#Eatorator Web Web application to support the eaterator app. Web application front end for AngularJS, as well as HTML/CSS for webapge are included. App is separated into four main sections.

* General website (`application/controller.py`)
* Auth for social and standard authorization, and issuance of JWT (`application/auth/~`)
* Recipe for recipes and ingredients (`application/recipe/~`)
* User for user data like search history and favourite recipes (`application/user/~`)

#Setup and Requirements Python version is 3.5.2 in a virtual environment. The requirements.txt have all requirements for the recipe_scraper and recipe_parser, but are not required to run the app, only to start the pipeline from text data -> recipe parser -> recipe app models. Some commands to interact with the dev server:

python manage.py init_db  # creates database/models
python manage.py migrate  # migrates database using alembic (untested at the moment)
python manage.py runserver  # starts the flask dev server
python manage.py insert_base_data  # untested function to populate base data for the application
python manage.py insert_update_recipe_data  # developing at the moment -> to process raw recipes in DB
python run.py  # starts either the gevent web server or application server depending on config

To work properly, a config_dev.py has to be made in the root folder that specifies a connection to a PostgreSQL instance. Note that not having a config_dev.py will throw an ImportError.

HOST = 'localhost'
PORT = '5432'
USER = 'user'
PASSWORD = 'password'
DATABASE = 'database'
USE_GEVENT = False  # for windows as gevent isn't compatible

In production or locally, some data from the nltk package needs to be downloaded. Activate the virtualenv and then run the following commands:

nltk.download('punkt') nltk.download('wordnet') nltk.download('maxent_treebank_tagger')

Note that this will be copied to the /home/<user>/nltk_data folder by default, and therefore, in production, this folder must be copied to /var/www/nltk_data. Error messages from nltk will also point to the locations it searches by default. When running the python manage.py ~ commands from the virtual env as non www-data user, the nltk_data folder must also exist in the /home//nltk_data folder. Maybe look into symlinking this or changing permissions.

##Windows Notes When using windows, change the database engine in the config_dev.py to:

DATABASE_ENGINE = 'sqlite'
USE_GEVENT = False

and the Flask development server will run with an SQLite database in ~/application/database/'localdb.sqlite'.

#Deployment Pipeline: Data Creation and Insertion The following commands are needed to insert data and make the API functional. Remember to set the EATERATOR_DATA_SCRAPING_PATH to point the scripts to the proper data source.

(virtual-env)$ python manage.py init_db # creates db
(virtual-env)$ python manage.py db migrate # if DB exists and there are model changes
(virtual-env)$ python manage.py insert_base_data  # insert base source url, and roles/permissions data
(virtual-env)$ python manage.py insert_recipe_data  # inserts recipe data
(virtual-env)$ python manage.py update_fulltext_fields  # populates fulltext fields for fulltext search API
(virtual-env)$ python manage.py create_indexes  # creates indexes for to expedite fulltext search API