Upload folder using huggingface_hub

.github/workflows/update_space.yml

@@ -0,0 +1,28 @@
+name: Run Python script
+
+on:
+  push:
+    branches:
+      - main
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+
+    steps:
+    - name: Checkout
+      uses: actions/checkout@v2
+
+    - name: Set up Python
+      uses: actions/setup-python@v2
+      with:
+        python-version: '3.9'
+
+    - name: Install Gradio
+      run: python -m pip install gradio
+
+    - name: Log in to Hugging Face
+      run: python -c 'import huggingface_hub; huggingface_hub.login(token="${{ secrets.hf_token }}")'
+
+    - name: Deploy to Spaces
+      run: gradio deploy

.gitignore

@@ -0,0 +1,64 @@
+# Python
+__pycache__/
+*.py[cod]
+*$py.class
+*.so
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+
+# Virtual Environment
+venv/
+ENV/
+env/
+.env
+
+# IDE
+.idea/
+.vscode/
+*.swp
+*.swo
+
+# Jupyter Notebook
+.ipynb_checkpoints
+*.ipynb
+
+# Model files and cache
+*.h5
+*.keras
+*.joblib
+*.pkl
+.cache/
+.pytest_cache/
+.coverage
+htmlcov/
+.transformers/
+cached_*
+
+# Gradio
+flagged/
+gradio_cached_examples/
+.gradio/
+*.pem
+
+# Logs
+*.log
+logs/
+lightning_logs/
+
+# OS
+.DS_Store
+Thumbs.db

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 Luciano Ayres
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -1,12 +1,95 @@
 ---
-title: Sentiment Analysis
-emoji: ⚡
-colorFrom: pink
-colorTo: purple
+title: sentiment-analysis
+app_file: app.py
 sdk: gradio
 sdk_version: 5.25.0
-app_file: app.py
-pinned: false
 ---
+# 🎯 Análise de Sentimento em Avaliações de Produtos
+
+Este sistema analisa o sentimento em avaliações de produtos em português usando o modelo BERT com fine-tuning em dados do e-commerce brasileiro.
+
+## 🤖 Modelo
+Utiliza o modelo [BERT fine-tuned para análise de sentimentos](https://huggingface.co/layers2024/bert-sentiment), treinado com o dataset [Olist Store](https://www.kaggle.com/datasets/olistbr/brazilian-ecommerce/data), um conjunto público de mais de 100 mil avaliações de e-commerce brasileiro feitas entre 2016 e 2018.
+
+## 🎯 Projeto
+Desenvolvido como parte do projeto NLP-Sentinel por [Luciano Ayres](https://linkedin.com/in/lucianoayres).
+
+## 💻 Instalação Local
+
+### Pré-requisitos
+- Python 3.10+
+- Git (opcional)
+
+### Instalação
+
+1. Clone o repositório:
+```bash
+git clone [email protected]:lucianoayres/sentiment-analysis-app.git
+cd sentiment-analysis-app
+```
+
+2. Execute o script de instalação e inicialização:
+```bash
+./run.sh
+```
+
+O script irá:
+- Criar um ambiente virtual Python
+- Instalar as dependências necessárias
+- Iniciar a aplicação
+
+## 🌐 Demo Online
+
+Você pode acessar uma versão online da aplicação em:
+[https://huggingface.co/spaces/layers2024/analise-de-sentimentos-avaliacao-de-produtos](https://huggingface.co/spaces/layers2024/analise-de-sentimentos-avaliacao-de-produtos)
+
+Gradio will:
+- Start a local server (usually accessible at [http://localhost:7860](http://localhost:7860))
+- Print a shareable public URL (if `share=True` is set) so that you can try your app in your browser.
+
+## Deploying Your Gradio App to Hugging Face Spaces
+
+Hugging Face Spaces provides a free and permanent hosting option for your Gradio demo. Follow the steps below to deploy your app using the terminal method:
+
+### 1. Ensure You Have a Hugging Face Account
+Make sure you have a free Hugging Face account. If not, [create one here](https://huggingface.co/join).
+
+### 2. Deploy via Terminal
+
+From your app's directory (where your `app.py` and `requirements.txt` reside), simply run:
+
+```bash
+gradio deploy
+```
+
+This command will gather basic metadata from your project, automatically create a new Space for you, and deploy your Gradio app.  
+- **To Update Your Space:** Re-run the `gradio deploy` command, or you can enable GitHub Actions to automatically update your Space on git push.
+
+### 3. Access and Share Your App
+
+Once deployed, your app will be live at a URL in the following format:
+
+```
+https://<your-username>-<your-space-name>.hf.space
+```
+
+Share this URL with others to allow them to interact with your Gradio demo directly from their browsers.
+
+## Additional Information
+
+- **Model Updates:** If you update your model on Hugging Face, the next time your app runs (locally or on Spaces), it will load the latest version.
+- **Hot Reload (Local Development):** Instead of running `python app.py`, you can run:
+  ```bash
+  gradio app.py
+  ```
+  This enables hot reloading so your changes are automatically reflected in your demo.
+- **Troubleshooting:**
+  - Ensure your virtual environment is activated before installing dependencies and running your script.
+  - Verify that the package versions in your `requirements.txt` file are compatible with your code.
+  - The initial launch might take extra time as your model files download from Hugging Face.
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+For further details, please refer to the [Gradio Documentation](https://gradio.app/docs/) and the [Hugging Face Transformers Documentation](https://huggingface.co/docs/transformers).
+=======
+# sentiment-analysis-app
+App de anális de sentimento em avaliações de produtos em português usando BERT com fine-tuning em dados do e-commerce brasileiro.
+>>>>>>> 42cb5fa7402ec14e53cdffc7568dcf02fc9750fe

