Facility Bot

Org and Bot Git:

https://github.com/HumanbiOS/

Python async framework for telegram bots: https://github.com/aiogram/aiogram

Join our developer group for telegram: https://t.me/joinchat/KTRuxhl949aHv0atuWPtyw
Write us on Facebook: https://www.facebook.com/humanbios.org/

Prototype

Features to be specified:

• Translation API's

worker = medical/social personnel

Goal is to connect people to doctors, psychologists, social workers via a bot. The bot welcomes the person and asks the user questions:

MVP Bot Questions:

- Are you feeling okay? --> Yes --> Thats is great! Do you wanna help --> Yes --> Ask for a state description --> new members room
- Are you feeling okay? --> Yes --> Thats is great! Do you wanna help --> No -> Okay, please tell your friends about humanbios
- Are you feeling okay? --> No --> Oh no, I'm sorry about that! Do you have a cough or fever? --> Yes --> Ask for a state description --> Doctors room
- Are you feeling okay? --> No --> Oh no, I'm sorry about that! Do you have a cough or fever? --> No --> Are you feeling stressed or anxious?  --> Yes --> Ask for a state description --> Psychologist room

Extensibility of the diagnosis graph

The Bot should be implemented, s.t. the diagnosis graph generated by the user interaction can be easiliy extended as more information becomes available.
-> Extensability is Nr. 1 priority here

###  provide WHO recommandations to worker's that pick up medical cases for the first time. Repeat after 10 cases. (https://apps.who.int/iris/rest/bitstreams/1272288/retrieve)
###  provide emergeny heuristics  to worker's that pick up medical cases for the first time. Repeat after 10 cases. (https://hackmd.io/IFreJzwgRi-ZjGHCTUtBaQ)
###  provide "how to calm someone down text" to worker's that pick up social/psychological cases  (https://hackmd.io/52ptNl2QTQSn-MI_AWORog?both)
###  Rate limiting messages when cases wait longer then 15 Minutes
### Hard cap if more than 100 cases per worker that is online:

This is a limitation for `number_of_cases/number_of_available_workers.` 

If availability is not available by the API, use the following heuristic: 
- add "sign out" button for user <-> worker conversation
- `last_seen: last message send by worker `
- after x (default_10) minutes: send message to worker: Hello, are you still working on cases? If not please "sign out"
- if no response after y (default=5) minutes, assume worker offline. 

State descriptions template:

New members room:
Welcome new member. We ware so glad you're here! Please provide a short description of what you would like to help with and what you can do. Keep it brief and professionell.

Doctors room:
Dear patient, we will try to help you as much as we can. What is your body temperature?

Social room:
Hello! Please tell us a little about your current situation. How are you feeling? Are you afraid? Take a minute to relax and breath. Tell us also about your friends and family:

Bonus: Find emoticons/images to make the questions more emphatic [consult with psychologists, Bonus: Use local conventions]

2. Digital Hospitals:

Overview

Bot can be deployed as a docker container, care taking personell is selforganised in groups (medical/social/psychological)
Users provide information by Q&A bot, describe their current state in the last step.

Care workers frontend:

Care taking frontend:
-> accept/I am not able to handle this case/report abuse
-> establishes chat link: User <-> bot <-> Care taker conversation to user
-> Bot provides option to open voice channel
-> Bot provides option to open video channel

Backend:
-> alert if requests start queing > x minutes (config variable)]
-> Persistency: Store decision trees and state descriptions (low hanging fruit: python pickle )

Server measures incoming getUpdate load:
If load gets to high, dynamically resizes hosting instance (or reassigns to different instance)

UX Questions:

When should we allow partial responses (when people have fever, they might not be able to type correctly? What are good local conventions to keep people calm? done

Find emoticons to make the questions more emphatic [consult with psychologists, Bonus: Use local conventions depending on country]

Care taking personell has to be able to decide if they wanna see treatment timers, should we pick a default? Let the users decide!
Should the user see a treatment timer?

Architecture overview: