app.py

import gradio as gr
from PIL import Image
import os

from chatbot_updated import (
    chatbot_updated,
    detect_artifact,
    text_to_speech,
    cleanup_audio_file
)
import os

print("CURRENT DIR:", os.getcwd())
print("FILES:", os.listdir())
print("DATA EXISTS:", os.path.exists("data"))

# =========================
# CORE FUNCTION
# =========================
def process(input_text, image, audio):
    """
    Handles text + image + audio input
    """

    # =========================
    # 1. PRIORITY: AUDIO
    # =========================
    question = input_text

    if audio is not None:
        # gradio gives file path for audio
        audio_bytes = open(audio, "rb").read()
        question = audio_bytes

    # =========================
    # 2. IMAGE HANDLING
    # =========================
    img = None
    detected_name = None

    if image is not None:
        img = Image.open(image).convert("RGB")
        detected_name, annotated = detect_artifact(img)
    else:
        annotated = None

    # =========================
    # 3. CHATBOT CALL
    # =========================
    answer = chatbot_updated(question, image=img)
    answer = str(answer).strip()

    # =========================
    # 4. TEXT TO SPEECH
    # =========================
    audio_file = text_to_speech(answer)
    audio_out = None

    if audio_file and os.path.exists(audio_file):
        audio_out = audio_file

    # =========================
    # RETURN UI OUTPUTS
    # =========================
    return (
        answer,          # text output
        annotated,       # image output
        audio_out       # audio output
    )


# =========================
# GRADIO UI
# =========================
with gr.Blocks(title="🏺 Egyptian Artifact Chatbot") as app:

    gr.Markdown("# 🏺 Egyptian Artifact Chatbot (Gradio Version)")
    gr.Markdown("Ask using **text, voice, or image**")

    with gr.Row():
        with gr.Column():
            input_text = gr.Textbox(label="💬 Ask a question")

            audio_input = gr.Audio(
                sources=["microphone"],
                type="filepath",
                label="🎤 Voice Input"
            )

            image_input = gr.Image(
                type="filepath",
                label="📷 Upload Image"
            )

            btn = gr.Button("🚀 Get Answer")

        with gr.Column():
            output_text = gr.Textbox(label="🤖 Answer")
            output_image = gr.Image(label="🖼️ Detected Artifact")
            output_audio = gr.Audio(label="🔊 Speech Output")

    # =========================
    # BUTTON CLICK
    # =========================
    btn.click(
        fn=process,
        inputs=[input_text, image_input, audio_input],
        outputs=[output_text, output_image, output_audio]
    )

# =========================
# RUN APP
# =========================
if __name__ == "__main__":
    app.launch()