This is a work-in-progress site to visualize the data from https://twitter.com/hpavocadoprice. Tweets are downloaded using Selenium, checking for new posts every 5 minutes. Prices are manually updated by the site maintainers via an admin interface. The site is currently deployed at https://avocado-prices.herokuapp.com. The visualization is in progress, but the data is available at /data and /data/<tweet_id>, for any <tweet_id> listed in /data.
- Python 3.6+
- virtualenv and virtualenvwrapper (highly suggested)
- Pip
- Postgres, available at the project website or from a package manager like brew.
- Appropriate Selenium web driver for your choice of browser.
- Web drivers can also be installed using brew.
- Currently, this project uses Chrome for compatibility with Heroku. If you wish to use something else, you will have to modify the TweetFetcher init statement.
- Clone the repo.
- Make a new virtual environment
$ mkvirtualenv avocado
- Run
pip install -r requirements.txt - Create the postgres database
createdb <db_name>
The server requires the following environment variables to be set:
- A hashed version of the admin password
- The port on which to run
- A Flask secret key
- A CSRF secret key
- Whether or not to run in debug mode
The fabfile can generate and set all of these.
- secret keys are always automatically generated
- the port defaults to 5000
- debug mode defaults to false
- the password can be configured or generated automatically To use defaults, run the below command from within your virtual environment. The generated admin password will be printed to the shell:
$ fab local
PLEASE RECORD YOUR ADMIN PASSWORD: Z7Q9jZLvzjTslwB
Starting server locally in debug mode at port 5000.
The port, password, and debug mode are all configurable:
$ fab local --password=admin --port=8000 --debug=False