Support for group of kids

[pmithrandir]

Hi,

I saw that some people think about using OSD in institution.
In order to do that, the current system is not working that great. One phone can only listen to one patient.

I'm wondering if there is such a need, because building an app / a website that will list all devices, and track them, should not be hard.

We could imagine, like in the hospital in carers room, a monitor with all seizure charts visible next to each other, and the capacity to visualize alarms their.

(in the end each phone expose an API)

I'm just wondering if there is such a need.

[jones139]

If someone does have a use for it, I'd be inclined to use a small computer to serve a web page that would monitor and display several OSD instances - it would be easier than making it an android app, unless there is a use for it being on a mobile device....but even using a normal computer, the web page could be viewed on a mobile device if wanted.
I think it was @jbmalfilattre that was looking at this.

[pmithrandir]

Agree... a website seems much easier to manage and create.

[jbmalfilattre]

Question; wouldn't it be easier to just improve the existing website (see screenshot below)? As each monitoring device has its own IP address, it should be easy for the carer to have multiple browser sessions or multiple tabs open.

[jones139]

I was thinking that having multiple browser tabs open would be a pain - if we make it a single web page, we could have multiple 'tiles', showing an instance of OSD on each, all visible on the same page - and probably just displaying the equivalent of the 'Basic Mode' user interface in V5 - the OK/WARNING/ALARM status as a colour coded block so it stands out, and maybe a mute button beneath it to shut it up if it is going off constantly. Maybe something like this:

[jbmalfilattre]

Absolutely cool and very impressive! If that is not too much work, then please go ahead. Maybe adding the remaining battery power percentage of the monitoring device to that display would make sense?

[jones139]

That is, of course only a picture drawn in Powerpoint :).

[jones139]

So, if we are to do, this, what architecture?

I'd be inclined to go for:

• Backend - Python3, but which framework? I would probably go for Flask, but I think @pmithrandir has more experience in this area so might suggest something else.
• Configuration - for a proof of concept, a simple JSON configuration file that defines each instance that is to be monitored - will put a starter at the bottom of this post.
• Web page framework - something like bootstrap for the styling and a javascript framework to talk to the server. Now I'm old school and if I was just going to make something that works, I'd use jquery :) That is what the web page in the OSD app uses. I think React is the modern way, but I've never built a react app.

To make this a production system though we'd need to consider authentication (both between the web server and the OSD instances) and for access to the web server) and ease of configuration - a user interface to add/edit/delete instances to monitor.

The configuration file would be something like:

{
"title": "OSD Server",
"alarm_sound": "alarm.mp3",
"warning_sound": "warning.mp3",
"fault_sound": "fault.mp3",
"monitors": [
{
"title": "NAME_1",
"url":"http://xxx,yyy,zzz,aaa:8080/data",
"active": true
},
...
{
"title": "NAME_6",
"url":"http://xxx,yyy,zzz,aaa:8080/data",
"active": true
}
]
}

This is the sort of thing that is pretty straightforward, so I'd be inclined to give an AI coding tool the job of implementing it if we sort out a specification.

[pmithrandir]

Hi,

1st we need to know if there is a real need, and I mean real because it's going to be some work.
Then we must define the need exactly, because displaying the website might be already a first version.

In my opinion, what could be interesting is:
Alarm status (ML)
HR data
And mouvment in general. (power ?)

Because if you think about it, from the experience with my son, what worries me is:

• him being dead following a seizure (rare i agree, but real with some syndrome) or a pb with breathing requiring immediate action and emergency call.
• him moving a lot, or not. Because a seizure without moving, somehow worry me less (maybe not the same for all people)
• of course seizure to warn me of something happening.

Also a BIG question... do we want instant... or history.

If it's instant, maybe enough for a POC, then we could display each cellphone website using iframe and pure FE app.

If not, we will need to find a way to call each API on a regular basis, and save data locally.

But that's much more work and question to be raised.

[jbmalfilattre]

I tend to agree with @pmithrandir; if the work to integrate this is not straightforward, wouldn't there be more important items on the current "To Do" list?

Regarding "I think it was @jbmalfilattre that was looking at this"; I have unfortunately forgotten when I asked about that. Currently using Virtual Machines to view the screens of the phones on my main PC (which has no WLAN connection or Android OS, but is cable-connected to the LAN) is sufficient for my needs.

[jones139]

Ah, well if your virtual machine setup is working for you at the moment, let's park this one as a future development, because it is not something that I would use, but would do if someone wanted it :). Will keep this open so people can find it and comment on it in the future.

I was just thinking that if we don't have one, I'll create a 'Development Priorities' discussion page, so people can see what I am keenest on and they can add their thoughts.

Thanks!

[pmithrandir]

@jones139 do you think you could update the embedded homepage http somehow with the last alarms ?
Or remove it for good if not used.

That look like deprecated stuff

[jones139]

Yes, good idea - I haven't updated that page for a long time (because I never look at it!). I'll create a separate issue for it to include it in V5.1.