Upload 5 files

AlanVideos app.ipynb

@@ -0,0 +1,412 @@
+{
+  "cells": [
+    {
+      "cell_type": "code",
+      "source": [
+        "!pip install -q -r requirements.txt"
+      ],
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "_r4UUCGMTYac",
+        "outputId": "ceb295aa-f5cb-48ce-d194-5322d319461d"
+      },
+      "execution_count": 1,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "  Installing build dependencies ... \u001b[?25l\u001b[?25hdone\n",
+            "  Getting requirements to build wheel ... \u001b[?25l\u001b[?25hdone\n",
+            "  Preparing metadata (pyproject.toml) ... \u001b[?25l\u001b[?25hdone\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m57.6/57.6 kB\u001b[0m \u001b[31m1.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m32.8/32.8 MB\u001b[0m \u001b[31m1.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m85.5/85.5 MB\u001b[0m \u001b[31m1.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m817.7/817.7 kB\u001b[0m \u001b[31m1.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m312.3/312.3 kB\u001b[0m \u001b[31m1.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m8.1/8.1 MB\u001b[0m \u001b[31m1.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.8/1.8 MB\u001b[0m \u001b[31m1.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.9/1.9 MB\u001b[0m \u001b[31m902.6 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m299.3/299.3 kB\u001b[0m \u001b[31m788.9 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m116.4/116.4 kB\u001b[0m \u001b[31m959.8 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m75.6/75.6 kB\u001b[0m \u001b[31m1.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m207.3/207.3 kB\u001b[0m \u001b[31m1.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m6.9/6.9 MB\u001b[0m \u001b[31m1.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m83.0/83.0 kB\u001b[0m \u001b[31m648.6 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m49.4/49.4 kB\u001b[0m \u001b[31m601.8 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m62.7/62.7 kB\u001b[0m \u001b[31m500.1 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━��━━━\u001b[0m \u001b[32m77.9/77.9 kB\u001b[0m \u001b[31m473.3 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m58.3/58.3 kB\u001b[0m \u001b[31m556.6 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m53.0/53.0 kB\u001b[0m \u001b[31m573.6 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m141.1/141.1 kB\u001b[0m \u001b[31m668.0 kB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+            "\u001b[?25h  Building wheel for openai-whisper (pyproject.toml) ... \u001b[?25l\u001b[?25hdone\n"
+          ]
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 2,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "R1QTyYcpIJuw",
+        "outputId": "6060c0f3-621e-43ba-823d-676256cb406a"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "Writing app.py\n"
+          ]
+        }
+      ],
+      "source": [
+        "%%writefile app.py\n",
+        "\n",
+        "import os\n",
+        "import re\n",
+        "import streamlit as st\n",
+        "import transformers\n",
+        "from transformers import pipeline\n",
+        "from transformers import AutoTokenizer\n",
+        "import nltk\n",
+        "from PIL import Image\n",
+        "import torch\n",
+        "import ffmpeg\n",
+        "import speech_recognition as sr\n",
+        "from pytube import YouTube\n",
+        "import string\n",
+        "import whisper\n",
+        "from moviepy.editor import AudioFileClip\n",
+        "from langchain.text_splitter import CharacterTextSplitter\n",
+        "from langchain_core.messages import AIMessage, HumanMessage, SystemMessage\n",
+        "from langchain.embeddings import OpenAIEmbeddings\n",
+        "from langchain.vectorstores import FAISS\n",
+        "from langchain.chat_models import ChatOpenAI\n",
+        "from langchain.memory import ConversationBufferWindowMemory\n",
