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

10th
Sep 2018
PieterGit
@PieterGit
Sep 10 2018 06:15
@tynbendad what version of oref0 do you use? I know there are some bluetooth issues with the old kernel (kernel PANIC's). I don't think its explorer board hardware related, but old kernel / ubilinux related. Do you have a newer explorer board, so that you can check that it's hardware related (use the same edison on older and newer Explorer board)? After reboot check the logs in /var/log to see if you can see errors
benfiscal
@benfiscal
Sep 10 2018 06:55
Those of you using Raspberry PI, what operating system do you use? Is that just a matter of personal taste?
tynbendad
@tynbendad
Sep 10 2018 06:56
@PieterGit 0.6.2 oref0, 6/23/17 jubilinux on the ones that hang and are 2016... i have a 2017 explorer but it has an older jubilinux 4/4/17 on it and doesn't hang
i guess i should try that jubilinux on the 2016's
Giuseppe
@giupo
Sep 10 2018 07:01
@benfiscal I use raspbian, but keep in mind that with some tweaks here-and-there I could loop from a virtualbox instance with debian hosted on macOS
So basically any linux based OS could do
mkericson
@mkericson
Sep 10 2018 10:56
Help Troubleshooting. My rig is looping and getting glucose values on my home network via wifi. But I'm not seeing pill information on NS. With the rig having communication what could be the cause?
Paper
@PaperT1D
Sep 10 2018 10:57
Are you underclocking?
mkericson
@mkericson
Sep 10 2018 11:09
Edison Rig
Scott Leibrand
@scottleibrand
Sep 10 2018 14:32
@benfiscal the docs, scripts, and code are written for and tested with raspbian lite.
benfiscal
@benfiscal
Sep 10 2018 14:55
@scottleibrand ok! Probably safest to go with raspbian lite.
dmelliott20
@dmelliott20
Sep 10 2018 17:29
Hi all wonder if someone can point me in the right direction. Im trying to understand bgi and dev in my loop and have read the docs but i am stil strugglig. Are these values in mmol or mg? i have a current bgi of -0.86 and a dev of 7.9 - i cant quite work out how these are relating to my BG. E.g. Does the bgi above mean my blood should be dropping by 0.86 every 5 mins? How does the 7.9 relate to that?
Tim Street
@tim2000s
Sep 10 2018 18:36
@scottleibrand - been seeing what appears to be odd carb absorption on 0.7.0. Carbs seem to be taking a long time to absorb (12g/hr rate roughly) then disappearing altogether randomly. Feel free to have a look at my NightScout if it’s of interest, but I’m struggling to make sense of it. The last three hours are a good example.
Scott Leibrand
@scottleibrand
Sep 10 2018 19:57
traveling in France right now, so probably won't have a chance to dig into it myself this week
@dmelliott20 for historical reasons, dev is in units of mg/dL/30m, not per 5m as you'd expect
I believe BGI is in mg/dL/5m even if you're using mmol.
Tim Street
@tim2000s
Sep 10 2018 20:02
Okay, no worries. Will keep an eye on it.
Andrew Kirkby
@kirkbysound
Sep 10 2018 20:14
I have a quick query. I've set up logger and am looking at options for entering carbs. Is there any way at all of doing so via the command line? I have a 522 without the capture event option and don't want to switch out of Oref1 SMB functions owing to its performance.. Thanks in advance!
Scott Leibrand
@scottleibrand
Sep 10 2018 20:24
You can use the bolus wizard if you learn all its nuances and figure out how to do so safely, minimize A52 risk, and safely recover from an A52 if one occurs.
