API for predicting telephone price range
app
├── data - contains train and validation data
│ ├── train.csv - train set
│ └── val.csv - validation set (must contain target values)
├── models - this folder contains a trained estimator.
│ └── finalized_model.pickle - trained estimator.
│
├── settings - here you can store different constant values, connection parameters, etc.
│ ├── constants.py - multiple constants storage for their convenient usage.
│ └── specifications.json - specifications of your data preprocessing operations.
│
├── utils - this folder contains instruments we'll use to work with dataset.
│ ├── __init__.py - init file for the package.
│ ├── dataloader.py - dataloader.
│ ├── dataset.py - class dedicated for giving info about the dataset.
│ ├── predictor.py - predictor.
│ └── trainer.py - train script.
│
├── app.py - route, app.
│
├── requirements.txt
│
└── Dockerfile
API has only one command: /predict GET and data for prediction.
Data format: battery_power, blue, dual_sim, fc, int_memory, mobile_wt, pc, px_height, px_width, ram, sc_h, sc_w, talk_time, three_g, touch_screen, wifi. Price_range is predicted value.
For test the app run run.py. Output:
