Python Postprocessing Engine

[afproy]

The Python post process engine will receive as input: ID, location and the status (open/closed) for each umbrella. The script itself will be implemented in a MQTT fashion and it will act as a subscriber to receive the data. It will then send notifications to the interested users based on the processed data.

Using the post process engine we can isolate the raw data that will be sent by umbrellas and the GUI on which notifications will appear (Telegram). So, the system is more flexible and new features (i.e. new interface for notifications) can be added without modifying the entire platform. This will be the actor in charge of sending rain warnings to the users.

The work will be split into the following smaller bits:

• [x] documentation, identification of the needs, definition of the scope of the microservice
• [x] definition of the API
• [x] implementation of the MQTT subscriber
• [x] implementation of a simple first version of intelligent engine that acts on the on_message callback to update the location of the users, turn the map of Turin into a grid of neighborhoods, map the users onto that grid, if more than five users have their umbrella open then send notification to the users in that area that have their umbrellas closed to remind them to use their umbrella.

The branch associated with this issue is the following: :pizza: py_pp_engine :pizza:.

[afproy]

:warning: [UPDATE] Regarding step 4: for the moment I only print that I send a notification, the actual notification sent to the bot will be the subject of a different issue. Moreover, the number of users that makes the threshold is a parameter that can be set (doesn't have to be 5).

[afproy]

:warning: [UPDATE] Implementation finished! :muscle: Will write comments and make sure the code is fully clean and I expect to push my work on the branch at the beginning of the afternoon. @louisepe once it is on the branch, can you pull it and tell me what you think of it before I merge the branch with develop please?