OpenAPS will send ESC key presses before attempting to bolus, but if you’re holding down an arrow key, those won’t go through (they’ll just stop the scrolling), so you have to manually ESC our quickly
Zach Gohr
@zgohr
Sep 10 2018 20:26
It would be cool if there were a way from the command line. I’d like to have the ability from the Explorer HAT menu
(trivial to add to the menu if there was some command line way)
Scott Leibrand
@scottleibrand
Sep 10 2018 20:27
Alternatively you can do a small easy bolus first, wait for it to go through, and then more safely use the bolus wizard for carb entry, as OpenAPS won’t try to bolus within a couple minutes after a bolus.
The problem there is guaranteeing synchronization with NS.
Zach Gohr
@zgohr
Sep 10 2018 20:29
This is a true problem. Preferably it would work in camping mode
I could see it as an option such that when enabled, OpenAPS wouldn’t pull carbs from NS (instead only push them up, if possible/online.)
Jon Cluck
@cluckj
Sep 10 2018 20:49
Bolusing from the HAT is a really bad idea
(from a safety standpoint)
philipgo
@philipgo
Sep 10 2018 20:50
I just installed dev for the first time in a while. Calculated IOB is too low, apparently because OpenAPS fails to see some manual boluses. The boluses in question were uploaded to NS by OpenAPS, but now they are not included in pumphistory-24h-zoned.json Any hints how to track down the error?
The boluses are included in bolus history on the pump, I do not see any difference between those boluses and the ones still recognized by OpenAPS
Zach Gohr
@zgohr
Sep 10 2018 20:52
@cluckj I was talking about entering carbs
Jon Cluck
@cluckj
Sep 10 2018 20:53
Oh I misread; sorry!
philipgo
@philipgo
Sep 10 2018 20:59
Is there any way to manually force OpenAPS dev to re-read the entire pump history? It seems that deleting pumphistory-24h-zoned.json is not enough
alimhassam
@alimhassam
Sep 10 2018 21:07
@philipgo deleting pumphistory should do the trick, did that not work?
What issue are you having with the pump history?
Jeremy Cunningham
@jpcunningh
Sep 10 2018 21:07
I have had good luck with deleting settings/profile.json and settings/settings.json
That forces oref0-pump-loop to call its get_settings function.
alimhassam
@alimhassam
Sep 10 2018 21:08
Right that would be for re-reading settings if that's what you want to force update
philipgo
@philipgo
Sep 10 2018 21:09
@alimhassam How? By deleting which file? My issue is 4 posts above your reply, OpenAPS dev calculates IOB way too low and I noticed that some boluses are missing in pumphistory-24h-zoned.json. Strangely, they made their way to Nightscout before.
Jeremy Cunningham
@jpcunningh
Sep 10 2018 21:09
Yeah, it makes it reread everything
philipgo
@philipgo
Sep 10 2018 21:10
Is there any file I should preserve to allow tracking down the error in reading pump history (if that is the cause of the problem)
alimhassam
@alimhassam
Sep 10 2018 21:11
Would you mind posting the excerpt pumphistory that has missing bolus entry somewhere I'm curious how that happened
Fyi I have a pr to avoid such issues
Yes monitor/pumphistory-24h-zoned.json
Where bolus is Missing
Btw where the bolus done using the wizard?
