Update app.py

app.py

@@ -1,169 +1,169 @@
-import os
-from PyPDF2 import PdfReader
-from langchain.text_splitter import RecursiveCharacterTextSplitter
-from langchain.vectorstores import FAISS
-from langchain.prompts import PromptTemplate
-from langchain.chains.question_answering import load_qa_chain
-from langchain_google_genai import GoogleGenerativeAIEmbeddings, ChatGoogleGenerativeAI
-from dotenv import load_dotenv
-import streamlit as st
-import re
-from google.generativeai import configure
-
-# Load environment variables
-load_dotenv()
-os.getenv("GOOGLE_API_KEY")
-
-# Configure Google Generative AI
-configure(api_key=os.getenv("GOOGLE_API_KEY"))
-
-def get_pdf_text(pdf_docs):
-    """Extract text and course details (title, description, and link) from the PDF."""
-    text = ""
-    course_details = []
-    for pdf in pdf_docs:
-        pdf_reader = PdfReader(pdf)
-        for page in pdf_reader.pages:
-            page_text = page.extract_text()
-            text += page_text
-            
-            # Extract the course titles and links (bold titles and underlined links)
-            courses = extract_course_details(page_text)
-            course_details.extend(courses)
-
-    return text, course_details
-
-def extract_course_details(page_text):
-    """Extract course title and link from the page text."""
-    course_details = []
-    # Regex to find bold titles and underlined links
-    title_pattern = r"(\*\*([A-Z\s]+)\*\*)(.*?)(http[s]?://[^\s]+)"
-    matches = re.findall(title_pattern, page_text)
-    
-    for match in matches:
-        title = match[1].strip()
-        description = match[2].strip()
-        link = match[3].strip()
-        
-        # Return tuple of course title, description, and link
-        course_details.append({
-            "title": title,
-            "description": description,
-            "link": link
-        })
-    
-    return course_details
-
-def get_text_chunks(text, course_details):
-    """Split the extracted text into chunks and append course links to the description."""
-    text_splitter = RecursiveCharacterTextSplitter(chunk_size=10000, chunk_overlap=1000)
-    
-    chunks = []
-    for course in course_details:
-        course_text = f"**Course Title**: {course['title']}\n"
-        course_text += f"[Course Link]({course['link']})\n"
-        course_text += f"**Description**: {course['description']}\n"
-        
-        text_chunks = text_splitter.split_text(course_text)
-        chunks.extend(text_chunks)
-    
-    return chunks
-
-def get_vector_store(text_chunks):
-    """Generate and store embeddings in a vector store."""
-    embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
-    vector_store = FAISS.from_texts(text_chunks, embedding=embeddings)
-    vector_store.save_local("embedding")
-    return vector_store
-
-def generate_embeddings_from_pdf(pdf_docs):
-    """Generate and save embeddings from the PDF course file."""
-    # Extract text and course details from the PDF
-    raw_text, course_details = get_pdf_text(pdf_docs)
-    
-    # Get text chunks with course title, link, and description
-    text_chunks = get_text_chunks(raw_text, course_details)
-    
-    # Generate and save the vector store
-    vector_store = get_vector_store(text_chunks)
-    
-    print(f"Embeddings generated and saved successfully.")
-    return vector_store
-
-def load_vector_store():
-    """Load pre-generated embeddings from FAISS."""
-    embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
-    vector_store = FAISS.load_local("embedding", embeddings, allow_dangerous_deserialization=True)
-    return vector_store
-
-def get_conversational_chain():
-    """Create a structured chain for processing search queries."""
-    system_prompt = """
-    You are an intelligent assistant helping users find the best free courses on data science, machine learning, and related fields.
-    When given a query, recommend courses by analyzing their relevance based on:
-    - Keywords
-    - Topics of interest
-    - User's goals (if provided)
-
-    Format your responses as:
-    - **Course Title**: <Title>
-    - [Course Link](<Link>)
-    - **Description**: <Brief Description>
-    - **Relevance**: <Why it is recommended>
-    Provide concise and actionable recommendations.
