Create a new environment using the requirements.txt file:
$ conda create --name <env> --file requirements.txt
... (some output messages)
Set the Keras trained neural network path in server.py:
NEURAL_NET_MODEL_PATH = os.environ['NEURAL_NET_MODEL_PATH']
# or
NEURAL_NET_MODEL_PATH = "/path/to/models/SouqNet128v2_gpu.h5"
Set a secret key (since Flask needs this when posting forms) in server.py:
app.config['SECRET_KEY'] = os.environ['SECRET_KEY']
# or
app.config['SECRET_KEY'] = "SOMEreallyR@ND0M5TR1NGtoK33pYouSeKYUR"
Set an upload folder path in server.py. /tmp/ folders will delete all
uploaded images:
app.config['UPLOAD_FOLDER'] = os.environ['UPLOAD_FOLDER']
# or
app.config['UPLOAD_FOLDER'] = "/tmp/souqnet/"
Use the following command prefaced by the FLASK_APP setting
$ FLASK_APP=server.py flask run
... (some output messages)
An article was originally posted on March 26, 2018 in tandem with this repository to explain how the app was built. The original article can be view on StackSchool.io. In efforts to keep the article alive on the internet, the article is reuploaded into this repository under the ARTICLE.md file.