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

22nd
Jun 2016
Sarah Hodson Grady
@shgrady
Jun 22 2016 00:07
This message was deleted
This message was deleted
Dragan Maksimovic
@dm61
Jun 22 2016 00:13
@Spazholio looks like you have communication failures during pump history reading; I've seen such errors, but not very often, a few times a day
Paul Martin
@Spazholio
Jun 22 2016 00:15
@dm61 Oh, I’m used to seeing errors like that occasionally. This one was persistently the SAME error. I’ve switched the battery, as I was down to 1.38v and that’s where it gets iffy.
Dragan Maksimovic
@dm61
Jun 22 2016 00:20
I have not noticed that frequency of pump comm errors depends on the pump battery voltage (until it just shuts down the radio, of course); will look more carefully - this would be good to know
Paul Martin
@Spazholio
Jun 22 2016 00:21
Well, I’m still getting the error, so it wasn’t that. But my battery went from 4 bars to 3, and I’m using lithium which makes me think it’s about to die. May as well swap it out.
I’ve rebooted the Edison as well, and it seems to not be making a diffference.
I usually go 3m-5m between my loop running, and now it’s going 12m-15m.
Dragan Maksimovic
@dm61
Jun 22 2016 00:23
what's the signal level reported by mmtune?
Paul Martin
@Spazholio
Jun 22 2016 00:26
Admittedly not great. Gimme a minute, let me see if I can make a quick change...
Moved the rig just in time to catch the last preflight:
STARTING PREFLIGHT
Listening: .No comms detected
{
  "scanDetails": [
    [
      "916.300", 
      0, 
      -99
    ], 
    [
      "916.324", 
      0, 
      -99
    ], 
    [
      "916.348", 
      1, 
      -98
    ], 
    [
      "916.372", 
      0, 
      -99
    ], 
    [
      "916.396", 
      0, 
      -99
    ], 
    [
      "916.420", 
      0, 
      -99
    ], 
    [
      "916.444", 
      0, 
      -99
    ], 
    [
      "916.468", 
      0, 
      -99
    ], 
    [
      "916.492", 
      0, 
      -99
    ], 
    [
      "916.516", 
      0, 
      -99
    ], 
    [
      "916.540", 
      0, 
      -99
    ], 
    [
      "916.564", 
      0, 
      -99
    ], 
    [
      "916.588", 
      0, 
      -99
    ], 
    [
      "916.612", 
      0, 
      -99
    ], 
    [
      "916.636", 
      0, 
      -99
    ], 
    [
      "916.660", 
      0, 
      -99
    ], 
    [
      "916.684", 
      5, 
      -64
    ], 
    [
      "916.708", 
      5, 
      -57
    ], 
    [
      "916.732", 
      5, 
      -56
    ], 
    [
      "916.756", 
      5, 
      -56
    ], 
    [
      "916.780", 
      5, 
      -56
    ], 
    [
      "916.804", 
      5, 
      -57
    ], 
    [
      "916.828", 
      5, 
      -64
    ], 
    [
      "916.852", 
      0, 
      -99
    ], 
    [
      "916.876", 
      0, 
      -99
    ]
  ], 
  "setFreq": 916.78, 
  "usedDefault": false
}PREFLIGHT OK
And when it was failing, it was WAY worse.
I switched pockets that my rig was in. I mean…it’s a TI stick. I can get it to cover my whole house! I figured it could work through my legs.
Still failing though, even with good signal. Hmm.
ryanchen26
@ryanchen26
Jun 22 2016 00:30
@scottleibrand Auto sensitivity and meal assist are now running in the loop! Thanks for the info. We will now watch it like hawks with smartwatches and tablets...
Dana Lewis
@danamlewis
Jun 22 2016 00:30
:+1:
Dragan Maksimovic
@dm61
Jun 22 2016 00:33
@Spazholio -56 should be just fine; but, the TI stick, while being so much better than CareLInk, also seems to have more randomness. I've done tests with exact same fixed distance - about 6ft - between the stick and the pump and the signal level in repeated mmtune's would vary a lot, between -40 to -70. A successful preflight in the loop does not mean at all that subsequent pump comms will be successful. Any single one can fail, fortunately not that often. The pump history reading is the longest comm, probably most likely to fail.
