Update app.py

app.py

@@ -1,162 +1,43 @@
+import os
 import gradio as gr
-import joblib
-import nltk
-from nltk.stem import WordNetLemmatizer
-from nltk.tokenize import word_tokenize
-from nltk.corpus import stopwords
-import string
-import numpy as np
+import google.generativeai as genai
 
-# Download required NLTK data
-nltk.download('punkt')
-nltk.download('wordnet')
-nltk.download('stopwords')
+# Configure the Gemini API with environment variable
+GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY")
+if not GOOGLE_API_KEY:
+    raise ValueError("GOOGLE_API_KEY environment variable not set. Please configure it in the Hugging Face Space settings.")
 
-# Define positive and negative word lists
-POSITIVE_WORDS = [
-    'great', 'excellent', 'amazing', 'wonderful', 'fantastic', 'superb', 'brilliant',
-    'awesome', 'outstanding', 'perfect', 'terrific', 'fabulous', 'marvelous', 'stellar',
-    'enjoyable', 'engaging', 'incredible', 'delightful', 'impressive', 'inspiring',
-    'touching', 'heartwarming', 'fun', 'entertaining', 'captivating', 'masterpiece',
-    'beautiful', 'charming', 'uplifting', 'memorable', 'well-done', 'well-made',
-    'well-acted', 'well-written', 'clever', 'smart', 'thoughtful', 'moving', 'hilarious',
-    'funny', 'laugh', 'loved', 'love', 'favorite', 'recommend', 'must-see', 'enjoyed',
-    'spectacular', 'breathtaking', 'refreshing', 'unique', 'creative', 'original',
-    'strong', 'powerful', 'emotional', 'satisfying', 'rewarding', 'solid', 'top-notch',
-    'phenomenal', 'riveting', 'absorbing', 'gripping', 'remarkable', 'exceptional',
-    'flawless', 'genius', 'well-crafted', 'well-executed', 'well-directed', 'well-cast',
-    'well-paced', 'well-shot', 'well-produced', 'well-developed', 'well-performed',
-    'enchanting', 'magical', 'bravo', 'enriching', 'enlightening', 'enjoy', 'pleasure',
-    'pleasing', 'pleased', 'sublime', 'exquisite', 'glorious', 'outstanding', 'superior',
-    'commendable', 'noteworthy', 'notable', 'admirable', 'commend', 'praise', 'applaud',
-    'acclaimed', 'acclaim', 'favorite', 'best', 'top', 'winner', 'winning', 'award-winning',
-    'blockbuster', 'hit', 'crowd-pleaser', 'must', 'mustwatch', 'mustsee', 'worthwhile',
-    'worth', 'enjoyment', 'joy', 'cheerful', 'positive', 'optimistic', 'hopeful',
-    'heartfelt', 'satisfy', 'satisfying', 'satisfied', 'impressed', 'impress', 'impressive',
-    'commend', 'commendable', 'commendation', 'applause', 'applaud', 'applauded',
-    'recommendation', 'recommended', 'recommend', 'favorite', 'favorites', 'fav', 'fave',
-    'gem', 'hidden gem', 'classic', 'timeless', 'iconic', 'legendary', 'epic', 'must-own',
-    'must-have', 'must-see', 'must-watch', 'must experience', 'must try', 'must buy',
-    'must read', 'must listen', 'must play', 'must visit', 'must go', 'must do',
-    'must attend', 'must eat', 'must drink', 'must taste', 'must feel', 'must love',
-    'must enjoy', 'must appreciate', 'must admire', 'must respect', 'must cherish',
-    'must treasure', 'must value', 'must honor', 'must celebrate', 'must embrace',
-    'must support', 'must encourage', 'must inspire', 'must motivate', 'must uplift',
-    'must empower', 'must enlighten', 'must educate', 'must inform', 'must entertain',
-    'must amuse', 'must delight', 'must please', 'must satisfy', 'must gratify',
-    'must fulfill', 'must enrich', 'must enhance', 'must improve', 'must better',
-    'must advance', 'must progress', 'must develop', 'must grow', 'must evolve',
-    'must transform', 'must change', 'must innovate', 'must create', 'must build',
-    'must make', 'must produce', 'must generate', 'must invent', 'must discover',
-    'must explore', 'must learn', 'must teach', 'must share', 'must give', 'must help',
-    'must serve', 'must care', 'must love', 'must like', 'must prefer', 'must choose',
-    'must select', 'must pick', 'must opt', 'must decide', 'must determine', 'must resolve',
-    'must solve', 'must fix', 'must repair', 'must mend', 'must heal', 'must cure',
-    'must treat', 'must prevent', 'must protect', 'must defend', 'must guard', 'must shield',
-    'must save', 'must rescue', 'must recover', 'must restore', 'must revive', 'must renew',
-    'must refresh', 'must rejuvenate', 'must revitalize', 'must energize', 'must invigorate',
-    'must stimulate', 'must excite', 'must thrill', 'must exhilarate', 'must inspire',
-    'must motivate', 'must encourage', 'must support', 'must help', 'must assist',
-    'must aid', 'must benefit', 'must profit', 'must gain', 'must win', 'must succeed',
-    'must achieve', 'must accomplish', 'must attain', 'must reach', 'must realize',
-    'must fulfill', 'must complete', 'must finish', 'must end', 'must conclude',
-    'must close', 'must wrap', 'must finalize', 'must settle', 'must resolve',
-    'must solve', 'must fix', 'must repair', 'must mend', 'must heal', 'must cure',
-    'must treat', 'must prevent', 'must protect', 'must defend', 'must guard', 'must shield',
-]
+genai.configure(api_key=GOOGLE_API_KEY)
 
