Update app.py

app.py

@@ -0,0 +1,96 @@
+import streamlit as st
+from langchain_groq import ChatGroq
+from langchain_ollama import ChatOllama
+from langchain_community.document_loaders import PyMuPDFLoader
+from langchain.schema import Document
+from langchain_core.prompts import PromptTemplate
+from langchain_core.output_parsers import JsonOutputParser
+from pydantic import BaseModel, Field
+import os
+import json
+from prompt import REAG_SYSTEM_PROMPT, rag_prompt  # Import prompts
+
+# Hugging Face Spaces API Key Handling
+st.set_page_config(page_title="ReAG", layout="centered")
+
+# Set API Key using Hugging Face Secrets
+os.environ["GROQ_API_KEY"] = st.secrets["GROQ_API_KEY"]
+
+# Initialize LLM models
+llm_relevancy = ChatGroq(model="llama-3.3-70b-versatile", temperature=0)
+llm = ChatOllama(model="deepseek-r1:14b", temperature=0.6, max_tokens=3000)
+
+# Define schema for extracted content
+class ResponseSchema(BaseModel):
+    content: str = Field(..., description="Relevant content from the document")
+    reasoning: str = Field(..., description="Why this content was selected")
+    is_irrelevant: bool = Field(..., description="True if the content is irrelevant")
+
+class RelevancySchemaMessage(BaseModel):
+    source: ResponseSchema
+
+relevancy_parser = JsonOutputParser(pydantic_object=RelevancySchemaMessage)
+
+# Function to format document
+def format_doc(doc: Document) -> str:
+    return f"Document_Title: {doc.metadata.get('title', 'Unknown')}\nPage: {doc.metadata.get('page', 'Unknown')}\nContent: {doc.page_content}"
+
+# Extract relevant context function
+def extract_relevant_context(question, documents):
+    result = []
+    with st.spinner("🔍 Extracting relevant content from document..."):
+        for doc in documents:
+            formatted_documents = format_doc(doc)
+            system_prompt = f"{REAG_SYSTEM_PROMPT}\n\n# Available source\n\n{formatted_documents}"
+            prompt = f"""Determine if the 'Available source' content is sufficient to answer the QUESTION.
+            QUESTION: {question}
+            RESPONSE FORMAT (Strict JSON):
+            ```json
+            {{
+                "content": "Extracted relevant content",
+                "reasoning": "Why this was chosen",
+                "is_irrelevant": false
+            }}
+            ```
+            """
+            messages = [{"role": "system", "content": system_prompt},
+                        {"role": "user", "content": prompt}]
+            response = llm_relevancy.invoke(messages)
+            formatted_response = relevancy_parser.parse(response.content)
+            result.append(formatted_response)
+    
+    final_context = [item['content'] for item in result if not item['is_irrelevant']]
+    return final_context
+
+# Generate response using RAG Prompt
+def generate_response(question, final_context):
+    with st.spinner("📝 Generating AI response..."):
+        prompt = PromptTemplate(template=rag_prompt, input_variables=["question", "context"])
+        chain = prompt | llm
+        response = chain.invoke({"question": question, "context": final_context})
+        return response.content.split("\n\n")[-1]
+
+# Streamlit UI
+st.title("📚 RAG-based Knowledge Retrieval App")
+st.markdown("Upload a PDF and ask questions based on its content.")
+
+uploaded_file = st.file_uploader("📂 Upload PDF", type=["pdf"])
+if uploaded_file:
+    with st.spinner("📥 Uploading and processing PDF..."):
+        file_path = f"/tmp/{uploaded_file.name}"  
+        with open(file_path, "wb") as f:
+            f.write(uploaded_file.getbuffer())
+        
+        loader = PyMuPDFLoader(file_path)
+        docs = loader.load()
+        st.success("✅ PDF uploaded and processed successfully!")
+
+        question = st.text_input("❓ Ask a question about the document:")
+        if st.button("🚀 Get Answer"):
+            if question:
+                final_context = extract_relevant_context(question, docs)
+                if final_context:
+                    answer = generate_response(question, final_context)
+                    st.success(f"🧠 **Response:**\n\n{answer}")
+                else:
+                    st.warning("⚠️ No relevant information found in the document.")