+
+    return float(score)
+
+
+# Test Route
+@app.route('/')
+def index():
+    return "<h1>Review Prediction Flask API !!</h1>"
+
+# Review Prediction Route
+@app.route('/review_prediction',methods=['POST'])
+def review_prediction():
+    '''
+    For direct API calls trought request
+    '''
+    data = request.get_json(force=True)
+    start = time.time()
+    
+    reviewResultsList = []
+    
+    for var in data['review']:
+        reviewResultsList.append(predict(conbinenegativeWords(var)));
+	
+    # Getting the prediciton
+    #prediction_new = predict(conbinenegativeWords(data['review']));
+    
+    end = time.time()
+    
+    print('Prediction Time: ',end - start)
+
+    
+    return jsonify({"data":data, "result":reviewResultsList})
+
+
+# Review Prediction Route Get Final value to Tutor
+@app.route('/review_prediction_for_tutor',methods=['POST'])
+def review_prediction_for_tutor():
+    '''
+    For direct API calls trought request
+    '''
+    data = request.get_json(force=True)
+    start = time.time()
+    
+    reviewResultsList = []
+    totalScore = 0
+    
+    for var in data:
+        print(var);
+        reviewResultsList.append(predict_OnlyAccuracy(conbinenegativeWords(var)));
+	
+    for item in reviewResultsList:
+        totalScore = totalScore + item
+    
+    finalyPrecentage = (totalScore / len(reviewResultsList)) * 100
+    end = time.time()
+    
+    print('Prediction Time: ',end - start)
+
+    
+    return jsonify({"result":finalyPrecentage})

FlaskAPI/review_prediction/requirements.txt

+Flask
+numpy
+tensorflow==2.0
+Pillow
+flask_cors
+gunicorn
\ No newline at end of file

FlaskAPI/review_prediction/wsgi.py

+from app.main import app
+
+if __name__ == "__main__":
+    
+    app.run(debug=True)