-    """
-
-    prompt_template = PromptTemplate(
-        template=system_prompt + "\nContext: {context}\nQuery: {query}",
-        input_variables=["context", "query"]
-    )
-    model = ChatGoogleGenerativeAI(model="gemini-pro", temperature=0.3)
-    chain = load_qa_chain(model, chain_type="stuff", prompt=prompt_template)
-
-    return chain
-
-def user_input(user_query, keywords):
-    """Process user input and search for relevant courses."""
-    vector_store = load_vector_store()
-    chain = get_conversational_chain()
-
-    # Construct search query
-    query_context = f"Keywords: {keywords}. Query: {user_query}."
-    docs = vector_store.similarity_search(query_context)
-
-    # Get recommendations
-    response = chain({
-        "input_documents": docs,
-        "context": "Analytics Vidhya free courses database.",
-        "query": query_context
-    }, return_only_outputs=True)
-
-    return response["output_text"]
-
-def main():
-    # Streamlit app UI
-    st.set_page_config("Smart Course Search", layout="wide")
-    st.title("Smart Course Search Tool")
-    st.write("Search for the most relevant free courses using natural language or keywords.")
-
-    # User inputs
-    user_query = st.text_input("Enter your search query or context (e.g., 'I want to learn deep learning')")
-    keywords = st.text_input("Enter specific keywords (comma-separated, e.g., 'NLP, data visualization')")
-
-    if st.button("Search Courses"):
-        if user_query or keywords:
-            with st.spinner("Searching for the best courses..."):
-                results = user_input(user_query, keywords)
-                st.success("Search Complete!")
-                
-                # Beautify and display the results
-                if results:
-                    formatted_results = results.replace("**", "<b>").replace("**", "</b>").replace("[", "<u>").replace("]", "</u>")
-                    st.markdown(formatted_results, unsafe_allow_html=True)
-                else:
-                    st.warning("No relevant courses found.")
-        else:
-            st.error("Please provide a query or keywords for searching.")
-
-if __name__ == "__main__":
-    main()
+import os
+from PyPDF2 import PdfReader
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.vectorstores import FAISS
+from langchain.prompts import PromptTemplate
+from langchain.chains.question_answering import load_qa_chain
+from langchain_google_genai import GoogleGenerativeAIEmbeddings, ChatGoogleGenerativeAI
+from dotenv import load_dotenv
+import streamlit as st
+import re
+from google.generativeai import configure
+
+# Load environment variables
+load_dotenv()
+os.getenv("GOOGLE_API_KEY")
+
+# Configure Google Generative AI
+configure(api_key=os.getenv("GOOGLE_API_KEY"))
+
+def get_pdf_text(pdf_docs):
+    """Extract text and course details (title, description, and link) from the PDF."""
+    text = ""
+    course_details = []
+    for pdf in pdf_docs:
+        pdf_reader = PdfReader(pdf)
+        for page in pdf_reader.pages:
+            page_text = page.extract_text()
+            text += page_text
+            
+            # Extract the course titles and links (bold titles and underlined links)
+            courses = extract_course_details(page_text)
+            course_details.extend(courses)
+
+    return text, course_details
+
+def extract_course_details(page_text):
+    """Extract course title and link from the page text."""
+    course_details = []
+    # Regex to find bold titles and underlined links
+    title_pattern = r"(\*\*([A-Z\s]+)\*\*)(.*?)(http[s]?://[^\s]+)"
+    matches = re.findall(title_pattern, page_text)
+    
+    for match in matches:
+        title = match[1].strip()
+        description = match[2].strip()
+        link = match[3].strip()
+        
+        # Return tuple of course title, description, and link
+        course_details.append({
+            "title": title,
+            "description": description,
+            "link": link
+        })
+    
+    return course_details
+
+def get_text_chunks(text, course_details):
+    """Split the extracted text into chunks and append course links to the description."""