+        "from langchain.chains import ConversationalRetrievalChain\n",
+        "from langchain.callbacks import get_openai_callback\n",
+        "from key import openai_api_key\n",
+        "from langchain import PromptTemplate\n",
+        "from langchain import LLMChain\n",
+        "os.environ['OPENAI_API_KEY'] = openai_api_key\n",
+        "\n",
+        "\n",
+        "  # Criando a função que corta o texto em chunks:\n",
+        "def get_chunks(texto_transcrito):\n",
+        "    \"\"\"Função que recebe uma string (um texto transcrito) e devolve\n",
+        "    uma lista de strings, em que cada string (\"chunk\") contém um\n",
+        "    tamanho específico, com overlap entre os pedaços. \"\"\"\n",
+        "    text_splitter=CharacterTextSplitter(separator='.', chunk_size=1000, chunk_overlap=100, length_function=len)\n",
+        "    chunks = text_splitter.split_text(texto_transcrito)\n",
+        "    return chunks\n",
+        "\n",
+        "  # Criando a função de vectorstore para transformar os chunks de texto em embeddings:\n",
+        "def get_vectorstore(chunks):\n",
+        "    \"\"\"Função que recebe uma lista de chunks de texto e\n",
+        "    os converte em embeddings (representações numéricas)\n",
+        "    usando um modelo de embeddings pré-treinado. \"\"\"\n",
+        "    embeddings = OpenAIEmbeddings()\n",
+        "    vectorstore = FAISS.from_texts(texts=chunks, embedding=embeddings)\n",
+        "    return vectorstore\n",
+        "\n",
+        "\n",
+        "  # Criando a função para converter o vídeo para o formato adequado:\n",
+        "@st.cache_data\n",
+        "def convert_mp4_to_wav(mp4_file, wav_file):\n",
+        "    \"\"\"Função que converte um arquivo de vídeo no formato MP4 para um arquivo de\n",
+        "    áudio no formato WAV a partir das seguintes entradas: o caminho do arquivo\n",
+        "    de vídeo no formato MP4 que se deseja converter; o caminho onde o arquivo\n",
+        "    de áudio WAV resultante será salvo. Nenhuma saída explícita é retornada.\"\"\"\n",
+        "    video = AudioFileClip(mp4_file)\n",
+        "    video.write_audiofile(wav_file)\n",
+        "\n",
+        "  # Criando a função que gera a transcrição:\n",
+        "@st.cache_data\n",
+        "def get_transcriptions(url):\n",
+        "    \"\"\" Função que recebe um link de um vídeo no YouTube e\n",
+        "    devolve um dicionário contendo o título do vídeo (key)\n",
+        "    e a transcrição de seu áudio (value). \"\"\"\n",
+        "    dicionario = {}\n",
+        "      # Baixando o áudio:\n",
+        "    youtube_content = YouTube(url)\n",
+        "    audio_streams = youtube_content.streams.filter(only_audio=True)\n",
+        "    audio_streams[0].download()\n",
+        "    title = youtube_content.title\n",
+        "      # Convertendo para Wav:\n",
+        "    mp4_file = '/content/'+''.join(char for char in title if char not in string.punctuation.replace('/', '').replace('-', '').replace('!', '').replace('(', '').replace(')', ''))+'.mp4'\n",
+        "    wav_file = '/content/'+''.join(char for char in title if char not in string.punctuation.replace('/', '').replace('-', '').replace('!', '').replace('(', '').replace(')', ''))+'.wav'\n",
+        "    convert_mp4_to_wav(mp4_file, wav_file)\n",
+        "      # Inicializando o reconhecedor de fala:\n",
+        "    r = sr.Recognizer()\n",
+        "      # Carregando o áudio gravado pelo Whisper em um objeto de áudio:\n",
+        "    with sr.AudioFile(wav_file) as source:\n",
+        "        audio = r.record(source)\n",
+        "      # Transcrevendo:\n",
+        "    texto_transcrito = r.recognize_whisper(audio)\n",
+        "      # Adicionando ao dicionario:\n",
+        "    dicionario[title] = texto_transcrito\n",
+        "    return dicionario\n",
+        "\n",
+        "\n",
