app.py

import pandas as pd
from sentence_transformers import SentenceTransformer
from sklearn.metrics.pairwise import cosine_similarity
import gradio as gr

# Debug: Print start of application
print("Starting the application...")

# Load the dataset from the same directory
print("Loading dataset...")
df = pd.read_csv('courses.csv')  # Assuming courses.csv is in the same directory as app.py
print(f"Dataset loaded. Number of rows: {df.shape[0]}")

# Load a pre-trained sentence transformer model
print("Loading Sentence Transformer model...")
model = SentenceTransformer('all-MiniLM-L6-v2')
print("Model loaded successfully.")

# Create a combined column for embedding (e.g., title + description + keywords)
print("Generating embeddings for courses...")
df['combined_text'] = df['title'] + " " + df['description'] + " " + df['keywords']
course_embeddings = model.encode(df['combined_text'].tolist(), convert_to_tensor=True)
print("Embeddings generated successfully.")

def search_courses(user_query):
    print(f"Received query: {user_query}")
    
    # Encode the user query
    query_embedding = model.encode(user_query, convert_to_tensor=True)

    # Compute cosine similarities between the query and each course embedding
    print("Calculating cosine similarities...")
    similarities = cosine_similarity(
        query_embedding.cpu().detach().numpy().reshape(1, -1),
        course_embeddings.cpu().detach().numpy()
    )

    # Get indices of top matching courses (top 5 results)
    top_matches = similarities.argsort()[0][-5:][::-1]
    
    # Retrieve top matching courses
    results = [{"title": df.iloc[i]["title"], "description": df.iloc[i]["description"]} for i in top_matches]
    print(f"Found {len(results)} results.")
    return results

# Define Gradio function for user interaction
def gradio_search(query):
    results = search_courses(query)
    return results

# Set up Gradio interface
print("Setting up Gradio interface...")
iface = gr.Interface(
    fn=gradio_search,
    inputs="text",
    outputs="json",
    title="Smart Course Search",
    description="Find the most relevant courses based on your query."
)

# Launch the app (for local testing or deploying in Hugging Face Spaces)
print("Launching the app...")
iface.launch()

# Debug: Print end of application
print("Application launched successfully.")