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=10000, chunk_overlap=1000)
+    
+    chunks = []
+    for course in course_details:
+        course_text = f"**Course Title**: {course['title']}\n"
+        course_text += f"[Course Link]({course['link']})\n"
+        course_text += f"**Description**: {course['description']}\n"
+        
+        text_chunks = text_splitter.split_text(course_text)
+        chunks.extend(text_chunks)
+    
+    return chunks
+
+def get_vector_store(text_chunks):
+    """Generate and store embeddings in a vector store."""
+    embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
+    vector_store = FAISS.from_texts(text_chunks, embedding=embeddings)
+    vector_store.save_local("embedding")
+    return vector_store
+
+def generate_embeddings_from_pdf(pdf_docs):
+    """Generate and save embeddings from the PDF course file."""
+    # Extract text and course details from the PDF
+    raw_text, course_details = get_pdf_text(pdf_docs)
+    
+    # Get text chunks with course title, link, and description
+    text_chunks = get_text_chunks(raw_text, course_details)
+    
+    # Generate and save the vector store
+    vector_store = get_vector_store(text_chunks)
+    
+    print(f"Embeddings generated and saved successfully.")
+    return vector_store
+
+def load_vector_store():
+    """Load pre-generated embeddings from FAISS."""
+    embeddings = GoogleGenerativeAIEmbeddings(model="models/embedding-001")
+    vector_store = FAISS.load_local("./", embeddings, allow_dangerous_deserialization=True)
+    return vector_store
+
+def get_conversational_chain():
+    """Create a structured chain for processing search queries."""
+    system_prompt = """
+    You are an intelligent assistant helping users find the best free courses on data science, machine learning, and related fields.
+    When given a query, recommend courses by analyzing their relevance based on:
+    - Keywords
+    - Topics of interest
+    - User's goals (if provided)
+
+    Format your responses as:
+    - **Course Title**: <Title>
+    - [Course Link](<Link>)
+    - **Description**: <Brief Description>
+    - **Relevance**: <Why it is recommended>
+    Provide concise and actionable recommendations.
+    """
+
+    prompt_template = PromptTemplate(
+        template=system_prompt + "\nContext: {context}\nQuery: {query}",
+        input_variables=["context", "query"]
+    )
+    model = ChatGoogleGenerativeAI(model="gemini-pro", temperature=0.3)
+    chain = load_qa_chain(model, chain_type="stuff", prompt=prompt_template)
+
+    return chain
+
+def user_input(user_query, keywords):
+    """Process user input and search for relevant courses."""
+    vector_store = load_vector_store()
+    chain = get_conversational_chain()
+
+    # Construct search query
+    query_context = f"Keywords: {keywords}. Query: {user_query}."
+    docs = vector_store.similarity_search(query_context)
+
+    # Get recommendations
+    response = chain({
+        "input_documents": docs,
+        "context": "Analytics Vidhya free courses database.",
+        "query": query_context
+    }, return_only_outputs=True)
+
+    return response["output_text"]
+
+def main():
+    # Streamlit app UI
+    st.set_page_config("Smart Course Search", layout="wide")
+    st.title("Smart Course Search Tool")
+    st.write("Search for the most relevant free courses using natural language or keywords.")
+
+    # User inputs
+    user_query = st.text_input("Enter your search query or context (e.g., 'I want to learn deep learning')")
+    keywords = st.text_input("Enter specific keywords (comma-separated, e.g., 'NLP, data visualization')")
+
+    if st.button("Search Courses"):
+        if user_query or keywords:
+            with st.spinner("Searching for the best courses..."):
+                results = user_input(user_query, keywords)
+                st.success("Search Complete!")
+                
+                # Beautify and display the results
+                if results:
+                    formatted_results = results.replace("**", "<b>").replace("**", "</b>").replace("[", "<u>").replace("]", "</u>")
+                    st.markdown(formatted_results, unsafe_allow_html=True)
+                else:
+                    st.warning("No relevant courses found.")
+        else:
+            st.error("Please provide a query or keywords for searching.")
+
+if __name__ == "__main__":
+    main()