app.py

import json
import numpy as np
from sentence_transformers import SentenceTransformer, util
import gradio as gr
import requests
from bs4 import BeautifulSoup
import time
import torch
import traceback

# Extended sample data with business analyst courses
SAMPLE_COURSES = [
    {
        "title": "Business Analysis and Process Management Specialization",
        "description": "Learn core business analysis skills, process mapping, and improvement techniques for organizational efficiency.",
        "link": "https://www.coursera.org/specializations/business-analysis-process-management",
        "source": "coursera"
    },
    {
        "title": "Business Analytics Fundamentals",
        "description": "Master data-driven decision making, Excel, SQL, and visualization tools for business analysis.",
        "link": "https://www.udemy.com/course/business-analytics-fundamentals",
        "source": "udemy"
    },
    {
        "title": "Agile Business Analysis Professional",
        "description": "Learn agile methodologies, user stories, and modern BA practices for software projects.",
        "link": "https://www.coursera.org/professional-certificates/agile-business-analysis",
        "source": "coursera"
    },
    # Original sample courses...
    {
        "title": "Python Programming for Beginners",
        "description": "Learn Python from scratch. Covers basic concepts, data structures, and programming fundamentals.",
        "link": "https://www.udemy.com/course/python-for-beginners",
        "source": "udemy"
    },
    {
        "title": "Machine Learning Specialization",
        "description": "Comprehensive machine learning course covering supervised learning, neural networks, and practical ML projects.",
        "link": "https://www.coursera.org/specializations/machine-learning",
        "source": "coursera"
    }
]

def scrape_courses(query):
    """
    Scrape courses based on search query from multiple sources
    """
    courses = []
    
    # Udemy API endpoint (you would need to register for API access)
    udemy_url = f"https://www.udemy.com/api-2.0/courses/?search={query}&price=price-free"
    
    # Coursera API endpoint (you would need to register for API access)
    coursera_url = f"https://api.coursera.org/api/courses.v1?q=search&query={query}&includes=free"
    
    try:
        # Here you would implement the actual API calls
        # For now, we'll filter the sample courses based on the query
        query_terms = query.lower().split()
        for course in SAMPLE_COURSES:
            if any(term in course['title'].lower() or term in course['description'].lower() 
                  for term in query_terms):
                courses.append(course)
        
        return courses
    except Exception as e:
        print(f"Error scraping courses: {e}")
        return []

def search_courses(query):
    if not query.strip():
        return "Please enter a search query."

    try:
        # Get relevant courses based on the query
        relevant_courses = scrape_courses(query)
        
        if not relevant_courses:
            return "No courses found for your search query."
            
        # Initialize the model
        device = torch.device('cpu')
        model = SentenceTransformer("all-MiniLM-L6-v2")
        model = model.to(device)
        
        # Generate embeddings
        course_descriptions = [f"{course['title']} {course['description']}" for course in relevant_courses]
        course_embeddings = model.encode(course_descriptions, convert_to_tensor=True)
        
        # Generate query embedding
        query_embedding = model.encode(f"course about {query}", convert_to_tensor=True)
        
        # Calculate similarities
        similarities = util.pytorch_cos_sim(query_embedding, course_embeddings)[0]
        top_indices = np.argsort((-similarities).numpy())[:5]
        
        results = []
        for idx in top_indices:
            similarity_score = similarities[idx].item()
            course = relevant_courses[idx]
            results.append({
                "Title": course["title"],
                "Description": course["description"],
                "Link": course["link"],
                "Source": course["source"],
                "Relevance": f"{similarity_score:.2%}"
            })
        
        return results
    except Exception as e:
        print(f"Search error: {str(e)}")
        traceback.print_exc()
        return []

def search_interface(query):
    try:
        print(f"\nSearching for: {query}")
        results = search_courses(query)
        
        if isinstance(results, str):
            return results
        if not results:
            return "No matching courses found. Please try a different search term."
            
        display_text = "\n\n".join(
            [f"**Title**: {result['Title']}\n\n**Description**: {result['Description']}\n\n**Source:** {result['Source']}\n\n**Relevance:** {result['Relevance']}\n\n[Go to course]({result['Link']})" 
             for result in results]
        )
        return display_text
    except Exception as e:
        traceback.print_exc()
        return f"An error occurred: {str(e)}"

# Create and launch the Gradio interface
iface = gr.Interface(
    fn=search_interface,
    inputs="text",
    outputs="markdown",
    title="Free Course Search Engine",
    description="Enter a topic or keywords to find relevant free courses from Udemy and Coursera.",
    examples=["Python", "Business Analyst", "Data Science", "Web Development"]
)

iface.launch(share=True)