app.py

import os
from dotenv import load_dotenv
import gradio as gr
from llama_index.core import StorageContext, load_index_from_storage, VectorStoreIndex, SimpleDirectoryReader, ChatPromptTemplate, Settings
from llama_index.llms.huggingface import HuggingFaceInferenceAPI
from llama_index.embeddings.huggingface import HuggingFaceEmbedding
from sentence_transformers import SentenceTransformer

load_dotenv()

# Configure the Llama index settings
Settings.llm = HuggingFaceInferenceAPI(
    model_name="meta-llama/Meta-Llama-3-8B-Instruct",
    tokenizer_name="meta-llama/Meta-Llama-3-8B-Instruct",
    context_window=3000,
    token=os.getenv("HF_TOKEN"),
    max_new_tokens=512,
    generate_kwargs={"temperature": 0.1},
)
Settings.embed_model = HuggingFaceEmbedding(
    model_name="BAAI/bge-small-en-v1.5"
)

# Define the directory for persistent storage and data
PERSIST_DIR = "db"
PDF_DIRECTORY = 'data'  # Changed to the directory containing PDFs

# Ensure directories exist
os.makedirs(PDF_DIRECTORY, exist_ok=True)
os.makedirs(PERSIST_DIR, exist_ok=True)

# Variable to store current chat conversation
current_chat_history = []

def data_ingestion_from_directory():
    # Use SimpleDirectoryReader on the directory containing the PDF files
    documents = SimpleDirectoryReader(PDF_DIRECTORY).load_data()
    storage_context = StorageContext.from_defaults()
    index = VectorStoreIndex.from_documents(documents)
    index.storage_context.persist(persist_dir=PERSIST_DIR)

def handle_query(query):
    chat_text_qa_msgs = [
        (
            "user",
            """
            You are E-Dermatologist, a friendly and knowledgeable AI assistant specializing in skin diseases. Your role is to provide helpful, clear, and precise responses to queries related to:
            - Diagnosis and detection of skin diseases
            - Symptoms and their possible causes
            - Prevention and safety measures for various skin conditions
            - Treatment options and remedies (including medical and home-based)
            - Guidance on when to consult a dermatologist

            Always ensure that your responses are:
            - User-friendly and easy to understand
            - Based on general medical knowledge (avoid prescribing medications)
            - Focused on awareness and prevention
            
            Use phrases like "From what I understand," or "Based on general guidelines," to make your responses conversational yet professional. 
            
            Quick tips:
            - Encourage users to consult a qualified dermatologist for a detailed diagnosis.
            - Avoid making definitive diagnoses or treatment recommendations.
            
            {context_str}
            Question:
            {query_str}
            """
        )
    ]


    text_qa_template = ChatPromptTemplate.from_messages(chat_text_qa_msgs)

    # Load index from storage
    storage_context = StorageContext.from_defaults(persist_dir=PERSIST_DIR)
    index = load_index_from_storage(storage_context)

    # Use chat history to enhance response
    context_str = ""
    for past_query, response in reversed(current_chat_history):
        if past_query.strip():
            context_str += f"User asked: '{past_query}'\nBot answered: '{response}'\n"

    query_engine = index.as_query_engine(text_qa_template=text_qa_template, context_str=context_str)
    answer = query_engine.query(query)

    if hasattr(answer, 'response'):
        response = answer.response
    elif isinstance(answer, dict) and 'response' in answer:
        response = answer['response']
    else:
          response = "Sorry, as per my current knowledge I am unable to answer this question. Is there anything else I can help you with?"

    # Remove sensitive information and unwanted sections from the response
    sensitive_keywords = [PERSIST_DIR, PDF_DIRECTORY, "/", "\\", ".pdf", ".doc", ".txt"]
    for keyword in sensitive_keywords:
        response = response.replace(keyword, "")

    # Remove sections starting with specific keywords
    unwanted_sections = ["Page Label","Page Label:","page_label","page_label:","file_path:","file_path",]
    for section in unwanted_sections:
        if section in response:
            response = response.split(section)[0]

    # Additional cleanup for any remaining artifacts from replacements
    response = ' '.join(response.split())

    # Update current chat history
    current_chat_history.append((query, response))

    return response

# Example usage: Process PDF ingestion from directory
print("Processing PDF ingestion from directory:", PDF_DIRECTORY)
data_ingestion_from_directory()

# Define the input and output components for the Gradio interface
input_component = gr.Textbox(
    show_label=False,
    placeholder="Your Query regarding Skin Diseases..."
)

output_component = gr.Textbox()

# Function to handle queries
def chatbot_handler(query):
    response = handle_query(query)
    return response

# Create the Gradio interface
interface = gr.Interface(
    fn=chatbot_handler,
    inputs=input_component,
    outputs=output_component,
    title="Welcome to E-Dermatologist",
    description="I am here to assist you with any questions you have about your skin Diseaases. How can I help you today?"
)

# Launch the Gradio interface
interface.launch()