Chicken Disease Classification

This project aims to classify diseases in chickens using a deep learning model. The repository contains all the necessary code, configurations, and instructions to train, evaluate, and deploy the model using CI/CD pipelines.

Interface

1. 
2. 

Getting Started

Prerequisites

• Python 3.8+
• Git
• DVC (Data Version Control)
• Virtual environment tool (e.g., venv, conda)

Setting Up the Environment

1. Create and Activate a Conda Environment:

   conda create -n chicken python=3.8 -y

conda activate chicken 2. Install Dependencies: ```bash pip install -r requirements.txt

Project Initialization

1. Create Necessary Files and Directories: Run template.py to create all necessary directories and files: ```bash python template.py

2.Add, Commit, and Push to GitHub: ```bash git add . git commit -m "Initial commit" git push origin main

Developing the Package

1. Editable Installation:

To use setup.py as a package in editable mode, include -e . in requirements.txt.

2. Logging Setup: Logs are stored using:

     ```bash
       logging.basicConfig(level=logging.INFO, format='[%(asctime)s]: %(message)s:')
   

3. Creating cnnClassifier Package:

Logging setup for cnnClassifier: ```bash logging_str = "[%(asctime)s: %(levelname)s: %(module)s: %(message)s]"

4. Testing the Package:

Create main.py and test the cnnClassifier package:

   ```bash
   from cnnClassifier import logger
   logger.info("Welcome to my log")

Data Ingestion

1. Download and Extract Dataset:

   a. Update config.yaml with the dataset URL. b. Implement DataIngestion class in data_ingestion.py.

2. Pipeline Stages:

   a. Implement pipeline stages in pipeline/stage_01_data_ingestion.py.

Model Training and Evaluation

1. Prepare Base Model:

   a. Store necessary parameters in params.yaml. b. Implement prepare_base_model.py.

2. Training and Evaluation:

   Follow the pipeline structure to implement training and evaluation scripts.

Deployment

1. CI/CD Pipeline with Azure:

a. Create a Dockerfile and push the image to Azure Container Registry. b. Deploy the web app using Azure Web App Service.

2. DVC Pipeline:

a. Initialize DVC and run the pipeline stages with dvc repro.

Additional Notes

a. Artifact Management: Include artifacts/* in .gitignore to avoid adding generated files to the repository. b. App Integration: Flask API is used for training, prediction, and root endpoints. c. Azure CI/CD: Detailed steps for setting up continuous deployment using Azure.

Workflow

1.Update config.yaml. 2. Update params.yaml. 3. Implement pipeline stages in the order: Data Ingestion → Prepare Base Model → Training → Evaluation. 4. Deploy using DVC and Azure.

Running the Project

1. DVC Pipeline:

    dvc init
    dvc repro

2. Flask App:

    python app.py
   

Conclusion

This project provides a comprehensive structure for chicken disease classification using deep learning, with CI/CD integration for deployment. Follow the steps above to clone, develop, and deploy the project.