+        "  # Criando a função que carrega o sumarizador extrativo (para português e inglês) via HuggingFace:\n",
+        "@st.cache_resource\n",
+        "def load_extractive():\n",
+        "    \"\"\" Função sem valores de entrada que carrega o modelo\n",
+        "     de sumarização extrativa via HuggingFace. \"\"\"\n",
+        "    return pipeline(\"summarization\", model=\"NotXia/longformer-bio-ext-summ\", tokenizer=AutoTokenizer.from_pretrained(\"NotXia/longformer-bio-ext-summ\"), trust_remote_code=True)\n",
+        "\n",
+        "  # Criando a função que gera a sumarização extrativa:\n",
+        "@st.cache_data\n",
+        "def get_summarization(_model_pipeline, full_text, ratio):\n",
+        "    \"\"\" Função que recebe um texto completo a ser resumido, juntamente\n",
+        "    à taxa desejada de sumarização e a pipeline do modelo que fará o\n",
+        "    sumário, e devolve a versão resumida desse texto. \"\"\"\n",
+        "    sentences = nltk.sent_tokenize(full_text)\n",
+        "    extractive_sentences = _model_pipeline({\"sentences\": sentences}, strategy=\"ratio\", strategy_args=ratio)\n",
+        "    extractive_text = \" \".join(extractive_sentences[0])\n",
+        "    return extractive_text\n",
+        "\n",
+        "\n",
+        "  # Criando a função que gera a tradução do texto a partir do ChatGPT:\n",
+        "@st.cache_data\n",
+        "def get_translation(summarized_text):\n",
+        "    \"\"\" Função que recebe um resumo a ser traduzido,\n",
+        "    juntamente à pipeline do modelo que fará a tradução,\n",
+        "    e devolve a versão traduzida do texto. \"\"\"\n",
+        "    LLM = ChatOpenAI(openai_api_key=openai_api_key, temperature=0.25, model_name=\"gpt-3.5-turbo\")\n",
+        "    messages = [SystemMessage(content='Você é um especialista em traduções de texto do inglês para português'), HumanMessage(content=summarized_text)]\n",
+        "    translation = LLM(messages)\n",
+        "    return translation.content\n",
+        "\n",
+        "\n",
+        "  # Criando a função que corrige os erros de transcrição do texto a partir do ChatGPT:\n",
+        "@st.cache_data\n",
+        "def get_correction(transcription):\n",
+        "    \"\"\" Função que recebe a transcrição (em PT-BR) e corrige eventuais\n",
+        "    erros do whisper - por exemplo, a troca de palavras que, embora\n",
+        "    semelhantes na pronúncia, são totalmente absurdas no contexto. \"\"\"\n",
+        "    correction_prompt = PromptTemplate.from_template(template = 'Você é um especialista em correção de textos. Você receberá uma transcrição de um áudio e deverá substituir as palavras transcritas erroneamente, considerando apenas a pronúncia, por palavras adequadas que façam sentido no contexto do texto. Além disso, corrija questões gramáticais para facilitar a compreensão. Corrija a seguinte transcrição: {transcription}')\n",
+        "    LLM = ChatOpenAI(openai_api_key=openai_api_key, temperature=0.1, model_name=\"gpt-3.5-turbo\")\n",
+        "    corrector = LLMChain(prompt=correction_prompt, llm=LLM)\n",
+        "    with get_openai_callback() as cb:\n",
+        "        corrected_transcription = corrector.run({\"transcription\": transcription})\n",
+        "        total_t = cb.total_tokens\n",
+        "        prompt_t = cb.prompt_tokens\n",
+        "        completion_t = cb.completion_tokens\n",
+        "        total_cost = cb.total_cost\n",
+        "    return corrected_transcription\n",
+        "\n",
+        "\n",
+        "  # Criando o chatbot\n",
+        "def alan_videos(vectorstore):\n",
+        "    \"\"\" Função que inicializa e configura um LLM da OpenAI\n",
+        "    e retorna um chatbot configurado pronto para uso. \"\"\"\n",
+        "    memory = ConversationBufferWindowMemory(memory_key='chat_history', return_messages=True, k=3)\n",
+        "    LLM = ChatOpenAI(openai_api_key=openai_api_key, temperature=0.25, model_name=\"gpt-3.5-turbo\")\n",
