Remote ML Model Serving Component for TrainDB
Python 3.8 on Ubuntu 20.04
# git clone https://github.com/kihyuk-nam/traindb-ml.git
# cd traindb-ml
// create a venv as 'venv'(or whatever you want)
# python3 -m venv venv
// activate it
# source venv/bin/activate
(venv) #
// install dependencies. (pip == pip3) For example,
(venv) # pip install numpy pandas tables fastapi uvicorn spflow sqlparse psycopg2 scipy
// which is the same as the following:
//(venv) # pip install -r requirements.txt
// Possible errors due to the spflow package:
// 1. PyQT5: use PyQT5=5.13 if the error occurs for 5.15.x
// 2. sklearn erorr: 'sklearn' is deprecated. It should be 'scikit-learn'. See the error message that contains the solution. (set env var)
(venv) # python3 main.py
The default host address and port (https://0.0.0.0:8000) will be applied if no args specified.
For setting up your own address/port (e.g., https://127.0.0.1:8080):
(venv) # python3 main.py --rest_host 127.0.0.1 --rest_port 8080
This project is dual-licensed. Apache 2.0 is for traindb-ml, and MIT is for the RSPN codes from deepdb(https://github.com/DataManagementLab/deepdb-public)