config.py

"""
Configuration file for the Multi-Agent Medical Chatbot

This file contains all the configuration parameters for the project.

If you want to change the LLM and Embedding model:

you can do it by changing all 'llm' and 'embedding_model' variables present in multiple classes below.

Each llm definition has unique temperature value relevant to the specific class. 
"""

import os
from dotenv import load_dotenv
from langchain_openai import AzureOpenAIEmbeddings, AzureChatOpenAI

# Load environment variables from .env file
load_dotenv()

class AgentDecisoinConfig:
    def __init__(self):
        self.llm = AzureChatOpenAI(
            deployment_name = os.getenv("deployment_name"),  # Replace with your Azure deployment name
            model_name = os.getenv("model_name"),  # Replace with your Azure model name
            azure_endpoint = os.getenv("azure_endpoint"),  # Replace with your Azure endpoint
            openai_api_key = os.getenv("openai_api_key"),  # Replace with your Azure OpenAI API key
            openai_api_version = os.getenv("openai_api_version"),  # Ensure this matches your API version
            temperature = 0.1  # Deterministic
        )

class ConversationConfig:
    def __init__(self):
        self.llm = AzureChatOpenAI(
            deployment_name = os.getenv("deployment_name"),  # Replace with your Azure deployment name
            model_name = os.getenv("model_name"),  # Replace with your Azure model name
            azure_endpoint = os.getenv("azure_endpoint"),  # Replace with your Azure endpoint
            openai_api_key = os.getenv("openai_api_key"),  # Replace with your Azure OpenAI API key
            openai_api_version = os.getenv("openai_api_version"),  # Ensure this matches your API version
            temperature = 0.7  # Creative but factual
        )

class WebSearchConfig:
    def __init__(self):
        self.llm = AzureChatOpenAI(
            deployment_name = os.getenv("deployment_name"),  # Replace with your Azure deployment name
            model_name = os.getenv("model_name"),  # Replace with your Azure model name
            azure_endpoint = os.getenv("azure_endpoint"),  # Replace with your Azure endpoint
            openai_api_key = os.getenv("openai_api_key"),  # Replace with your Azure OpenAI API key
            openai_api_version = os.getenv("openai_api_version"),  # Ensure this matches your API version
            temperature = 0.3  # Slightly creative but factual
        )
        self.context_limit = 20     # include last 20 messsages (10 Q&A pairs) in history

class RAGConfig:
    def __init__(self):
        self.vector_db_type = "qdrant"
        self.embedding_dim = 1536  # Add the embedding dimension here
        self.distance_metric = "Cosine"  # Add this with a default value
        self.use_local = True  # Add this with a default value
        self.vector_local_path = "./data/qdrant_db"  # Add this with a default value
        self.doc_local_path = "./data/docs_db"
        self.parsed_content_dir = "./data/parsed_docs"
        self.url = os.getenv("QDRANT_URL")
        self.api_key = os.getenv("QDRANT_API_KEY")
        self.collection_name = "medical_assistance_rag"  # Ensure a valid name
        self.chunk_size = 512  # Modify based on documents and performance
        self.chunk_overlap = 50  # Modify based on documents and performance
        # self.embedding_model = "text-embedding-3-large"
        # Initialize Azure OpenAI Embeddings
        self.embedding_model = AzureOpenAIEmbeddings(
            deployment = os.getenv("embedding_deployment_name"),  # Replace with your Azure deployment name
            model = os.getenv("embedding_model_name"),  # Replace with your Azure model name
            azure_endpoint = os.getenv("embedding_azure_endpoint"),  # Replace with your Azure endpoint
            openai_api_key = os.getenv("embedding_openai_api_key"),  # Replace with your Azure OpenAI API key
            openai_api_version = os.getenv("embedding_openai_api_version")  # Ensure this matches your API version
        )
        self.llm = AzureChatOpenAI(
            deployment_name = os.getenv("deployment_name"),  # Replace with your Azure deployment name
            model_name = os.getenv("model_name"),  # Replace with your Azure model name
            azure_endpoint = os.getenv("azure_endpoint"),  # Replace with your Azure endpoint
            openai_api_key = os.getenv("openai_api_key"),  # Replace with your Azure OpenAI API key
            openai_api_version = os.getenv("openai_api_version"),  # Ensure this matches your API version
            temperature = 0.3  # Slightly creative but factual
        )
        self.summarizer_model = AzureChatOpenAI(
            deployment_name = os.getenv("deployment_name"),  # Replace with your Azure deployment name
            model_name = os.getenv("model_name"),  # Replace with your Azure model name
            azure_endpoint = os.getenv("azure_endpoint"),  # Replace with your Azure endpoint
            openai_api_key = os.getenv("openai_api_key"),  # Replace with your Azure OpenAI API key
            openai_api_version = os.getenv("openai_api_version"),  # Ensure this matches your API version
            temperature = 0.5  # Slightly creative but factual
        )
        self.chunker_model = AzureChatOpenAI(
            deployment_name = os.getenv("deployment_name"),  # Replace with your Azure deployment name
            model_name = os.getenv("model_name"),  # Replace with your Azure model name
            azure_endpoint = os.getenv("azure_endpoint"),  # Replace with your Azure endpoint
            openai_api_key = os.getenv("openai_api_key"),  # Replace with your Azure OpenAI API key
            openai_api_version = os.getenv("openai_api_version"),  # Ensure this matches your API version
            temperature = 0.0  # factual
        )
        self.response_generator_model = AzureChatOpenAI(
            deployment_name = os.getenv("deployment_name"),  # Replace with your Azure deployment name
            model_name = os.getenv("model_name"),  # Replace with your Azure model name
            azure_endpoint = os.getenv("azure_endpoint"),  # Replace with your Azure endpoint
            openai_api_key = os.getenv("openai_api_key"),  # Replace with your Azure OpenAI API key
            openai_api_version = os.getenv("openai_api_version"),  # Ensure this matches your API version
            temperature = 0.3  # Slightly creative but factual
        )
        self.top_k = 5
        self.vector_search_type = 'similarity'  # or 'mmr'