+        "    retriever=vectorstore.as_retriever()\n",
+        "    chatbot = ConversationalRetrievalChain.from_llm(llm=LLM, retriever=retriever, memory=memory)\n",
+        "    return chatbot\n",
+        "\n",
+        "  # Criando um modelo de chat:\n",
+        "def chat(pergunta):\n",
+        "    \"\"\" Função que processa uma pergunta utilizando o chatbot\n",
+        "    configurado (alan_videos) e retorna sua resposta. \"\"\"\n",
+        "    with get_openai_callback() as cb:\n",
+        "        resposta = st.session_state.alanvideos.invoke({\"question\": pergunta})\n",
+        "        total_t = cb.total_tokens\n",
+        "        prompt_t = cb.prompt_tokens\n",
+        "        completion_t = cb.completion_tokens\n",
+        "        total_cost = cb.total_cost\n",
+        "    return resposta['answer']\n",
+        "\n",
+        "\n",
+        "  # Criando a função para customização com CSS:\n",
+        "def local_css(file_name):\n",
+        "    \"\"\" Função que carrega um arquivo CSS local e aplica o estilo\n",
+        "    ao Streamlit app, personalizando a interface do usuário. \"\"\"\n",
+        "    with open(file_name, \"r\") as f:\n",
+        "        st.markdown(f\"<style>{f.read()}</style>\", unsafe_allow_html=True)\n",
+        "\n",
+        "  # Configurando a aba do site:\n",
+        "icon = Image.open(\"Traçado laranja #f1863d.png\")\n",
+        "st.set_page_config(page_title=\"AlanVideos\", page_icon=icon, layout=\"wide\", initial_sidebar_state=\"auto\")\n",
+        "\n",
+        "\n",
+        "  # Configurando o site:\n",
+        "def main():\n",
+        "    local_css(\"style.css\")\n",
+        "\n",
+        "    header = st.container()\n",
+        "    model = st.container()\n",
+        "    model_1 = st.container()\n",
+        "    model_2 = st.container()\n",
+        "\n",
+        "      # Configurando a barra lateral:\n",
+        "    with st.sidebar:\n",
+        "        with st.form(\"data_collection\"):\n",
+        "            link = st.text_area(label=\"Coloque o link do seu vídeo do YouTube:\", height=25, placeholder=\"Digite seu link...\")\n",
+        "            language = st.selectbox('Qual a linguagem do seu vídeo?', ('Português (pt)', 'Inglês (en)'))\n",
+        "            translate = st.selectbox('Você deseja que o seu vídeo seja traduzido para Português?', ('Não', 'Sim'))\n",
+        "            compression_rate = st.slider(label=\"Selecione o percentual de sumarização:\", min_value=0.05, max_value=0.50, value=0.25, step=0.05)\n",
+        "            submitted = st.form_submit_button(\"Enviar\")\n",
+        "\n",
+        "            # Verificando se o link do vídeo é válido:\n",
+        "            if link != '':\n",
+        "                if re.match((r'(https?://)?(www\\.)?''(youtube|youtu|youtube-nocookie)\\.(com|be)/''(watch\\?v=|embed/|v/|.+\\?v=)?([^&=%\\?]{11})'), link):\n",
+        "                    st.success('Dados coletados!', icon=\"✅\")\n",
+        "                else:\n",
+        "                    st.error('Link inválido. Por favor, insira um link do YouTube.', icon=\"🚨\")\n",
+        "\n",
+        "    if \"alanvideos\" not in st.session_state:\n",
+        "        st.session_state.alanvideos = None\n",
+        "\n",
+        "      # Configurando o cabeçalho:\n",
+        "    with header:\n",
+        "        st.title(\":orange[Alan]Chat\")\n",
+        "        st.subheader(\"Olá, usuário! Este é um projeto que utiliza técnicas de inteligência artificial para simplificar e acelerar a compreensão de conteúdo audiovisual. Por favor, preencha o formulário ao lado para que possamos responder as suas dúvidas a respeito de um vídeo do YouTube!  :)\", divider = \"orange\")\n",
+        "\n",
+        "      # Configurando os modelos:\n",
+        "    with model:\n",
