deven96 / imageq_api

Prediction API powering the imageQ search engine

ImageQ API

Flask API for predicting classes made originally for ImageQ

• ImageQ API
  • Getting Started
    • Installation
  • Running Locally
  • Deploy
  • Documentation
  • Todo

Getting Started

Clone the repo

    # SSH
    git clone [email protected]:bisoncorps/imageQ_API.git
    # HTTPS
    git clone https://github.com/bisoncorps/imageQ_API.git

Activate virtual environment. All project work should be done in virtualenvs and virtualenv names must be added to gitignore

Installation

• Install the requirements

    # install pipenv
    sudo pip3 install pipenv

    # install requirements
    pipenv install

Running Locally

• With flask dev server

    python flask_api/run_keras_server.py

• With Gunicorn (port 8008)

    gunicorn -b :8008 flask_api:app

• With deployed Docker image from docker hub

    docker run deven96/imageqapi

Upon running image, docker container port is bound to localhost

Deploy

The /predict endpoint on master branch of the repo is linked to automatically deploy to be hosted on Google Cloud and as a docker container deven96/imageqapi

Documentation

Documentation including example use are available on hosted version

Todo

See TODO