These are chat archives for nightscout/intend-to-bolus

27th
Sep 2018
Scott Leibrand
@scottleibrand
Sep 27 2018 00:40
@Wallys_world9_twitter yes, that is probably the rig hostname, so you’ll likely need to reflash it.
@tzachi-dar if you enter carbs into NS it’ll get those. But any boluses < DIA hours before you start looping will need to also be given on the loopable pump: it won’t pull insulin from NS.
@Diadon81 not sure exactly what that would look like, or how it would help.
Abigail Cember
@acember
Sep 27 2018 02:17
image.png
Sorry, that screenshot is completely illegible. Basically I have an error, the root of which (last line in there) is "TypeError: unsupported operand type(s) for -=: 'Retry' and 'int'". This doesn't look to me like the kind of issue that will get solved just by rerunning curl -s! Does anyone have any better advice?
Scott Leibrand
@scottleibrand
Sep 27 2018 02:31
Not sure what that is, but it does say timeout, so you might retry, or do some Googling.
Abigail Cember
@acember
Sep 27 2018 02:38
By golly, I actually did just rerun curl -s a bunch of times, and eventually it worked! Thanks, though, @scottleibrand for responding.
Diadon81
@Diadon81
Sep 27 2018 07:45
@scottleibrand Let me explain in more detail. Imagine, for example, that I want to write a cgm module that allows you to query data straight from miaomiao using bluetooth and bypassing the phone. Now in order to do this, I need to study the code oref0 in detail and integrate this module into it.
It would be much simpler if, in the framework of oref0, the communication with cgm and the pump was allocated to separate modules and the API would be clearly defined (as a set of calls that should be handled by these modules, for example get_bg (), etc.) thus this would simplify the development of individual modules to the basic algorithm oref0 and would allow them to be made more diverse. So anyone who want to develop their own cgm or pump interaction modules will have possibility to do this using any development tool (Python, Node, C and etc.)
tzachi-dar
@tzachi-dar
Sep 27 2018 09:22
@Diadon81 I was not able to follow your proposed design, but wanted to say that there is already code that uses python to read the miaomiao device: see here: https://github.com/tzachi-dar/LibreAllHouse
Diadon81
@Diadon81
Sep 27 2018 09:39
@tzachi-dar thank you, i will check and try it with my rig. I am just trying to say, that the oref0 code is huge and developed by different people, maybe it is a time to make it modular and more flexible. As for the LibreAllHouse, without modular design I should integrate LibreAllHouse code into the oref0 code and check all the code lines one by one to not miss anything. But CGM interaction and pump interaction can be easily simplified in term of oref0 core algorithm (get_bg, start_a_bolus, check_reservour and etc.) all background work can be made by separate entity. As for example external call for openaps to communicate with mdt pump. Main problem, that these calls has not been standardized from both side, so if i want for example my own code to interact with omnipod or other pumps in future i should spend a lot of time for integration with oref0.
tzachi-dar
@tzachi-dar
Sep 27 2018 10:03
@Diadon81 unfortunately I don't know oref0 design at all, so I can't really comment. but doesn't it have a place where it is taking BG data from? Looks like this is the place which you will have to change.
Diadon81
@Diadon81
Sep 27 2018 10:10
@tzachi-dar yes I can, but as I mention two messages before, I just share my thoughts on design just as a part of common oref0 discussion.
alimhassam
@alimhassam
Sep 27 2018 11:21
@Diadon81 to support a new cgm, the api is just to publish the bg entries to nightscout. Alternatively you could just try to produce a glucose.json file.
tzachi-dar
@tzachi-dar
Sep 27 2018 12:23
@scottleibrand Thanks for your answer.
Two following questions:
1) What will happen once the rig pump is not connected to his body? Does it mean that the rig will still try to give commands to the pump? In that case, will it have wrong values of IOB (I might have to stop it's working completely? but in that case, how do I give the boluses before I connect it?)
s.
@samberger
Sep 27 2018 12:35
Hello! Could anyone please point me to the ExplorerHAT gitter channel? Thanks! :)
Ebgineer
@Ebgineer
Sep 27 2018 13:30
@tzachi-dar you're proposing an unusual scenario with two pumps, one of which is always disconnected. So you're going to find unusual conditions that have to be addressed by hand. If the non-OpenAPS pump has given a bolus, and you immediately give the same bolus to the OpenAPS pump while it is disconnected, the IOB calculation will be close to correct. The only difference would be the fixed rate basal schedule on the non-OpenAPS vs. the presumably different and variable basal rate the rig would be sending to the OpenAPS pump. You'd have to calculate that difference by hand. Wouldn't it be easier to use a single pump and just turn off OpenAPS when you don't want it running?
alimhassam
@alimhassam
Sep 27 2018 14:22
@tzachi-dar just curious, whats the reason for this multiple pump setup? Why couldn't you just turn off the rig during the day and use the same pump day and night?
tzachi-dar
@tzachi-dar
Sep 27 2018 14:24
@Ebgineer, @alimhassam I just thought about it today, and indeed I always find myself in non standard conditions.