+        "        if submitted and re.match((r'(https?://)?(www\\.)?''(youtube|youtu|youtube-nocookie)\\.(com|be)/''(watch\\?v=|embed/|v/|.+\\?v=)?([^&=%\\?]{11})'), link):\n",
+        "            with st.spinner(\"Carregando modelos...\"):\n",
+        "                nltk.download(\"punkt\")\n",
+        "                extractive = load_extractive()\n",
+        "            with st.spinner(\"Transcrevendo texto...\"):\n",
+        "                transcription = get_transcriptions(link)\n",
+        "                texto_cru = ''.join(list(transcription.values()))\n",
+        "              # Preparando o modelo de sumarização após o envio de um link:\n",
+        "            with model_1:\n",
+        "                st.header(\"Texto Sumarizado:\")\n",
+        "                with st.spinner(\"Carregando sumarização...\"):\n",
+        "                    summary = get_summarization(extractive, texto_cru, compression_rate)\n",
+        "                      # Usando o GPT para corrigir a transcrição, caso o vídeo seja em PT-BR:\n",
+        "                    if language == 'Português (pt)':\n",
+        "                        summary = get_correction(summary)\n",
+        "                      # Usando o GPT para traduzir a transcrição para PT-BR, caso o usuário prefira:\n",
+        "                    elif language == 'Inglês (en)' and translate == 'Sim':\n",
+        "                        with st.spinner(\"Traduzindo sumarização...\"):\n",
+        "                            summary = get_translation(summary)\n",
+        "                    st.session_state['summary'] = summary\n",
+        "              # Preparando o AlanVideos após o envio de um link:\n",
+        "            with model_2:\n",
+        "                st.header(\"Resposta das perguntas:\")\n",
+        "                with st.spinner(\"Preparando AlanVideoss...\"):\n",
+        "                      # Separando o texto em chunks:\n",
+        "                    chunks = get_chunks(texto_cru)\n",
+        "                      # Armazenando os pedaços de texto em embeddings:\n",
+        "                    vectorstore = get_vectorstore(chunks)\n",
+        "                      # Criando o chatbot:\n",
+        "                    st.session_state.alanvideos = alan_videos(vectorstore)\n",
+        "\n",
+        "      # Apresentando os resultados:\n",
+        "    with model:\n",
+        "          # Resultado do modelo de sumarização:\n",
+        "        with model_1:\n",
+        "            if 'summary' in st.session_state:\n",
+        "                st.write(st.session_state['summary'])\n",
+        "      # Resultado do AlanVideos:\n",
+        "    if 'summary' in st.session_state:\n",
+        "        if \"chat_history\" not in st.session_state:\n",
+        "            st.session_state.chat_history = [AIMessage(content=f\"Olá, meu nome é Alan, e estou aqui para responder perguntas sobre o vídeo. Como posso ajudar?\")]\n",
+        "        pergunta = st.chat_input('Faça uma pergunta sobre o vídeo: ')\n",
+        "        if pergunta is not None and pergunta != \"\":\n",
+        "            resposta = chat(pergunta)\n",
+        "            st.session_state.chat_history.append(HumanMessage(content=pergunta))\n",
+        "            st.session_state.chat_history.append(AIMessage(content=resposta))\n",
+        "        for message in st.session_state.chat_history:\n",
+        "            if isinstance(message, AIMessage):\n",
+        "                with st.chat_message(\"AI\"):\n",
+        "                    st.write(message.content)\n",
+        "            elif isinstance(message, HumanMessage):\n",
+        "                with st.chat_message(\"Human\"):\n",
+        "                    st.write(message.content)\n",
+        "\n",
+        "main()"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 3,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "tPfm2EHyimcF",
