Update app.py

app.py

@@ -1,38 +1,71 @@
-from flask import Flask, request, render_template
-import transformers
-from parrot import Parrot
-import torch
-parrot = Parrot(model_tag="prithivida/parrot_paraphraser_on_T5")
-
-#from flask import Flask, request, render_template
-#import parrot
-#import transformers
-import time
+from flask import Flask, render_template, request
+from textblob import TextBlob
+import re
+import nltk
+from nltk.translate.bleu_score import sentence_bleu
+from nltk.corpus import wordnet
+from nltk.corpus import sentiwordnet as swn
+from nltk.sentiment import SentimentIntensityAnalyzer
 
 app = Flask(__name__)
 
-@app.route("/")
+@app.route('/')
 def index():
-    return render_template("index.html")
+    return render_template('index.html')
 
-@app.route("/paraphrase", methods=["POST"])
+@app.route('/paraphrase', methods=['POST'])
 def paraphrase():
-    text = request.form["text"]
-    model = request.form["model"]
-    
-    if model == "parrot":
-        start_time = time.time()
-        paraphrased_text = parrot.augment(input_phrase=text, use_gpu=False)
-        processing_time = time.time() - start_time
-        
-    elif model == "transformers":
-        start_time = time.time()
-        paraphrased_text = transformers.paraphrase(text)
-        processing_time = time.time() - start_time
-    
-    processing_percentage = 100 * processing_time / 30
-    
-    return render_template("paraphrase.html", original_text=text, paraphrased_text=paraphrased_text, processing_percentage=processing_percentage)
-
-if __name__ == "__main__":
-    app.run(host="0.0.0.0",port=7860)
+    input_text = request.form['input_text']
+    input_text = re.sub(r'[^\w\s]', '', input_text) # remove special characters
+
+    # Correct grammar using TextBlob
+    input_text = str(TextBlob(input_text).correct())
+
+    # Summarize the text using TextBlob
+    summarized_text = str(TextBlob(input_text).summarize())
+
+    # Paraphrase the text
+    paraphrased_text = generate_paraphrase(input_text)
+
+    # Emotion detection
+    emotion = detect_emotion(input_text)
+
+    # Named Entity Recognition
+    entities = nltk.ne_chunk(nltk.pos_tag(nltk.word_tokenize(input_text)))
+
+    # Part-of-Speech Tagging
+    pos_tags = nltk.pos_tag(nltk.word_tokenize(input_text))
+
+    # Sentiment Analysis
+    sentiment = SentimentIntensityAnalyzer().polarity_scores(input_text)
+
+    return render_template('index.html', input_text=input_text, summarized_text=summarized_text, paraphrased_text=paraphrased_text, entities=entities, pos_tags=pos_tags, sentiment=sentiment, emotion=emotion)
+
+def generate_paraphrase(text):
+    # Use TextBlob to generate paraphrased text
+    paraphrased_text = str(TextBlob(text).words)
+
+    # Custom synonyms
+    custom_synonyms = [('happy', 'joyful'), ('sad', 'unhappy')]
+    for syn in custom_synonyms:
+        paraphrased_text = paraphrased_text.replace(syn[0], syn[1])
+
+    return paraphrased_text
+
+def detect_emotion(text):
+    # Use SentiWordNet to detect emotion in text
+    emotions = []
+    words = nltk.word_tokenize(text)
+    for word in words:
+        synset = swn.senti_synsets(word)
+        if len(synset) > 0:
+            emotions.append(synset[0].pos_score() - synset[0].neg_score())
+    if emotions:
+        emotion = max(emotions)
+    else:
+        emotion = 0
+
+    return 'positive' if emotion > 0 else 'negative' if emotion < 0 else 'neutral'
+
+if __name__ == '__main__':
+    app.run(host="0.0.0.0",port=7860,debug=True)