app.py

import gradio as gr
import matplotlib.pyplot as plt
import pandas as np
import tempfile
from datetime import datetime
from langchain_core.messages import HumanMessage
from tools import tools
from agents import *
from config import *
from workflow import create_workflow
import logging
import threading
import queue

# Global timeout variable
TIMEOUT_SECONDS = 300

# Initialize workflow
graph = create_workflow()

# Configure logging
logging.basicConfig(level=logging.DEBUG)
logger = logging.getLogger(__name__)

# Helper Functions
def run_graph(input_message, history, user_details):
    def invoke_workflow(q):
        try:
            # Determine if the query is general or personal
            if "how" in input_message.lower() or "what" in input_message.lower():
                general_query = True if "general" in input_message.lower() else False
            else:
                general_query = False

            system_prompt = (
                "You are a fitness and health assistant. "
                "If the user's query is personal, provide tailored advice based on their details, including BMI and caloric needs. "
                "If the query is general, respond broadly without relying on personal metrics. "
                "Encourage the user to ask follow-up questions and maintain a conversational tone."
            )

            # Summarize user details
            user_details_summary = (
                f"Name: {user_details.get('name', 'Unknown')}, "
                f"Age: {user_details.get('age', 'Unknown')}, "
                f"Gender: {user_details.get('gender', 'Unknown')}, "
                f"Weight: {user_details.get('weight', 'Unknown')} kg, "
                f"Height: {user_details.get('height', 'Unknown')} cm, "
                f"Activity Level: {user_details.get('activity_level', 'Unknown')}"
            )

            # Messages for the workflow
            messages = [
                {"role": "system", "content": system_prompt},
                {"role": "user", "content": f"User details: {user_details_summary}" if not general_query else "General query"},
                {"role": "user", "content": input_message}
            ]

            response = graph.invoke({"messages": messages})

            # Extract LLM response
            llm_response = None
            for msg in response.get("messages", []):
                if isinstance(msg, HumanMessage) and msg.name in ["nutritionist", "workout_coach", "general_expert"]:
                    llm_response = msg.content
                    break

            if llm_response:
                q.put(llm_response)
            else:
                q.put("The workflow did not return a valid response. Please try again.")
        except Exception as e:
            q.put(f"An error occurred: {e}")

    q = queue.Queue()
    thread = threading.Thread(target=invoke_workflow, args=(q,))
    thread.start()
    thread.join(timeout=TIMEOUT_SECONDS)

    if thread.is_alive():
        return f"The request took longer than {TIMEOUT_SECONDS} seconds and timed out. Please try again."
    return q.get()


def calculate_bmi(height, weight):
    height_m = height / 100
    bmi = weight / (height_m ** 2)
    if bmi < 18.5:
        status = "underweight"
    elif 18.5 <= bmi < 24.9:
        status = "normal weight"
    elif 25 <= bmi < 29.9:
        status = "overweight"
    else:
        status = "obese"
    return bmi, status


def visualize_bmi_and_calories(bmi, calories):
    categories = ["Underweight", "Normal Weight", "Overweight", "Obese"]
    bmi_values = [18.5, 24.9, 29.9, 40]
    calorie_range = [1500, 2000, 2500, 3000]

    fig, ax1 = plt.subplots(figsize=(10, 6))

    # BMI Visualization
    ax1.bar(categories, bmi_values, color=['blue', 'green', 'orange', 'red'], alpha=0.6, label="BMI Ranges")
    ax1.axhline(y=bmi, color='purple', linestyle='--', linewidth=2, label=f"Your BMI: {bmi:.2f}")
    ax1.set_ylabel("BMI Value")
    ax1.set_title("BMI and Caloric Needs Visualization")
    ax1.legend(loc="upper left")

    # Calorie Visualization
    ax2 = ax1.twinx()
    ax2.plot(categories, calorie_range, 'o-', color='magenta', label="Calorie Ranges")
    ax2.axhline(y=calories, color='cyan', linestyle='--', linewidth=2, label=f"Your Calorie Needs: {calories:.2f} kcal")
    ax2.set_ylabel("Calories")
    ax2.legend(loc="upper right")

    plt.tight_layout()

    # Save visualization to a temporary file
    temp_file = tempfile.NamedTemporaryFile(suffix=".png", delete=False)
    try:
        plt.savefig(temp_file.name)
    finally:
        plt.close()

    return temp_file.name


def calculate_calories(age, weight, height, activity_level, gender):
    if gender.lower() == "male":
        bmr = 10 * weight + 6.25 * height - 5 * age + 5
    else:
        bmr = 10 * weight + 6.25 * height - 5 * age - 161

    activity_multipliers = {
        "sedentary": 1.2,
        "lightly active": 1.375,
        "moderately active": 1.55,
        "very active": 1.725,
        "extra active": 1.9,
    }

    activity_level = activity_level.lower()
    return bmr * activity_multipliers.get(activity_level, 1.2)

