Update app.py

app.py

@@ -2,8 +2,6 @@ import requests
 from bs4 import BeautifulSoup
 import pandas as pd
 import gradio as gr
-from sklearn.feature_extraction.text import TfidfVectorizer
-from sklearn.metrics.pairwise import cosine_similarity
 import os
 from groq import Groq
 import creds  # Assuming creds.py holds your API key as creds.api_key
@@ -39,40 +37,68 @@ for course_card in soup.find_all('header', class_='course-card__img-container'):
 df = pd.DataFrame(courses)
 
 # Step 3: Initialize the Groq client and set the API key
-client = Groq(api_key=creds.api_key)
+client = Groq(api_key=creds.api_key)  # Properly passing the API key
 
 def search_courses(query):
     try:
-        # Step 4: Preprocessing query and course titles for TF-IDF
-        course_titles = df['title'].tolist()
-        course_titles.append(query)  # Add the query to the list of titles
+        print(f"Searching for: {query}")
+        print(f"Number of courses in database: {len(df)}")
 
-        # Using TF-IDF to vectorize the course titles and query
-        tfidf_vectorizer = TfidfVectorizer(stop_words='english')
-        tfidf_matrix = tfidf_vectorizer.fit_transform(course_titles)
+        # Prepare the prompt for Groq
+        prompt = f"""Given the following query: "{query}"
+        Please analyze the query and rank the following courses based on their relevance to the query. 
+        Prioritize courses from Analytics Vidhya. Provide a relevance score from 0 to 1 for each course.
+        Only return courses with a relevance score of 0.5 or higher.
+        Return the results in the following format:
+        Title: [Course Title]
+        Relevance: [Score]
+        
+        Courses:
+        {df['title'].to_string(index=False)}
+        """
 
-        # Compute cosine similarity between the query and course titles
-        cosine_similarities = cosine_similarity(tfidf_matrix[-1:], tfidf_matrix[:-1]).flatten()
+        print("Sending request to Groq...")
+        # Get response from Groq
+        response = client.chat.completions.create(
+            model="mixtral-8x7b-32768",  # Use the appropriate model
+            messages=[{"role": "system", "content": "You are an AI assistant specialized in course recommendations."},
+                      {"role": "user", "content": prompt}],
+            temperature=0.2,
+            max_tokens=1000
+        )
+        print("Received response from Groq")
 
-        # Get the top 10 relevant courses based on cosine similarity
-        top_indices = cosine_similarities.argsort()[-10:][::-1]
-        
-        # Step 5: Build results
+        # Parse Groq's response
         results = []
-        for index in top_indices:
-            relevance = cosine_similarities[index]
-            if relevance >= 0.5:  # Only consider courses with at least 50% relevance
-                course = df.iloc[index]
-                results.append({
-                    'title': course['title'],
-                    'image_url': course['image_url'],
-                    'course_link': course['course_link'],
-                    'score': round(relevance * 100, 2)  # Show relevance as percentage
-                })
+        print("Groq response content:")
+        print(response.choices[0].message.content)
         
-        return results if results else []
+        for line in response.choices[0].message.content.split('\n'):
+            if line.startswith('Title:'):
+                title = line.split('Title:')[1].strip()
+                print(f"Found title: {title}")
+            elif line.startswith('Relevance:'):
+                relevance = float(line.split('Relevance:')[1].strip())
+                print(f"Relevance for {title}: {relevance}")
+                if relevance >= 0.5:
+                    matching_courses = df[df['title'] == title]
+                    if not matching_courses.empty:
+                        course = matching_courses.iloc[0]
+                        results.append({
+                            'title': title,
+                            'image_url': course['image_url'],
+                            'course_link': course['course_link'],
+                            'score': relevance
+                        })
+                        print(f"Added course: {title}")
+                    else:
+                        print(f"Warning: Course not found in database: {title}")
+
+        print(f"Number of results found: {len(results)}")
+        return sorted(results, key=lambda x: x['score'], reverse=True)[:10]  # Return top 10 results
 
     except Exception as e:
+        print(f"An error occurred in search_courses: {str(e)}")
         return []
 
 def gradio_search(query):
@@ -84,7 +110,7 @@ def gradio_search(query):
             course_title = item['title']
             course_image = item['image_url']
             course_link = item['course_link']
-            relevance_score = item['score']
+            relevance_score = round(item['score'] * 100, 2)
             
             html_output += f'''
             <div class="course-card">
@@ -100,18 +126,17 @@ def gradio_search(query):
     else:
         return '<p class="no-results">No results found. Please try a different query.</p>'
 
-# Dark-themed CSS
+# Custom CSS for the Gradio interface
 custom_css = """
 body {
     font-family: Arial, sans-serif;
-    background-color: #121212;  /* Dark background */
-    color: #E0E0E0;  /* Light text color for dark background */
+    background-color: #000000;  /* Set background to black */
+    color: #ffffff;  /* Set text color to white for contrast */
 }
 .container {
     max-width: 800px;
     margin: 0 auto;
     padding: 20px;
-    color: #E0E0E0;
 }
 .results-container {
     display: flex;
@@ -119,9 +144,9 @@ body {
     justify-content: space-between;
 }
 .course-card {
-    background-color: #1E1E1E;  /* Darker card background */
+    background-color: white;
     border-radius: 8px;
-    box-shadow: 0 2px 4px rgba(0, 0, 0, 0.5);
+    box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);
     margin-bottom: 20px;
     overflow: hidden;
     width: 48%;
@@ -141,10 +166,10 @@ body {
 .course-info h3 {
     margin-top: 0;
     font-size: 18px;
-    color: #E0E0E0;  /* Light text color */
+    color: #333;
 }
 .course-info p {
-    color: #B0B0B0;  /* Slightly darker text color for contrast */
+    color: #666;
     font-size: 14px;
     margin-bottom: 10px;
 }
@@ -163,7 +188,7 @@ body {
 }
 .no-results {
     text-align: center;
-    color: #B0B0B0;
+    color: #666;
     font-style: italic;
 }
 """
@@ -177,10 +202,11 @@ iface = gr.Interface(
     description="Find the most relevant courses from Analytics Vidhya based on your query.",
     theme="huggingface",
     css=custom_css,
-    examples=[["machine learning for beginners"],
-              ["advanced data visualization techniques"],
-              ["python programming basics"], 
-              ["Business Analytics"]
+    examples=[
+        ["machine learning for beginners"],
+        ["advanced data visualization techniques"],
+        ["python programming basics"], 
+        ["Business Analytics"]
     ],
 )