📎 Webclip
COVID-19 Helper Whatsapp Bot - A Twilio Hackathon Submission
The project is a WhatsApp bot built for the Twilio and DEV 2020 hackathon. It uses Twilio Autopilot and Python 3.8, with FastAPI, FaunaDB, Heroku, and external APIs to answer COVID-19 questions and run a symptom-based self-screening flow.
Reading notes#
- The bot returns statistical COVID-19 data by country, such as Spain or the USA.
- It also returns historical COVID-19 data by country, such as Ireland on March 15th.
- A self-screening flow asks about common symptoms and other variables, then advises whether the user should seek medical attention.
- The author says the self-checker is only informative and does not replace a health professional’s diagnosis.
- The stack includes Python 3.8.2, Poetry, FastAPI, Pydantic, Fauna DB, Heroku, Pytest, Loguru, Requests, Sentry, and Click.
- The project was started from cookiecutter-fastapi.
- The author wanted more features, such as more specific location queries, grocery store suggestions, and daily news, but some were left out because of time or technical limits.
- The author says FastAPI was a learning experience, especially because it was the first ASGI framework used.
- FaunaDB introduced a number of surprises compared with relational databases, including primary keys, array results, missing migrations, and repeated records.
- A parsing issue with application/x-www-form-urlencoded requests initially blocked Autopilot requests from reaching the API.
- The APIs used were Novelcovid API and Endless Medical API.
- The author concludes that the project taught several new concepts in a month and became the first post published on DEV.
