SmartSearch / app.py
abhiixxhek's picture
Update app.py
251f17c verified
import requests
from bs4 import BeautifulSoup
import pandas as pd
import gradio as gr
from groq import Groq
import os
from groq import Groq
from bs4 import BeautifulSoup
import requests
import pandas as pd
import gradio as gr
# Access the GROQ_API_KEY from environment variables
api_key = os.getenv("GROQ_API_KEY")
if not api_key:
raise ValueError("GROQ_API_KEY environment variable is not set.")
# Initialize Groq client with the API key
client = Groq(api_key=api_key)
# Rest of your code remains the same...
# Step 1: Scrape the free courses from Analytics Vidhya
url = "https://courses.analyticsvidhya.com/pages/all-free-courses"
response = requests.get(url)
soup = BeautifulSoup(response.content, 'html.parser')
courses = []
# Extracting course title, image, and course link
for course_card in soup.find_all('header', class_='course-card__img-container'):
img_tag = course_card.find('img', class_='course-card__img')
if img_tag:
title = img_tag.get('alt')
image_url = img_tag.get('src')
link_tag = course_card.find_previous('a')
if link_tag:
course_link = link_tag.get('href')
if not course_link.startswith('http'):
course_link = 'https://courses.analyticsvidhya.com' + course_link
courses.append({
'title': title,
'image_url': image_url,
'course_link': course_link
})
# Step 2: Create DataFrame
df = pd.DataFrame(courses)
client = Groq()
def search_courses(query):
try:
print(f"Searching for: {query}")
print(f"Number of courses in database: {len(df)}")
# 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)}
"""
print("Sending request to Groq...")
# Get response from Groq
response = client.chat.completions.create(
model="mixtral-8x7b-32768",
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")
# Parse Groq's response
results = []
print("Groq response content:")
print(response.choices[0].message.content)
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):
result_list = search_courses(query)
if result_list:
html_output = '<div class="results-container">'
for item in result_list:
course_title = item['title']
course_image = item['image_url']
course_link = item['course_link']
relevance_score = round(item['score'] * 100, 2)
html_output += f'''
<div class="course-card">
<img src="{course_image}" alt="{course_title}" class="course-image"/>
<div class="course-info">
<h3>{course_title}</h3>
<p>Relevance: {relevance_score}%</p>
<a href="{course_link}" target="_blank" class="course-link">View Course</a>
</div>
</div>'''
html_output += '</div>'
return html_output
else:
return '<p class="no-results">No results found. Please try a different query.</p>'
# Custom CSS for the Gradio interface
custom_css = """
body {
font-family: Arial, sans-serif;
background-color: #f0f2f5;
}
.container {
max-width: 800px;
margin: 0 auto;
padding: 20px;
}
.results-container {
display: flex;
flex-wrap: wrap;
justify-content: space-between;
}
.course-card {
background-color: white;
border-radius: 8px;
box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);
margin-bottom: 20px;
overflow: hidden;
width: 48%;
transition: transform 0.2s;
}
.course-card:hover {
transform: translateY(-5px);
}
.course-image {
width: 100%;
height: 150px;
object-fit: cover;
}
.course-info {
padding: 15px;
}
.course-info h3 {
margin-top: 0;
font-size: 18px;
color: #333;
}
.course-info p {
color: #666;
font-size: 14px;
margin-bottom: 10px;
}
.course-link {
display: inline-block;
background-color: #007bff;
color: white;
padding: 8px 12px;
text-decoration: none;
border-radius: 4px;
font-size: 14px;
transition: background-color 0.2s;
}
.course-link:hover {
background-color: #0056b3;
}
.no-results {
text-align: center;
color: #666;
font-style: italic;
}
"""
# Gradio interface
iface = gr.Interface(
fn=gradio_search,
inputs=gr.Textbox(label="Enter your search query", placeholder="e.g., machine learning, data science, python"),
outputs=gr.HTML(label="Search Results"),
title="Analytics Vidhya Smart Course Search",
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"]
],
)
if __name__ == "__main__":
iface.launch()