# coding=utf8
from gpt_index import SimpleDirectoryReader, GPTListIndex, GPTSimpleVectorIndex, LLMPredictor, PromptHelper
from langchain import OpenAI
import openai
import gradio as gr
import random
import time
import sys
import os
from transformers import pipeline

p = pipeline("automatic-speech-recognition")

#os.environ["OPENAI_API_KEY"] = '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'
os.environ["OPENAI_API_KEY"] = 'k-XxiwN5pp1VibrIo3Ntw5T3BlbkFJWFa8B5c9BIPzDwGwEUYB'

css = """

#component-2 {position: absolute; bottom: 0;    width: 100%;
}
.app.svelte-ac4rv4>.main.svelte-ac4rv4 {
    display: flex;
    flex-grow: 1;
    flex-direction: column;
    background-image: url(https://i.ibb.co/xj8R4r3/background-vertical.png);
}
div.svelte-1frtwj3 {
    display: inline-flex;
    align-items: center;}

div.float.svelte-1frtwj3 {
    position: absolute;
    opacity: 0;
    top: var(--block-label-margin);
    left: var(--block-label-margin);}

.wrap.svelte-6roggh.svelte-6roggh {
 adding: var(--block-padding);
    height: 100%;
    max-height: 800px;
    overflow-y: auto;
    }

.bot.svelte-6roggh.svelte-6roggh, .pending.svelte-6roggh.svelte-6roggh {
    border-color: var(--border-color-accent);
    background-color: var(--color-accent-soft);
    color: white;
    font-family: initial;
    font-style: italic;
    font: message-box;
    font-weight: bold;
}
div.svelte-1frtwj3 {
    display: inline-flex;
    align-items: center;
    z-index: var(--layer-2);
    box-shadow: var(--block-shadow);
    border: var(--block-label-border-width) solid #ffffff;
    border-top: none;
    border-left: none;
    border-radius: var(--block-label-radius);
    background: #eff6ff;
    padding: var(--block-label-padding);
    pointer-events: none;
    color: var(--block-label-text-color);
    font-weight: var(--block-label-text-weight);
    width: 100%;
    line-height: var(--line-sm);
    }
div.svelte-awbtu4 {
    display: flex;
    flex-direction: inherit;
    flex-wrap: wrap;
    gap: var(--form-gap-width);
    box-shadow: var(--block-shadow);
    border: var(--block-border-width) solid #5f0000;
    border-radius: var(--radius-lg);
    background: #ffffff;
    overflow: hidden;
    position: fixed;
    bottom: 0;
    margin-left: -16px;
}

img.svelte-ms5bsk {
    width: var(--size-full);
    height: 90px;
    object-fit: contain;
}
.app.svelte-ac4rv4.svelte-ac4rv4 {
    max-width: none;
    background-color: #ffffff;
}
.app.svelte-ac4rv4.svelte-ac4rv4{max-width:none}
.wrap.svelte-1o68geq.svelte-1o68geq {max-height: none}
.block.svelte-mppz8v {
    position: relative;
    margin: 0;
    box-shadow: var(--block-shadow);
    border-width: var(--block-border-width);
    border-color: white;
    border-radius: var(--block-radius);
    background: white;
    width: 100%;
    line-height: var(--line-sm);
}
div.bot.svelte-6roggh.svelte-6roggh {
    background: #D9A13D;
}
div.bot.svelte-17nzccn.svelte-17nzccn {
    background: #D9A13D;
}
div.user.svelte-6roggh.svelte-6roggh {
        background: #5F0000;
color: white;    


}
div.user.svelte-17nzccn.svelte-17nzccn {
    background: #5F0000;
}    
"""


def transcribe(audio):
    text = p(audio)["text"]
    return text


def construct_index(directory_path):
    max_input_size = 100000000
    num_outputs = 1000000000
    max_chunk_overlap = 200000000
    chunk_size_limit = 6000000000

    prompt_helper = PromptHelper(max_input_size, num_outputs, max_chunk_overlap, chunk_size_limit=chunk_size_limit)

    llm_predictor = LLMPredictor(llm=OpenAI(temperature=0.0, model_name="text-davinci-004", max_tokens=num_outputs))

    documents = SimpleDirectoryReader(directory_path).load_data()

    index = GPTSimpleVectorIndex.from_documents(documents)
    index.save_to_disk('index.json')

    return index


def chatbot(input_text):
    index = GPTSimpleVectorIndex.load_from_disk('index.json')
    response = index.query(input_text)
    return str(response.response)