+        "outputId": "80127d2b-8c4a-4475-bac9-a682fe37ff24"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "34.136.19.61\n"
+          ]
+        }
+      ],
+      "source": [
+        "!wget -q -O - ipv4.icanhazip.com"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 4,
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "dmF1j2csipF8",
+        "outputId": "2bc0bbbb-b2f6-4e4b-8e32-5423790888ff"
+      },
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "\n",
+            "Collecting usage statistics. To deactivate, set browser.gatherUsageStats to false.\n",
+            "\u001b[0m\n",
+            "\u001b[0m\n",
+            "\u001b[34m\u001b[1m  You can now view your Streamlit app in your browser.\u001b[0m\n",
+            "\u001b[0m\n",
+            "\u001b[34m  Network URL: \u001b[0m\u001b[1mhttp://172.28.0.12:8501\u001b[0m\n",
+            "\u001b[34m  External URL: \u001b[0m\u001b[1mhttp://34.136.19.61:8501\u001b[0m\n",
+            "\u001b[0m\n",
+            "\u001b[K\u001b[?25hnpx: installed 22 in 3.639s\n",
+            "your url is: https://tangy-heads-feel.loca.lt\n",
+            "\u001b[34m  Stopping...\u001b[0m\n",
+            "^C\n"
+          ]
+        }
+      ],
+      "source": [
+        "!streamlit run app.py & npx localtunnel --port 8501"
+      ]
+    }
+  ],
+  "metadata": {
+    "accelerator": "GPU",
+    "colab": {
+      "collapsed_sections": [
+        "ETBAYQ66Heo-",
+        "h2eI6F54GrDY"
+      ],
+      "gpuType": "T4",
+      "provenance": []
+    },
+    "kernelspec": {
+      "display_name": "Python 3",
+      "name": "python3"
+    },
+    "language_info": {
+      "name": "python"
+    }
+  },
+  "nbformat": 4,
+  "nbformat_minor": 0
+}

htmlTemplates.py

@@ -0,0 +1,44 @@
+css = '''
+<style>
+.chat-message {
+    padding: 1.5rem; border-radius: 0.5rem; margin-bottom: 1rem; display: flex
+}
+.chat-message.user {
+    background-color: #2b313e
+}
+.chat-message.bot {
+    background-color: #475063
+}
+.chat-message .avatar {
+  width: 20%;
+}
+.chat-message .avatar img {
+  max-width: 78px;
+  max-height: 78px;
+  border-radius: 50%;
+  object-fit: cover;
+}
+.chat-message .message {
+  width: 80%;
+  padding: 0 1.5rem;
+  color: #fff;
+}
+'''
+
+bot_template = '''
+<div class="chat-message bot">
+    <div class="avatar">
+        <img src="https://w7.pngwing.com/pngs/971/3/png-transparent-alan-turing-the-enigma-bletchley-park-the-imitation-game-alan-turing-law-actor-thumbnail.png" style="max-height: 78px; max-width: 78px; border-radius: 50%; object-fit: cover;">
+    </div>
+    <div class="message">{{MSG}}</div>
+</div>
+'''
+
+user_template = '''
+<div class="chat-message user">
+    <div class="avatar">
+        <img src="https://e7.pngegg.com/pngimages/901/543/png-clipart-computer-icons-person-random-icons-miscellaneous-image-file-formats.png">
+    </div>    
+    <div class="message">{{MSG}}</div>
+</div>
+'''

requirements.txt

@@ -0,0 +1,22 @@
+# Bibliotecas de Processamento de Vídeo
+ffmpeg-python
+pytube
+
+# Reconhecimento e Processamento de Fala
+git+https://github.com/openai/whisper.git
+SpeechRecognition
+
+# Frameworks de Machine Learning e Deep Learning
+torch
+transformers
+
+# Indexação de Dados e Busca por Similaridade
+faiss-gpu
+
+# Bibliotecas PLN e APIs de IA
+nltk
+langchain
+openai
+
+# Framework de Aplicação Web
+streamlit

style.css

@@ -0,0 +1,17 @@
+/* Customize the colors for interactive Streamlit elements */
+.stButton>button {
+    background-color: #f1863d; /* Orange background for buttons */
+    color: #ededf5;
+}
+
+.stTextInput>div>div>input {
+    background-color: #ededf5;
+    color: #383838; /* Dark text for contrast */
+}
+
+h1.changa-title {
+    font-family: "Changa", sans-serif;
+    font-weight: 700; /* Bold weight */
+    font-style: normal;
+    color: #ededf5 !important; /* Ensure the color applies overriding default */
+}