-NEGATIVE_WORDS = [
-    'bad', 'worst', 'awful', 'terrible', 'horrible', 'atrocious', 'dreadful',
-    'abysmal', 'lousy', 'poor', 'subpar', 'mediocre', 'unacceptable', 'disappointing',
-    'boring', 'predictable', 'uninteresting', 'forgettable', 'tedious', 'slow',
-    'annoying', 'unbearable', 'painful', 'mess', 'flawed', 'weak', 'unconvincing',
-    'overrated', 'underwhelming', 'cliché', 'cliche', 'ridiculous', 'nonsense',
-    'unrealistic', 'waste', 'pointless', 'dull', 'unimpressive', 'cringe', 'cringeworthy',
-    'unoriginal', 'incoherent', 'confusing', 'unfunny', 'forced', 'flat', 'shallow',
-    'forgettable', 'unnecessary', 'unpleasant', 'unwatchable', 'cheesy', 'corny',
-    'painstaking', 'tedium', 'drag', 'lackluster', 'insipid', 'overlong', 'bloated',
-    'contrived', 'derivative', 'sloppy', 'amateurish', 'unfocused', 'awkward',
-    'unresolved', 'unfulfilled', 'unremarkable', 'unengaging', 'unappealing',
-    'unbelievable', 'unbalanced', 'unpolished', 'unrefined', 'unprofessional',
-    'unmemorable', 'unrelatable', 'unreal', 'unconvincing', 'unbearable', 'unimaginative',
-    'unintelligent', 'unnecessary', 'unpleasant', 'unrewarding', 'unsettling',
-    'unsubtle', 'unworthy', 'unwelcome', 'unwise', 'unwieldy', 'unworthy', 'unjustified',
-    'unjust', 'unforgivable', 'unforgiving', 'unfortunate', 'unfriendly', 'unfulfilled',
-    'unimpressed', 'uninspired', 'uninspiring', 'uninteresting', 'unlikable', 'unlucky',
-    'unmotivated', 'unoriginal', 'unpleasant', 'unrealistic', 'unsatisfying', 'unsuccessful',
-    'unsuitable', 'unsurprising', 'untalented', 'unwatchable', 'upsetting', 'useless',
-    'vapid', 'weak', 'worthless', 'yawn', 'disaster', 'disastrous', 'disgusting',
-    'distasteful', 'disturbing', 'dreary', 'embarrassing', 'excruciating', 'fail',
-    'failure', 'flop', 'garbage', 'hackneyed', 'hated', 'hate', 'illogical', 'inferior',
-    'irritating', 'lame', 'lacking', 'letdown', 'messy', 'monotonous', 'nonsensical',
-    'offensive', 'painful', 'pathetic', 'poorly', 'regret', 'regrettable', 'repetitive',
-    'shame', 'shameful', 'stupid', 'tiresome', 'trash', 'trite', 'unbearable', 'unconvincing',
-    'unimpressive', 'uninteresting', 'unlikable', 'unoriginal', 'unpleasant', 'unwatchable',
-    'waste', 'worthless'
-]
+# Use Gemini 1.5 Flash model
+model = genai.GenerativeModel('gemini-1.5-flash-latest')
 