Paul Martin
@Spazholio
Jun 22 2016 00:38
Yeah, I’ve certainly encountered that. My concern is that it’s the EXACT same error message, and that’s one I’ve never seen before. Just weird is all.
Dragan Maksimovic
@dm61
Jun 22 2016 00:43
yes, weird. Given how TI stick works, I am thinking about adding a retry to each pump comm step in the loop instead of a retry of the entire loop. Just one retry attempt should reduce the probability of overall loop failure substantially, as long as the stick within reasonable range.
Garrett Webb
@garetis
Jun 22 2016 00:47
Hey guys, I feel like I'm staring at the finish line, but I just need a little help. I just got my dexcom today, plugged it in, and now that I have everything I need, when I run openaps report invoke enact/suggested.json, I get that BG data is too old, or clock set incorrectly. I haven't seen anything in the docs about this, and don't know where to start. Could someone point me in the right direction?
Dragan Maksimovic
@dm61
Jun 22 2016 00:55
@garetis your bg data needs to be fresh enough. you may try to enact again, but do so just after you've generated all the necessary reports.
Garrett Webb
@garetis
Jun 22 2016 00:56
Well, when I get the error, it drops in two different times that are different, so it makes me think it's the clock issue...
Sarah Hodson Grady
@shgrady
Jun 22 2016 00:56
Does anyone have an example of what you use patterns, profiles A and B for? Is it like school days v. weekends?
Garrett Webb
@garetis
Jun 22 2016 00:57
@dm61
pi@raspberrypi:~/openaps2 $ openaps gather
get-profile://text/shell/settings/profile.json
reporting settings/profile.json
cgm://JSON/iter_glucose/monitor/glucose.json
reporting monitor/glucose.json
pump://JSON/read_clock/monitor/clock.json
reporting monitor/clock.json
pump://JSON/read_temp_basal/monitor/temp_basal.json
reporting monitor/temp_basal.json
pump://JSON/iter_pump_hours/monitor/pumphistory.json
reporting monitor/pumphistory.json
calculate-iob://JSON/shell/monitor/iob.json
reporting monitor/iob.json
calculate-iob://JSON/shell/monitor/iob.json
reporting monitor/iob.json
pi@raspberrypi:~/openaps2 $ openaps report invoke enact/suggested.json
determine-basal://text/shell/enact/suggested.json
BG data is too old, or clock set incorrectly Tue Jun 21 2016 19:54:48 GMT-0500 (CDT) vs Mon Jun 20 2016 08:07:20 GMT-0500 (CDT)
reporting enact/suggested.json
Paul Martin
@Spazholio
Jun 22 2016 00:58
@dm61 I’d LOVE it if you were to set something like that up. I know it’s just one iteration of the loop to wait for, but man, if it could retry the one failed report instead of having to delay and loop again (only to possibly fail) that’d be sweet.
Garrett Webb
@garetis
Jun 22 2016 00:58
@shgrady I think that's the idea, but I never found any use for them
Paul Martin
@Spazholio
Jun 22 2016 01:00
@shgrady I’ve heard that people (I think @danamlewis, but I could be wrong) have patterns set for when they’re planning to be active vs. sedentary.
Dana Lewis
@danamlewis
Jun 22 2016 01:00
@shgrady Or active days vs sick days
jinx @Spazholio :)
Paul Martin
@Spazholio
Jun 22 2016 01:00
Weekdays vs. weekends are a legitimate use as well.
;)
Sarah Hodson Grady
@shgrady
Jun 22 2016 01:00
this Medtronic is sooooo retro, I use the 5 profiles on my current pump just to keep track of the date of any changes.
Dana Lewis
@danamlewis
Jun 22 2016 01:00
standard is for normal every day, I have pattern A as “super active days”, pattern B was before we got timezones figured out so it was basically my standard pattern shifted two hours to central time ;)
But OpenAPS and especially autosens has negated the need for switching profiles for the most part
Sarah Hodson Grady
@shgrady
Jun 22 2016 01:01
whoooo that would be cool if my Rheumatoid Arthritis had a predictable blood sugar effect
Paul Martin
@Spazholio
Jun 22 2016 01:01
@garetis If you’re having date issues like that…actually, it looks like your system time is off. What’s the output of the date command?
