mgr-tensorflow

Tensorflow python API

Development

export FLASK_APP=flaskapp.py
export FLASK_DEBUG=1
flask run

Deployment

1. Login to server

ssh -i cert/LeafProjectNew.pem ubuntu@ec2-35-157-132-97.eu-central-1.compute.amazonaws.com

2. Update & restart server

ON SERVER:

./update.sh

or

cd flaskapp/
git pull origin master
sudo apachectl restart

Server installing

http://www.datasciencebytes.com/bytes/2015/02/24/running-a-flask-app-on-aws-ec2/

cd flaskapp/
git pull origin master
sudo pip install -r requirements.txt
sudo apachectl restart

Test tensorflow version

python -c 'import tensorflow as tf; print(tf.__version__)'

Server logs (linux)

ERROR_LOGS

tail -f /var/log/apache2/error.log

APP_LOGS

tail -f /var/log/apache2/access.log

API Methods

Server

http://ec2-35-157-132-97.eu-central-1.compute.amazonaws.com/

[GET, POST] /api/photo-prediction

http://ec2-35-157-132-97.eu-central-1.compute.amazonaws.com/api/photo-prediction

[GET, POST] /api/photo-prediction-mock-<1/2/3>

Return results with 1x > 80% data. http://ec2-35-157-132-97.eu-central-1.compute.amazonaws.com/api/photo-prediction-mock-1

Return results with 2x > 80% data. http://ec2-35-157-132-97.eu-central-1.compute.amazonaws.com/api/photo-prediction-mock-2

Return results with 0x > 80% data. http://ec2-35-157-132-97.eu-central-1.compute.amazonaws.com/api/photo-prediction-mock-3