Specifically for the pump issue, well, I did a lot of searching and publishing in local facebook groups and was able to get a half dozen of old pumps, however, I now have 2 pumps with fw that is too new. one pump with good fw that does not work. One pump that is too old, and two pumps that are probably fine for the project. However, from the two pumps one is almost totally broken, both are not in hebrew.
So, going with the one pump that is working is problematic because 1) my son will not be able to use it (not in the local language) and 2) very big chance he will break it, so I'm trying to use it at night only.
I'll look at the implementation of oaps and see if there is something simple that can be done.
Aoife
@Stormybubbles
Sep 27 2018 15:14
Hi, am having issues I'm trying to resolve, I am seeing this in the logs mmeowlink.exceptions.InvalidPacketReceived on clock reset , does anyone have any suggestions???
Aoife
@Stormybubbles
Sep 27 2018 15:25
could it be a sign of the pump radio giving up?
Scott Leibrand
@scottleibrand
Sep 27 2018 15:53
What does mmtune say?
wfeddern
@wfeddern
Sep 27 2018 15:53
Anyone have issues with enhancedradio.com lately? They agreed to replace my pi HAT board (screen was no longer working and battery charging was acting strange). I sent them my broken board over a month ago (with tracking so know they received it), and all my attempts since then to get an update have been met with silence.
Scott Leibrand
@scottleibrand
Sep 27 2018 15:54
@s. Probably https://gitter.im/openaps/hardware-dev is the closest we have, although the ERD folks and Kelly’s don’t hang out there.
wfeddern
@wfeddern
Sep 27 2018 15:55
Considering trying to find an edison and build a second rig now, to avoid this kind of interruption to looping in the future
Scott Leibrand
@scottleibrand
Sep 27 2018 15:57
Yeah, backups are always a good idea.
Aoife
@Stormybubbles
Sep 27 2018 16:08
@scottleibrandep 27 15:36:28 Hal pump-loop.log: mmtune: "868.342", 3, -80 waiting for 40 second silence before continuing
Sep 27 15:41:34 Hal pump-loop.log: mmtune: No wait required.
Sep 27 15:44:38 Hal pump-loop.log: mmtune: "868.438", 1, -94 waiting for 68 second silence before continuing
Sep 27 15:50:10 Hal pump-loop.log: mmtune: No wait required.
Sep 27 15:54:11 Hal pump-loop.log: mmtune: No wait required.
Sep 27 15:57:50 Hal pump-loop.log: mmtune: "868.438", 5, -70 waiting for 20 second silence before continuing
Sep 27 16:08:53 Hal pump-loop.log: mmtune: "868.294", 5, -65 waiting for 10 second silence before continuing
Sep 27 16:22:58 Hal pump-loop.log: mmtune: "868.438", 1, -94 waiting for 68 second silence before continuing
Sep 27 16:38:31 Hal pump-loop.log: mmtune: No wait required.
Sep 27 16:43:06 Hal pump-loop.log: mmtune: "868.270", 5, -70 waiting for 20 second silence before continuing
Sep 27 16:49:33 Hal pump-loop.log: mmtune: "868.390", 5, -65 waiting for 10 second silence before continuing
Scott Leibrand
@scottleibrand
Sep 27 2018 16:37
How far was the pump from the rig during that timeframe?
-65 should be plenty good for looping. -94 not so much.
Aoife
@Stormybubbles
Sep 27 2018 16:40
sat right next to each other
Scott Leibrand
@scottleibrand
Sep 27 2018 16:44
Then yeah, if you’re getting worse than -80 at that range, there may be a problem with the pump or rig hardware. I would get a second rig or pump and see if that pair gets better results.
Aoife
@Stormybubbles
Sep 27 2018 16:46
thanks, I'm trying to get my hands on a second explorer board to rule that out, easier than trying to find a second pump!
Scott Leibrand
@scottleibrand
Sep 27 2018 16:47
If you have a non-loopable minimed pump you can also test with that.
Aoife
@Stormybubbles
Sep 27 2018 16:56
how? I have a non loopable pump
Eric
@ecc1
Sep 27 2018 17:16
@Stormybubbles all the read-only commands will still work, so you can run mmtune against it, as well as get pump history. You just can't change anything remotely, like temp basal
Aoife
@Stormybubbles
Sep 27 2018 17:24
perfect, thank you very much :)
adenis76
@adenis76
Sep 27 2018 22:12
I am instaled Edison driver. After connecting intel edison with explorer board to windows 10 pc device is shown like Merrifield with no driver. I cant connect to Edison thru Putty. Merrifield lose connection with pc after about 1 minute. Any one with simmilar trouble?
N3FM
@N3FM
Sep 27 2018 23:44
@adenis76 I had problems getting drivers to work in Windows 10. I had to use another pc on Windows 8.1, and run flashall.bat multiple times.