app.py

@@ -0,0 +1,141 @@
+import gradio as gr
+from transformers import AutoTokenizer, TFAutoModelForSequenceClassification
+import tensorflow as tf
+
+# Load BERT model
+def load_models():
+    models = {}
+    try:
+        models['BERT'] = {
+            'model': TFAutoModelForSequenceClassification.from_pretrained("layers2024/bert-sentiment"),
+            'tokenizer': AutoTokenizer.from_pretrained("neuralmind/bert-base-portuguese-cased")
+        }
+        print("✓ BERT loaded")
+    except Exception as e:
+        print(f"⚠️ Error loading BERT model: {str(e)}")
+    
+    return models
+
+# Exemplos de avaliações para teste
+EXAMPLES = [
+    # Avaliações Positivas (nota 5)
+    "Produto excelente! Entrega rápida e qualidade ótima. Recomendo!",
+    "Ótimo atendimento, produto bem embalado e conforme descrito.",
+    "Entrega antes do prazo. Produto de qualidade e preço justo.",
+    "Muito satisfeito! Produto original e entrega rápida.",
+    "Produto perfeito, bem embalado. Vendedor confiável.",
+    
+    # Avaliações Negativas (nota 1)
+    "Produto de baixa qualidade. Não recomendo.",
+    "Veio com defeito e sem suporte do vendedor.",
+    "Atrasou muito e veio diferente do anunciado.",
+    "Faltaram itens na entrega. Sem retorno do vendedor.",
+    "Produto muito ruim. Não funciona!"
+]
+
+def get_prediction(text, model_dict):
+    inputs = model_dict['tokenizer'](text, return_tensors="tf", truncation=True, padding=True)
+    outputs = model_dict['model'](**inputs)
+    probabilities = tf.nn.softmax(outputs.logits, axis=-1)
+    probs = probabilities.numpy()[0]
+    predicted_class = tf.math.argmax(probabilities, axis=-1).numpy()[0]
+    confidence = probs[predicted_class]
+    return predicted_class, confidence
+
+def predict_sentiment(review_text):
+    # Definir threshold de confiança
+    HIGH_CONFIDENCE = 0.90  # 90% de confiança mínima para classificação definitiva
+    
+    try:
+        predicted_class, confidence = get_prediction(review_text, MODELS['BERT'])
+        
+        # Determinar o sentimento baseado no threshold
+        if confidence >= HIGH_CONFIDENCE:
+            sentiment = "POSITIVO" if predicted_class == 1 else "NEGATIVO"
+            status = "✅ Alta confiança"
+        else:
+            if predicted_class == 1:
+                sentiment = "POSSIVELMENTE POSITIVO"
+            else:
+                sentiment = "POSSIVELMENTE NEGATIVO"
+            status = "⚠️ Baixa confiança"
+        
+        return f"BERT:\n{sentiment}\n{status}\nConfiança: {confidence:.1%}"
+            
+    except Exception as e:
+        return f"⚠️ Erro: {str(e)}"
+
+# Load all models
+print("Carregando modelos...")
+MODELS = load_models()
+print("Modelos carregados!")
+
+# Create the Gradio interface
+with gr.Blocks() as interface:
+    gr.Markdown("# 🎯 Análise de Sentimento em Avaliações de Produtos")
+    
+    with gr.Row():
