📦 Forecasting Service – Django-based Demand Prediction

This microservice forecasts product demand using pre-trained ML models and integrates with a warehouse service to determine restocking needs.

---

🗂️ Project Structure

Folder/File	Purpose
forecast_system/	Main Django project config (settings, URLs, wsgi/asgi)
forecaster/	App containing all forecast logic
feature_pipelines/	Per-product feature engineering modules
models/	Pre-trained .joblib models for each product
utils/warehouse_api.py	Makes HTTP requests to warehouse management service
forecast_runner.py	Orchestrates model loading, forecasting, warehouse fetch
views.py	Exposes the forecast API endpoint
migrations/	Django migration folder
manage.py	Django project entry point

---

⚙️ Setup Instructions

1. Clone the Repository

git clone https://github.com/IASSCMS/Forecasting-Service.git
cd forecast_system

2. Create and Activate a Virtual Environment

python -m venv venv
source venv/bin/activate          # Linux/macOS
# or
. venv\Scripts\activate             # Windows in GitBash

3. Install Dependencies

pip install -r requirements.txt

---

🚀 Run the Project

python manage.py migrate
python manage.py runserver

• Or use Make commands in the root dir

make run

• for migrations

make migrate

• Now can find the root view at http://localhost:8000/api/

---

🧪 Sample Forecast API

Endpoint

POST /api/forecast/

Request Body

{
  "product_SKU": "SKU001",
  "days" : 30
}

Response

{
    "product_SKU": "SKU001",
    "current_stock": 10000,
    "average_forecasted_demand": 14608.27,
    "maximum_forecast": 16425.71,
    "minimum_forecast": 12764.46,
    "stock_shortfall": 4608.27,
    "daily_predictions": [
        {
            "date": "2025-05-03",
            "predicted": 593.5,
            "lower_bound": 530.5,
            "upper_bound": 651.77
        },
        {
            "date": "2025-05-04",
            "predicted": 635.67,
            "lower_bound": 575.94,
            "upper_bound": 697.32
        },
        ...
  ]
}

---

🔗 Dependencies

• Django – Backend web framework
• Prophet – Time series forecasting
• Pandas – Data manipulation
• Joblib – Model serialization
• Requests – HTTP requests to warehouse service

---

📡 Warehouse Service Integration

The service calls the external warehouse API:

GET http://warehouse-service/api/stock/<product_SKU>/
→ Expected response: { "current_stock": 120 }

You can change the warehouse base URL in forecaster/utils/warehouse_api.py.

---

📁 Adding a New Product

To support a new product:

1. Save its trained model as forecaster/models/<product>.joblib
2. Add a feature engineering file: forecaster/feature_pipelines/<product>.py with a function:

def prepare_features(df):
    # Add custom columns to df
    return df

---