def chat(input_text, MaxToken=50, outputs=3):
    MSGS = [
        {"role": "user", "content": input_text},
    ]
    # We use the Chat Completion endpoint for chat like inputs
    response = openai.ChatCompletion.create(
    # model used here is ChatGPT
    # You can use all these models for this endpoint:
    # gpt-4, gpt-4-0314, gpt-4-32k, gpt-4-32k-0314,
    # gpt-3.5-turbo, gpt-3.5-turbo-0301
    model="gpt-3.5-turbo",
    messages=MSGS,
    # max_tokens generated by the AI model
    # maximu value can be 4096 tokens for "gpt-3.5-turbo"
    max_tokens = MaxToken,
    # number of output variations to be generated by AI model
    n = outputs,
    )
    return str(response.choices[0].message.content)
    
with gr.Blocks(css=css) as demo:
    realPath = str(os.path.dirname(os.path.realpath(__file__)))
    img1 = gr.Image("images/header.png", elem_classes=".img.svelte-ms5bsk", elem_id="img.svelte-ms5bsk").style(
        container=False)
    gpt = gr.Chatbot(label=".", elem_classes=".wrap.svelte-1o68geq.svelte-1o68geq", elem_id="chatbot").style(
        container=True)
    msg = gr.Textbox(elem_id="div.svelte-awbtu4", elem_classes="textBoxBot", show_label=False,
                     placeholder="Bem vindo ao assistente virtual OM30, em que posso ajuda-lo?",
                     ).style(container=False)


    # clear = gr.Button("Limpar Conversa")
    gr.Audio(source="microphone", type="filepath",label="ESTÁ COM DIFICULDADES EM ESCREVER? CLIQUE E ME DIGA O QUE DESEJA")
    def respond(message, chat_history):
        chat_history.append((message, chat(message)))
        time.sleep(1)
        vetor = []
        realPath = str(os.path.dirname(os.path.realpath(__file__)))

        #if str(message).upper() == "OLA" or str(message).upper() == "OLÁ" or str(message).upper() == "OI":
        #    vetor = vetor + [((realPath + "\\images\\hippo-apresentacao.mp4",), "")]
        #elif str(message).upper() == "VINHO CASA DEL RONCO PINOT GRIGIO":
        #    vetor = vetor + [((realPath + "\\images\\casa-del-ronco-branco.png",), "")]
        #elif str(message).upper() == "SURVIVOR CHENIN BLANC":
        #    vetor = vetor + [((realPath + "\\images\\survivor-branco.png",), "")]
        #    vetor = vetor + [((realPath + "\\images\\survivor.mp4",), "")]

        #elif str(message).upper() == "VINHO PORTO NOVA VERDE":
        #    vetor = vetor + [((realPath + "\\images\\porta-nova-branco.jpg",), "")]
        #    vetor = vetor + [((realPath + "\\images\\porta-nova-verde.mp4",), "")]

        #elif str(message).upper() == "VINHO QUINTA DO PINTO ARINTO BRANCO":
        #    vetor = vetor + [((realPath + "\\images\\quinta-pinto-arinto-branco.png",), "")]
        #elif str(message).upper() == "VINHO 1492 CHARDONNAY":
        #    vetor = vetor + [((realPath + "\\images\\chardonay-branco.jpg",), "")]
        #elif str(message).upper() == "ME SUGIRA UM VINHO TINTO BOM COM QUEIJO":
        #    vetor = vetor + [((realPath + "\\images\\TNT-CABERNET.png",), "")]
        #    vetor = vetor + [((realPath + "\\images\\vinho-queijo.mp4",), "")]

        #elif str(message).upper() == "VINHO BOM COM CHOCOLATE":
        #    vetor = vetor + [((realPath + "\\images\\TNT-CABERNET.png",), "")]
        #elif str(message).upper() == "VINHO BOM COM PEIXE":
        #    vetor = vetor + [((realPath + "\\images\\luson-branco.png",), "")]
        #    vetor = vetor + [((realPath + "\\images\\vinho-peixe.mp4",), "")]

        #elif str(message).upper() == "VINHAS DO LASSO COLHEITA SELECIONADA":
        #    vetor = vetor + [((realPath + "\\images\\lasso-colheita-rose.png",), "")]
        #elif str(message).upper() == "DOM CAMPOS MOSCATEL":
        #    vetor = vetor + [((realPath + "\\images\\dom-campos-rose.png",), "")]
        #elif str(message).upper() == "BECAS ROSE MEIO SECO":
        #    vetor = vetor + [((realPath + "\\images\\becas-rose.png",), "")]
        #elif str(message).upper() == "PORTA DA RAVESSA":
        #    vetor = vetor + [((realPath + "\\images\\luson-branco.png",), "")]

        return "", chat_history + vetor


    # clear.click(lambda:None, None, gpt, queue=False,)
    msg.submit(respond, [msg, gpt], [msg, gpt])

index = construct_index("docs")
demo.launch()