        self.huggingface_token = os.getenv("HUGGINGFACE_TOKEN")

        self.reranker_model = "cross-encoder/ms-marco-TinyBERT-L-6"
        self.reranker_top_k = 3

        self.max_context_length = 8192  # (Change based on your need) # 1024 proved to be too low (retrieved content length > context length = no context added) in formatting context in response_generator code

        self.include_sources = True  # Show links to reference documents and images along with corresponding query response

        # ADJUST ACCORDING TO ASSISTANT'S BEHAVIOUR BASED ON THE DATA INGESTED:
        self.min_retrieval_confidence = 0.40  # The auto routing from RAG agent to WEB_SEARCH agent is dependent on this value

        self.context_limit = 20     # include last 20 messsages (10 Q&A pairs) in history

class MedicalCVConfig:
    def __init__(self):
        self.brain_tumor_model_path = "./agents/image_analysis_agent/brain_tumor_agent/models/brain_tumor_segmentation.pth"
        self.chest_xray_model_path = "./agents/image_analysis_agent/chest_xray_agent/models/covid_chest_xray_model.pth"
        self.skin_lesion_model_path = "./agents/image_analysis_agent/skin_lesion_agent/models/checkpointN25_.pth.tar"
        self.skin_lesion_segmentation_output_path = "./uploads/skin_lesion_output/segmentation_plot.png"
        self.llm = AzureChatOpenAI(
            deployment_name = os.getenv("deployment_name"),  # Replace with your Azure deployment name
            model_name = os.getenv("model_name"),  # Replace with your Azure model name
            azure_endpoint = os.getenv("azure_endpoint"),  # Replace with your Azure endpoint
            openai_api_key = os.getenv("openai_api_key"),  # Replace with your Azure OpenAI API key
            openai_api_version = os.getenv("openai_api_version"),  # Ensure this matches your API version
            temperature = 0.1  # Keep deterministic for classification tasks
        )

class SpeechConfig:
    def __init__(self):
        self.eleven_labs_api_key = os.getenv("ELEVEN_LABS_API_KEY")  # Replace with your actual key
        self.eleven_labs_voice_id = "21m00Tcm4TlvDq8ikWAM"    # Default voice ID (Rachel)

class ValidationConfig:
    def __init__(self):
        self.require_validation = {
            "CONVERSATION_AGENT": False,
            "RAG_AGENT": False,
            "WEB_SEARCH_AGENT": False,
            "BRAIN_TUMOR_AGENT": True,
            "CHEST_XRAY_AGENT": True,
            "SKIN_LESION_AGENT": True
        }
        self.validation_timeout = 300
        self.default_action = "reject"

class APIConfig:
    def __init__(self):
        self.host = "0.0.0.0"
        self.port = 8000
        self.debug = True
        self.rate_limit = 10
        self.max_image_upload_size = 5  # max upload size in MB

class UIConfig:
    def __init__(self):
        self.theme = "light"
        # self.max_chat_history = 50
        self.enable_speech = True
        self.enable_image_upload = True

class Config:
    def __init__(self):
        self.agent_decision = AgentDecisoinConfig()
        self.conversation = ConversationConfig()
        self.rag = RAGConfig()
        self.medical_cv = MedicalCVConfig()
        self.web_search = WebSearchConfig()
        self.api = APIConfig()
        self.speech = SpeechConfig()
        self.validation = ValidationConfig()
        self.ui = UIConfig()
        self.eleven_labs_api_key = os.getenv("ELEVEN_LABS_API_KEY")
        self.tavily_api_key = os.getenv("TAVILY_API_KEY")
        self.max_conversation_history = 20  # Include last 20 messsages (10 Q&A pairs) in history

# # Example usage
# config = Config()