Skip to content

Latest commit

 

History

History
92 lines (73 loc) · 4.11 KB

README.md

File metadata and controls

92 lines (73 loc) · 4.11 KB
Raw

EnerKnol Code Exercise

Click Here

Source Code Organization

  • enerknol.py is our main entry point if we're debugging or running locally. It gives us access to a debug console as well as some contextual objects (mainly database connections) when running flask shell after setting FLASK_APP=enerknol.py. The flask-shell-ptpython ensures ptpython shell is used rather than the default python shell.

  • config.py is used for storing environment variables and gets called later in the app package.

  • app/__init__.py takes care of loading all of our flask modules in a global accessible context

  • search.py & models.py the latter stores our SQL models (postgres or sqlite depending on context) while the former handles our elasticsearches. In the future, I would write out models for the elastic indices and glue it to the mongoDB so that changes are faithfully reflected.

  • forms.py here are some very simple form objects to automate some validation as well as benefit from automatic rendering in templates.

  • routes.py

    • login very basic login view that redirects to the front page if auth'd, attempts to log in if a valid form is submitted, then proceeds to populate the login_user objects from flask_login which is just a wrapper around flask sessions and helps us manage user state and cookies.
    • flask_login module here serves as a wrapper for flask sessions. The next_page argument is checked to make sure the redirect stays on our domain.
    • register checks if user is authenticated, if not the form validation is executed and barring any conflicts a new user is created. werkzeug.security is tapped for hashing given passwords with sha256 encryption and 8 character long salts.
    • @login_required is used for gating content to authorized users only
    • search validates form then searches for movies using a search term. Under the hood, the Movies object is querying our elasticsearch index first, returning matching oid's, then returning the data as it is from the mongoDB database itself. Some pagination is included, but in the future I would add more details for navigating other pages, rather than just hitting next or previous.
    • mark is a jinja2 filter that simply wraps any instance of our search term in the results with <mark> tags causing the text to be highlighted.

Installation

Create your environment

  • virtualenv venv
  • . venv/bin/activate or source venv/bin/activate
  • pip install -r requirements.txt
  • deactivate

Flask implementation

export FLASK_APP=enerknol.py

flask shell

MongoDB

mongod --dbpath data/

ElasticSearch

brew install elasticsearch

elasticsearch

Data can be found in /usr/local/var/lib/elasticsearch/

bulk insert of json data

Requires data format of newline delimited JSON (NDJSON) with an action on the first line, and an optional source on the second.

curl -XPOST localhost:9200/movies/my_doc_type/_bulk -H "Content-Type: application/x-ndjson" --data-binary @media.movies.json
{ "index" : { "_index" : "test", "_type" : "_doc", "_id" : "1" } }
{ "field1" : "value1" }

Notes

  • elasticsearch is missing a number of records because the free heroku plan limits the database size to about 20mb.
  • Heroku Deployment
  • migrations folder is excellent for changing your SQL db schema in place should you make any changes to your sqlalchemy models

Documentations

Todo

  • Write a scraper?
  • Optimize index creation
  • Breakout authentication code
  • Use .env
  • Add tests
  • Flesh out pagination