# Interface Components
with gr.Blocks() as demo:
    gr.Markdown("<strong>FIT.AI - Your Fitness and Wellbeing Coach</strong>")

    with gr.Tabs():
        with gr.Tab("Visualization + Chat"):
            # User Input
            with gr.Row():
                user_name = gr.Textbox(placeholder="Enter your name", label="Name")
                user_age = gr.Number(label="Age (years)", value=25, precision=0)
                user_gender = gr.Dropdown(choices=["Male", "Female"], label="Gender", value="Male")
                user_weight = gr.Number(label="Weight (kg)", value=70, precision=1)
                user_height = gr.Number(label="Height (cm)", value=170, precision=1)
                activity_level = gr.Dropdown(
                    choices=["Sedentary", "Lightly active", "Moderately active", "Very active", "Extra active"],
                    label="Activity Level",
                    value="Moderately active"
                )

            # Visualization Output
            bmi_chart = gr.Image(label="BMI and Calorie Chart")

            # Chat Outputs
            with gr.Row():
                chatbot = gr.Chatbot(label="Chat with FIT.AI")
                text_input = gr.Textbox(placeholder="Type your question here...", label="Your Question")

            submit_button = gr.Button("Submit")
            clear_button = gr.Button("Clear Chat")

            def submit_message(message, history=[]):
                user_details = {
                    "name": user_name.value,
                    "age": user_age.value,
                    "weight": user_weight.value,
                    "height": user_height.value,
                    "activity_level": activity_level.value,
                    "gender": user_gender.value
                }

                bmi, status = calculate_bmi(user_details['height'], user_details['weight'])
                calories = calculate_calories(
                    user_details['age'], user_details['weight'], user_details['height'], user_details['activity_level'], user_details['gender']
                )
                chart_path = visualize_bmi_and_calories(bmi, calories)

                user_prompt = (
                    f"User wants advice on: {message}\n"
                    f"User Details:\n"
                    f"- Name: {user_details['name']}\n"
                    f"- Age: {user_details['age']}\n"
                    f"- Gender: {user_details['gender']}\n"
                    f"- Weight: {user_details['weight']} kg\n"
                    f"- Height: {user_details['height']} cm\n"
                    f"- Activity Level: {user_details['activity_level']}\n"
                    f"- BMI: {bmi:.2f} ({status})\n"
                    f"- Daily Caloric Needs: {calories:.2f} kcal\n"
                    f"\nProvide tailored advice based on these metrics."
                )

                response = run_graph(user_prompt, history, user_details)

                history.append(("User", message))
                if isinstance(response, str):
                    history.append(("FIT.AI", response + "\nLet me know if there's anything else you'd like to ask! 😊"))
                else:
                    history.append(("FIT.AI", "An unexpected response was received."))

                return history, chart_path

            submit_button.click(submit_message, inputs=[text_input, chatbot], outputs=[chatbot, bmi_chart])
            clear_button.click(lambda: ([], ""), inputs=None, outputs=[chatbot, bmi_chart])

        # Calculator + Visualization Tab
        with gr.Tab("Calculator + Visualization"):
            user_age_calc = gr.Number(label="Age (years)", value=25, precision=0)
            user_gender_calc = gr.Dropdown(choices=["Male", "Female"], label="Gender", value="Male")
            user_weight_calc = gr.Number(label="Weight (kg)", value=70, precision=1)
            user_height_calc = gr.Number(label="Height (cm)", value=170, precision=1)
            activity_level_calc = gr.Dropdown(
                choices=["Sedentary", "Lightly active", "Moderately active", "Very active", "Extra active"],
                label="Activity Level",
                value="Moderately active"
            )

            bmi_output = gr.Label(label="BMI Result")
            calorie_output = gr.Label(label="Calorie Needs")
            bmi_chart_calc = gr.Image(label="BMI and Calorie Chart")

            calculate_button = gr.Button("Calculate")

            def calculate_metrics(age, weight, height, gender, activity_level):
                bmi, status = calculate_bmi(height, weight)
                calories = calculate_calories(age, weight, height, activity_level, gender)
                chart_path = visualize_bmi_and_calories(bmi, calories)

                return f"Your BMI is {bmi:.2f}, considered {status}.", f"Daily calorie needs: {calories:.2f} kcal", chart_path

            calculate_button.click(
                calculate_metrics,
                inputs=[user_age_calc, user_weight_calc, user_height_calc, user_gender_calc, activity_level_calc],
                outputs=[bmi_output, calorie_output, bmi_chart_calc]
            )

demo.launch(share=True)