Update app.py

app.py

@@ -6,21 +6,21 @@ from PyPDF2 import PdfReader
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain.embeddings import HuggingFaceEmbeddings
 from langchain.vectorstores import FAISS
-from transformers import AutoModel, AutoTokenizer
+from transformers import pipeline, AutoModel, AutoTokenizer
 import torch
 
-# Set up Groq API key
-GROQ_API_KEY = os.getenv("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")
+# Load the summarization pipeline model
+@st.cache_resource
+def load_summarization_pipeline():
+    summarizer = pipeline("summarization", model="facebook/bart-large-cnn")  # Use a summarization model
+    return summarizer
 
-# Initialize embedding model (using sentence-transformers model)
-tokenizer = AutoTokenizer.from_pretrained("sentence-transformers/all-MiniLM-L6-v2")
-embedding_model = AutoModel.from_pretrained("sentence-transformers/all-MiniLM-L6-v2")
+summarizer = load_summarization_pipeline()
 
 # List of Hugging Face PDF URLs
 PDF_URLS = [
     "https://huggingface.co/spaces/tahirsher/GenAI_Lawyers_Guide/blob/main/administrator92ada0936848e501425591b4ad0cd417.pdf",
     "https://huggingface.co/spaces/tahirsher/GenAI_Lawyers_Guide/blob/main/Pakistan%20Penal%20Code.pdf",
-    # Add more document links as needed
 ]
 
 # Helper function to convert Hugging Face blob URLs to direct download URLs
@@ -33,7 +33,7 @@ def get_huggingface_raw_url(url):
 def fetch_pdf_text_from_huggingface(urls):
     text = ""
     for url in urls:
-        raw_url = get_huggingface_raw_url(url)  # Convert to direct download link
+        raw_url = get_huggingface_raw_url(url)
         response = requests.get(raw_url)
         if response.status_code == 200:
             pdf_file = BytesIO(response.content)
@@ -65,31 +65,17 @@ def load_or_create_vector_store(text_chunks):
     vector_store = FAISS.from_texts(text_chunks, embedding=embedding_function)
     return vector_store
 
-# Call Groq API for generating summary based on the query and retrieved text
-def generate_summary_with_groq(query, retrieved_text):
-    url = "https://api.groq.com/v1/chat/completions"
-    headers = {
-        "Authorization": f"Bearer {GROQ_API_KEY}",
-        "Content-Type": "application/json"
-    }
-    payload = {
-        "messages": [
-            {"role": "user", "content": f"{query}\n\nRelated information:\n{retrieved_text}"}
-        ],
-        "model": "llama3-8b-8192",
-    }
-    response = requests.post(url, headers=headers, json=payload)
-    if response.status_code == 200:
-        return response.json()["choices"][0]["message"]["content"]
-    else:
-        st.error("Failed to generate summary with Groq API")
-        return "Error in Groq API response"
+# Generate summary based on the retrieved text
+def generate_summary_with_huggingface(query, retrieved_text):
+    summarization_input = f"{query}\n\nRelated information:\n{retrieved_text}"
+    summary = summarizer(summarization_input, max_length=200, min_length=50, do_sample=False)
+    return summary[0]["summary_text"]
 
 # Generate response for user query
 def user_input(user_question, vector_store):
     docs = vector_store.similarity_search(user_question)
     context_text = " ".join([doc.page_content for doc in docs])
-    return generate_summary_with_groq(user_question, context_text)
+    return generate_summary_with_huggingface(user_question, context_text)
 
 # Main function to run the Streamlit app
 def main():