📘 Template for Text Classification NLP (Flask + TensorFlow)

This repository provides a Flask web application for sentiment analysis using an NLP text classification model. It includes setup instructions for both Windows and macOS.

---

🧱 Prerequisites

✅ Python Installed (Version 3.12 or Below)

TensorFlow may not support very new Python versions.

Recommended Python versions:

• 3.10
• 3.11
• 3.12

Download Python:
https://www.python.org/downloads/

---

📥 Installation Guide (Windows & macOS)

Follow these steps after cloning the repository from GitHub.

---

1️⃣ Clone the Repository

git clone <your-repo-url>
cd text-classification-nlp-training-flask-framework

2️⃣ Open Project in VS Code

code .

3️⃣ Create Virtual Environment

Windows:

python -m venv venv

macOS:

python3 -m venv venv

4️⃣ Activate the Virtual Environment

Windows:

venv\Scripts\activate

macOS:

source venv/bin/activate

You should now see the environment name:

(venv)

5️⃣ Install Dependencies

This project uses requirements.txt.

Windows:

pip install -r requirements.txt

macOS:

pip3 install -r requirements.txt

Note: Some systems use python / python3 or pip / pip3. Use what works on your machine.

6️⃣ Run the Flask Application

Windows:

python app.py

macOS:

python3 app.py

Once running, Flask will show something like:

 * Running on http://127.0.0.1:5000

7️⃣ Open the Application in a Browser

Open:

http://127.0.0.1:5000

You will now see the sentiment UI.

---

🧪 Testing the Sentiment Analysis Model

Try entering text such as:

• I love this product!
• This is terrible.
• It's okay, nothing special.

The model should return categories like:

• Positive
• Neutral
• Negative

If predictions appear correctly, your setup is complete.

---

🔄 Using Your Own Model (COMPULSORY)

If you have trained your own sentiment analysis model, you can easily replace the existing model files.

Steps to Replace the Model:

1. Place your model files in the root folder of the project

2. Replace these files:

   • sentiment_rnn_model.h5 - Your trained TensorFlow/Keras model
   • tokenizer.pkl - Your fitted tokenizer (pickled)

3. Ensure your model:

   • Uses the same input format (tokenized sequences)
   • Outputs predictions in the same format
   • Is compatible with TensorFlow/Keras

4. Restart the Flask application to load the new model

Note: Make sure your custom model and tokenizer were trained with compatible parameters (sequence length, vocabulary size, etc.)

---

📝 Troubleshooting (Optional only if facing any error)

Virtual Environment Not Activating (Windows)

If permission error appears:

Set-ExecutionPolicy Unrestricted

Run PowerShell as Administrator.

TensorFlow Errors

Make sure Python ≤ 3.12.

If issues persist:

pip install --upgrade pip
pip install tensorflow

Flask Not Starting

Check missing packages:

pip list

Ensure Flask, TensorFlow, NumPy are installed.

---

🎉 You're Ready for Presentation

Once:

• The app runs
• No installation errors
• Sentiment prediction works

You are ready to present the NLP project.