import streamlit as st
from gradio_client import Client
from st_audiorec import st_audiorec
from gtts import gTTS
from IPython.display import Audio, display

# Constants
TITLE = "AgriTure"
DESCRIPTION = """
----
This Project demonstrates a model fine-tuned by Achyuth. This Model is named as "AgriaTure". This Model helps the farmers and scientists to develop the art of agriculture and farming.
Hope this will be a Successful Project!!!
~Achyuth
----
"""

# Initialize client
whisper_client = Client("https://sanchit-gandhi-whisper-large-v2.hf.space/")

# Function to convert text to speech using gTTS
def text_to_speech(text, lang='en'):
    tts = gTTS(text=text, lang=lang, slow=False)
    tts.save("response.mp3")
    return "response.mp3"

# Function to transcribe audio
def transcribe(wav_path):
    return whisper_client.predict(
        wav_path,
        "transcribe",
        api_name="/predict"
    )

# Prediction function
def predict(message, system_prompt='Your name is AgriaTure...', temperature=0.7, max_new_tokens=4096, Topp=0.5, Repetitionpenalty=1.2):
    with st.status("Starting client"):
        client = Client("https://huggingface-projects-llama-2-7b-chat.hf.space/")
        st.write("Requesting Audio Transcriber")
    with st.status("Requesting AgriTure v1"):
        st.write("Requesting API")
        response = client.predict(
            message,
            system_prompt,
            max_new_tokens,
            temperature,
            Topp,
            500,
            Repetitionpenalty,
            api_name="/chat"
        )
        st.write("Done")
        return response

# Streamlit UI
st.title(TITLE)
st.write(DESCRIPTION)

if "messages" not in st.session_state:
    st.session_state.messages = []

# Display chat messages from history on app rerun
for message in st.session_state.messages:
    with st.chat_message(message["role"], avatar=("🧑‍💻" if message["role"] == 'human' else '🦙')):
        st.markdown(message["content"])

textinput = st.chat_input("Ask AgriTure anything...")
wav_audio_data = st_audiorec()

if wav_audio_data is not None:
    with st.status("Transcribing audio..."):
        # save audio
        with open("audio.wav", "wb") as f:
            f.write(wav_audio_data)
        prompt = transcribe("audio.wav")
        st.write("Transcribed Given Audio ✔")

    st.chat_message("human", avatar="🌿").markdown(prompt)
    st.session_state.messages.append({"role": "human", "content": prompt})

    # transcribe audio
    response = predict(message=prompt)

    with st.chat_message("assistant", avatar='🌿'):
        st.markdown(response)

    # Convert AI response to speech
    speech_file = text_to_speech(response)

    # Display assistant response in chat message container
    with st.chat_message("assistant", avatar='🌿'):
        st.markdown(response)

    # Play the generated speech
    display(Audio(speech_file, autoplay=True))

    # Add assistant response to chat history
    st.session_state.messages.append({"role": "assistant", "content": response})

# React to user input
if prompt := textinput:
    # Display user message in chat message container
    st.chat_message("human", avatar="🌿").markdown(prompt)
    # Add user message to chat history
    st.session_state.messages.append({"role": "human", "content": prompt})

    response = predict(message=prompt)

    # Convert AI response to speech
    speech_file = text_to_speech(response)

    # Display assistant response in chat message container
    with st.chat_message("assistant", avatar='🌿'):
        st.markdown(response)

    # Play the generated speech
    display(Audio(speech_file, autoplay=True))

    # Add assistant response to chat history
    st.session_state.messages.append({"role": "assistant", "content": response})