-# Sentiment analyzer class
-class EnhancedSentimentAnalyzer:
-    def __init__(self):
-        self.lemmatizer = WordNetLemmatizer()
-        self.stop_words = set(stopwords.words('english'))
-        self.punctuation = set(string.punctuation)
-        self.positive_words = set(POSITIVE_WORDS)
-        self.negative_words = set(NEGATIVE_WORDS)
-        
-    def preprocess(self, text):
-        tokens = word_tokenize(text.lower())
-        processed_tokens = [
-            self.lemmatizer.lemmatize(token) for token in tokens
-            if token not in self.punctuation and token not in self.stop_words and len(token) > 2
-        ]
-        return ' '.join(processed_tokens)
-
-# Load the pre-trained model
-model = joblib.load('sentiment_model.pkl')
-
-# Analyze sentiment function
-def analyze_sentiment(text):
+def sentiment_analysis(text):
     if not text.strip():
-        return "Error: Please enter a non-empty review text."
-    
-    preprocessor = EnhancedSentimentAnalyzer()
-    processed_text = preprocessor.preprocess(text)
-    
-    proba = model.predict_proba([processed_text])[0]
-    prediction = model.predict([processed_text])[0]
+        return "Error: Please enter a non-empty text for analysis."
     
-    tokens = processed_text.split()
-    pos_words = [w for w in tokens if w in POSITIVE_WORDS]
-    neg_words = [w for w in tokens if w in NEGATIVE_WORDS]
+    prompt = f"""
+    Analyze the sentiment of the following text and respond in this format:
     
-    result = {
-        "prediction": "Positive" if prediction == 1 else "Negative",
-        "confidence": float(max(proba)),
-        "positive_words": list(set(pos_words)),
-        "negative_words": list(set(neg_words))
-    }
-    
-    # Format output
-    output = f"Prediction: {result['prediction']} (Confidence: {result['confidence']:.2f})\n"
-    if result['positive_words']:
-        output += f"Positive words found: {', '.join(result['positive_words'])}\n"
-    else:
-        output += "Positive words found: None\n"
-    if result['negative_words']:
-        output += f"Negative words found: {', '.join(result['negative_words'])}\n"
-    else:
-        output += "Negative words found: None\n"
-    
-    return output
+    Sentiment: [Positive / Negative / Neutral]
+    Reason: [Brief explanation of why you classified it this way]
+
+    Text: \"{text}\"
+    """
+
+    try:
+        response = model.generate_content(prompt)
+        return response.text.strip()
+    except Exception as e:
+        return f"Error: {str(e)}\nTip: Ensure your API key is valid at https://aistudio.google.com/"
 
 # Define Gradio interface
 iface = gr.Interface(
-    fn=analyze_sentiment,
-    inputs=gr.Textbox(lines=5, placeholder="Enter your movie review here...", label="Movie Review"),
+    fn=sentiment_analysis,
+    inputs=gr.Textbox(lines=5, placeholder="Enter text for sentiment analysis...", label="Text Input"),
     outputs=gr.Textbox(label="Sentiment Analysis Result"),
-    title="Movie Review Sentiment Analyzer",
-    description="Enter a movie review to predict its sentiment (Positive or Negative) using a pre-trained model."
+    title="Gemini Sentiment Analyzer",
+    description="Enter text to analyze its sentiment using Google's Gemini 1.5 Flash model. Provide a valid GOOGLE_API_KEY in the Hugging Face Space settings."
 )
 
 # Launch the interface