philipgo
@philipgo
Sep 10 2018 21:15
NS.jpg
[
  {
    "timestamp": "2018-09-10T22:31:57+02:00",
    "_type": "TempBasalDuration",
    "duration (min)": 120
  },
  {
    "timestamp": "2018-09-10T22:31:57+02:00",
    "_type": "TempBasal",
    "temp": "absolute",
    "rate": 0
  },
  {
    "timestamp": "2018-09-10T22:20:39+02:00",
    "_type": "TempBasalDuration",
    "duration (min)": 120
  },
  {
    "timestamp": "2018-09-10T22:20:39+02:00",
    "_type": "TempBasal",
    "temp": "absolute",
    "rate": 0
  },
  {
    "timestamp": "2018-09-10T22:05:43+02:00",
    "_type": "TempBasalDuration",
    "duration (min)": 120
  },
  {
    "timestamp": "2018-09-10T22:05:43+02:00",
    "_type": "TempBasal",
    "temp": "absolute",
    "rate": 0
  },
  {
    "timestamp": "2018-09-10T22:01:11+02:00",
    "_type": "Bolus",
    "amount": 0.2,
    "programmed": 0.2,
    "unabsorbed": 4.8,
    "duration": 0
  },
  {
    "timestamp": "2018-09-10T22:00:50+02:00",
    "_type": "TempBasalDuration",
    "duration (min)": 60
  },
  {
    "timestamp": "2018-09-10T22:00:50+02:00",
    "_type": "TempBasal",
    "temp": "absolute",
    "rate": 0
  },
  {
    "timestamp": "2018-09-10T21:50:43+02:00",
    "_type": "TempBasalDuration",
    "duration (min)": 120
  },
  {
    "timestamp": "2018-09-10T21:50:43+02:00",
    "_type": "TempBasal",
    "temp": "absolute",
    "rate": 0
  },
  {
    "timestamp": "2018-09-10T21:45:45+02:00",
    "_type": "TempBasalDuration",
    "duration (min)": 90
  },
  {
    "timestamp": "2018-09-10T21:45:45+02:00",
    "_type": "TempBasal",
    "temp": "absolute",
    "rate": 0
  },
  {
    "timestamp": "2018-09-10T21:41:01+02:00",
    "_type": "Bolus",
    "amount": 0.175,
    "programmed": 0.175,
    "unabsorbed": 4.925,
    "duration": 0
  },
  {
    "timestamp": "2018-09-10T21:35:59+02:00",
    "_type": "Bolus",
    "amount": 0.325,
    "programmed": 0.325,
    "unabsorbed": 4.65,
    "duration": 0
  },
  {
    "timestamp": "2018-09-10T21:35:37+02:00",
    "_type": "TempBasalDuration",
    "duration (min)": 60
  },
  {
    "timestamp": "2018-09-10T21:35:37+02:00",
    "_type": "TempBasal",
    "temp": "absolute",
    "rate": 0
  },
  {
    "timestamp": "2018-09-10T21:26:15+02:00",
    "_type": "Bolus",
    "amount": 0.125,
    "programmed": 0.125,
    "unabsorbed": 4.625,
    "duration": 0
  },
  {
    "timestamp": "2018-09-10T21:25:51+02:00",
    "_type": "TempBasalDuration",
    "duration (min)": 60
  },
  {
    "timestamp": "2018-09-10T21:25:51+02:00",
    "_type": "TempBasal",
    "temp": "absolute",
    "rate": 0
  },
  {
    "timestamp": "2018-09-10T21:21:08+02:00",
    "_type": "Bolus",
    "amount": 0.5,
    "programmed": 0.5,
    "unabsorbed": 4.175,
    "duration": 0
  },
  {
    "timestamp": "2018-09-10T21:16:29+02:00",
    "_type": "Bolus",
    "amount": 0.4,
    "programmed": 0.4,
    "unabsorbed": 3.8,
    "duration": 0
  },
  {
    "timestamp": "2018-09-10T21:16:04+02:00",
    "_type": "TempBasalDuration",
    "duration (min)": 60
  },
  {
    "timestamp": "2018-09-10T21:16:04+02:00",
    "_type": "TempBasal",
    "temp": "absolute",
    "rate": 0
  },
  {
    "timestamp": "2018-09-10T21:11:09+02:00",
    "_type": "Bolus",
    "amount": 0.35,
    "programmed": 0.35,
    "unabsorbed": 3.5,
    "duration": 0
  }
]
Nope, regular boluses, no wizard. Pump history is relatively short since I switched pumps tonight. The boluses of 2 and 1U shown in Nightscout are not included in pumphistory-24h-zoned.json
philipgo
@philipgo
Sep 10 2018 21:21
I wonder if this might be have anything to do with using https://github.com/openaps/oref0/pull/1052/commits and the additional decimal places in boluses like 0.225U, but the boluses in question are full units
alimhassam
@alimhassam
Sep 10 2018 21:25
Really weird if they are in ns they must have been in pumphistory at some point
Ill
dmelliott20
@dmelliott20
Sep 10 2018 21:25
Thanks @scottleibrand - all makes sense now
alimhassam
@alimhassam
Sep 10 2018 21:26
Ah I guess the pump history is cut short
Can you manually run the pumphistory command and save that output as well
philipgo
@philipgo
Sep 10 2018 21:28
How would I do that?
