Create app.py

app.py

@@ -0,0 +1,50 @@
+from sklearn.feature_extraction.text import TfidfVectorizer
+from sklearn.metrics.pairwise import cosine_similarity
+import numpy as np
+import gradio as gr
+
+# Step 1: Collect FAQs - 
+faqs = {
+    "What is CodeAlpha?": "CodeAlpha is a tech platform offering free virtual internships in AI, Web Development, and Data Science to help students gain practical experience.",
+    "How long is CodeAlpha internship?": "CodeAlpha internships last 1 month. You must complete minimum 2 out of 4 tasks to receive the completion certificate and Letter of Recommendation.",
+    "When is CodeAlpha submission deadline?": "Submission window: 10 May 2026 to 10 June 2026. Last date: 10 June 2026. Certificates issued: 11 June 2026.",
+    "Is CodeAlpha certificate valuable?": "Yes, CodeAlpha certificates are recognized and the LOR helps in job applications. It's a verified virtual internship experience.",
+    "How to contact CodeAlpha?": "Website: www.codealpha.tech | WhatsApp: +91 9336576683 | Tasks are shared via WhatsApp group.",
+    "What is TF-IDF?": "TF-IDF stands for Term Frequency-Inverse Document Frequency. It's an NLP technique used to convert text to numerical vectors for similarity matching.",
+    "What is cosine similarity?": "Cosine similarity measures the cosine of angle between two vectors. Used in NLP to find how similar two text documents are. Score from 0 to 1.",
+    "What is NLTK used for?": "NLTK is Natural Language Toolkit. Used for text preprocessing: tokenization, stemming, removing stopwords, and cleaning text data."
+}
+
+questions = list(faqs.keys())
+answers = list(faqs.values())
+
+# Step 2: Preprocess using TF-IDF -// tokenize + clean automatically
+vectorizer = TfidfVectorizer(stop_words='english')
+tfidf_matrix = vectorizer.fit_transform(questions)
+
+# Step 3: Match user questions using cosine similarity
+def chatbot_response(message, history):
+    if not message.strip():
+        return "Please ask me a question about CodeAlpha or NLP concepts!"
+
+    user_vec = vectorizer.transform([message])
+    similarity = cosine_similarity(user_vec, tfidf_matrix)
+    idx = np.argmax(similarity)
+    confidence = similarity[0][idx]
+
+    # Step 4: Display best matching answer
+    if confidence > 0.3:
+        return f"🤖 {answers[idx]}\n\nMatch Score: {confidence:.0%}"
+    else:
+        return "🤖 I don't have info on that. Try asking about CodeAlpha internship, TF-IDF, or cosine similarity."
+
+# Step 5: Create simple chat UI
+demo = gr.ChatInterface(
+    fn=chatbot_response,
+    title="CodeAlpha FAQ Chatbot - Task 2",
+    description="Ask me about CodeAlpha internship or NLP concepts! Uses TF-IDF + Cosine Similarity",
+    examples=["What is CodeAlpha?", "When is submission deadline?", "What is TF-IDF?"],
+    theme=gr.themes.Soft()
+)
+
+demo.launch(share=True)