+        with gr.Column(scale=1):
+            input_text = gr.Textbox(
+                label="Texto da Avaliação",
+                placeholder="Digite aqui a avaliação do produto...",
+                lines=4
+            )
+            with gr.Row():
+                clear_btn = gr.Button("Limpar", variant="secondary")
+                analyze_btn = gr.Button("Analisar Sentimento", variant="primary")
+        
+        with gr.Column(scale=1):
+            output_text = gr.Textbox(
+                label="Resultados da Análise",
+                lines=4
+            )
+    
+    gr.Examples(
+        examples=[[ex] for ex in EXAMPLES],
+        inputs=[input_text],
+        label="Exemplos de Avaliações"
+    )
+    
+    with gr.Accordion("ℹ️ Sobre o Projeto", open=False):
+        gr.Markdown('''
+        ### Análise de Sentimento com BERT
+        Este sistema analisa o sentimento em avaliações de produtos em português usando BERT com fine-tuning em dados do e-commerce brasileiro.
+        
+        #### 🤖 Detalhes Técnicos
+        - **Modelo**: [BERT fine-tuned para análise de sentimentos](https://huggingface.co/layers2024/bert-sentiment)
+        - **Dataset**: [Olist Store](https://www.kaggle.com/datasets/olistbr/brazilian-ecommerce/data) (100k+ avaliações)
+        - **Projeto**: NLP-Sentinel por [Luciano Ayres](https://linkedin.com/in/lucianoayres)
+        ''')
+    
+    def clear_inputs():
+        return "", ""
+    
+    analyze_btn.click(
+        fn=predict_sentiment,
+        inputs=input_text,
+        outputs=output_text
+    )
+    
+    clear_btn.click(
+        fn=clear_inputs,
+        outputs=[input_text, output_text]
+    )
+
+interface.theme = gr.themes.Default()
+css = """
+    .gradio-container {max-width: 800px; margin: auto;}
+    .footer {display: none;}
+    a {color: #2196F3; text-decoration: none;}
+    a:hover {text-decoration: underline;}
+    .examples {margin-top: 20px; text-align: left;}
+    .examples-table {width: 100%; border-collapse: collapse;}
+    .example-row {cursor: pointer; transition: background 0.2s;}
+    .example-row:hover {background: #f0f0f0;}
+    .textbox {text-align: left;}
+"""
+interface.css = css
+
+# Launch the Gradio app
+if __name__ == "__main__":
+    interface.launch(share=True)

requirements.txt

@@ -0,0 +1,3 @@
+gradio==5.25.0
+tensorflow
+transformers

run.sh

@@ -0,0 +1,24 @@
+#!/bin/bash
+
+# Kill any running Python processes
+pkill -9 python3
+
+# Wait a moment to ensure ports are freed
+sleep 1
+
+# Create virtual environment if it doesn't exist
+if [ ! -d "venv" ]; then
+    echo "Creating virtual environment..."
+    python3 -m venv venv
+fi
+
+# Activate virtual environment
+source venv/bin/activate
+
+# Install/update dependencies
+echo "Installing/updating dependencies..."
+pip install -r requirements.txt
+
+# Run the app
+echo "Starting the application..."
+python3 app.py