alimhassam
@alimhassam
Sep 10 2018 21:28
When did you change pump?
philipgo
@philipgo
Sep 10 2018 21:31
First bolus on new pump was at 20:21, so pump history is indeed cut short and the first 4 boluses are now missing in OpenAPS pump history
I deleted settings/profile.json and settings/settings.json as @jpcunningh recommended and it did a full history refresh, but with the same result.
alimhassam
@alimhassam
Sep 10 2018 22:01
Type pumphistory in your shell
You might need to set the 2 environment variables
Sorry I can't paste exact command right now on mobile
philipgo
@philipgo
Sep 10 2018 22:17
@alimhassam Tried it and set the pump id, but apparently I also need to let it know that this is a WW pump. Couldn't figure out how to do that.
Just installed master on my second rig and interestingly master also cuts pump history short. So seems the pump or pump history is the cause.
alimhassam
@alimhassam
Sep 10 2018 22:21
Ww will be automatically set by setting the frequency variable
This message was deleted
export MEDTRONIC_PUMP_ID=`grep serial pump.ini | tr -cd 0-9`
export MEDTRONIC_FREQUENCY=`cat monitor/medtronic_frequency.ini`
Sorry trying to paste command on mobile
Eric
@ecc1
Sep 10 2018 22:25
The mobile version of this site is the worst. Unlike every other texting app, when your phone screen-blanks, or you switch to another task, it sends any partially-typed message, instead of leaving it there for you to finish. How brain-dead can you be? OK, done ranting for now
philipgo
@philipgo
Sep 10 2018 22:27
Hehe ;)
Thanks @alimhassam , that worked. Pump history is also cut short when using pumphistory
alimhassam
@alimhassam
Sep 10 2018 22:33
@ecc1 any ideas?
Try pumphistory -a
For all
(making sure you run it when loop is not blocking the radio)
philipgo
@philipgo
Sep 10 2018 22:45
@alimhassam pumphistory -a looks pretty much the same, the same entries are missing
alimhassam
@alimhassam
Sep 10 2018 22:52
I guess we'll need @ecc1 to help debug here no idea what to look at next
But in any case if it's missing in pumphistory -a I doubt even a full refresh would return the missing entries
Are you sure the bolus occurred in this pump?
philipgo
@philipgo
Sep 10 2018 22:59
@alimhassam Thanks a lot for your help! Yep, it is included in the bolus history on the pump itself. @ecc1 Is there any other output I should save before calling it a day? Where should I file the issue since this seems to be an issue with both master and dev?
Eric
@ecc1
Sep 10 2018 23:01
That's very strange. Can you fetch the relevant bolus entries from NS (using curl -gL .../api/v1/treatments.json?count=NNN" | jq . )
set NNN big enough to go back to the boluses you're interested in
and if it's both master and dev, then it's nothing to do with the Go pumphistory program (which is dev-only)
Also, please post (or PM me) the "pumphistory -a" output and tell me the time of the bolus entries that are on the pump itself but not in that output.
philipgo
@philipgo
Sep 10 2018 23:14
@ecc1 Thanks, I will PM you the details in order not to spam this room.
alimhassam
@alimhassam
Sep 10 2018 23:58
:+1: @ecc1 @philipgo
Without having looked at it much, Could there be some kind of off by one error when there are very few history page where page 0 is missed or something????