Module 4 | Team: Autumn Martin, Daniel Mulitauopele, Dina Caraballo, Nick Dambrosio
Cross-Pollination Capstone: a project where backend and frontend students collaborate and build an app together. This project encompasses separate backend and frontend apps that communicate to each other via API requests.
Would you like to find out the most recent attitudes towards a particular cryptocurrency? Teller AI retrieves relevant tweets, and runs Watson sentiment analysis to discover sentiment towards a coin.
Teller AI is deployed here.
Teller AI is a Python Django API. It consumes tweet data from Twitter and accesses Watson Sentiment Analysis to analyze tweets.
Teller AI is a microservice for our Teller project, which also includes a Ruby on Rails backend, Teller Api, and a React frontend, Teller.
Python 3.7.1, Django 2.1.4, Heroku, Django-Nose testing
Example Request:
GET /teller/watson_analysis?coin=dogecoin
Example Response:
{ "document_tones": ["joy", "tentative"] }
First, clone down this project by running git clone [email protected]:DanielMulitauopele/teller-ai.git in the CLI.
If you are new to Python and Django, this may be helpful for setting up.
This app uses Python 3.7.1O and Django 2.1.4. Other requirements for this app are located here.
For development, you may either start the Django development server or start the Gunicorn server server.
To start the Django server, run python manage.py runserver 8080. The default port is 8000, but this last command-line argument will change it to port 8080.
To start the Gunicorn server locally (that's right--it's a Green Unicorn!) run gunicorn tellerai.wsgi
This app is deployed to Heroku. We use beta-teller-ai for staging prior to production, and teller-ai for production.
Changes can only be pushed to beta-teller-ai by accepted collaborators from the master branch. From there, this build may be manually promoted to production on teller-ai.
To see if tests pass and what test coverage is, run python manage.py test.
This command already includes command line arguments, --with-coverage --cover-package=teller --verbosity=1, due to our settings. This runs coverage for files within the teller directory. The output will be similar to this one:
If you would like more detailed test coverage information, run coverage run manage.py. To view the results in terminal run coverage report.
For a more elegant presentation, run coverage html. This will create or update a folder called htmlcov. View the file path htmlcov/index.html, and open this file in your browser. For example, right click on index.html within htmlcov. Choose open with and click on your browser of choice (note: you may need to view it in finder first).