Dragan Maksimovic
@dm61
Jun 22 2016 01:01
@garetis, are the time/date on your CGM correct? I see a Tue and a Wed in that next to last line; one of the clocks must be off
Garrett Webb
@garetis
Jun 22 2016 01:02
It's off. Mon 20 Jun 08:11:48 CDT 2016
Paul Martin
@Spazholio
Jun 22 2016 01:02
Do you know how to set it properly? I’ve got a nifty script for you, if not. =)
Garrett Webb
@garetis
Jun 22 2016 01:02
I'm reading date --help
Paul Martin
@Spazholio
Jun 22 2016 01:03
Let me make your life easy. =)
In my ~ dir, I have a file called fixtime.sh. So nano fixtime.sh.
Then paste in the following:
sudo /etc/init.d/ntp stop                                                                                                                                                                                                                                                                                             
sudo ntpd -q -g
sudo /etc/init.d/ntp start
(I’m assuming you have ntp installed, but you probably do.
Once that file’s created/saved, you have to run an additional command. Let me know when you’re there.
Ben West
@bewest
Jun 22 2016 01:04
@Spazholio if you put it in a gist, you could tell people to wget/curl it
Paul Martin
@Spazholio
Jun 22 2016 01:04
...
Showoff.
=)
Garrett Webb
@garetis
Jun 22 2016 01:04
I may, but ntp is not ringing a bell...
Ben West
@bewest
Jun 22 2016 01:04
then there's even less complexity, or you could nominate it for oref0 in a PR :-)
Sarah Hodson Grady
@shgrady
Jun 22 2016 01:04
@danamlewis thanks, those are all good ideas.
Paul Martin
@Spazholio
Jun 22 2016 01:04
What’s a “PR”?
Spazholio @Spazholio ducks and runs
Dana Lewis
@danamlewis
Jun 22 2016 01:04
hahahahahaha
danamlewis @danamlewis takes back the nice things she was about to say about @spazholio paying it forward ;)
Paul Martin
@Spazholio
Jun 22 2016 01:05
Totally worth it. =)
Dana Lewis
@danamlewis
Jun 22 2016 01:05
(I’m kidding. But I’ll only say a few nice things ;))
Garrett Webb
@garetis
Jun 22 2016 01:05
I'll do a first-pass on a PR as a "thank you" =)
Dana Lewis
@danamlewis
Jun 22 2016 01:05
:D
BTW, I’m just going to leave this here… :microphone: (drop)
Paul Martin
@Spazholio
Jun 22 2016 01:06
Once that file is completed, you need to make it executable. You do that by typing chmod +x ./fixtime.sh
Garrett Webb
@garetis
Jun 22 2016 01:06
How could I check if I have ntp?
Paul Martin
@Spazholio
Jun 22 2016 01:07
Run this: sudo /etc/init.d/ntp stop
Does it error out?
Brian Bonner
@itsn0tm3
Jun 22 2016 01:08
@Spazholio I've got the exact same problem with my pump with pumphistory
Garrett Webb
@garetis
Jun 22 2016 01:08
I got an [ok] and some other stuff
So, looks like it
Brian Bonner
@itsn0tm3
Jun 22 2016 01:09
@Spazholio but I don't think it's the battery :( I think something else is wonky because I can run it manually and it runs just fine, but when I run it as a report it fails
repeatedly
driving me batty
what no :crazy: emoticon!?
Paul Martin
@Spazholio
Jun 22 2016 01:11
@itsn0tm3 Yeah, it’s working now. Ish. Sorta. Sometimes.
Garrett Webb
@garetis
Jun 22 2016 01:11
@Spazholio I think I just made it executable.
Paul Martin
@Spazholio
Jun 22 2016 01:11
Ok, now do this: ./fixtime.sh
Brian Bonner
@itsn0tm3
Jun 22 2016 01:12
@Spazholio heh I changed mine to 14 hours instead of 28 so it still can do the autosens somewhat OK but it fails a lot less
Paul Martin
@Spazholio
Jun 22 2016 01:12
That’ll run it. You might get a slight error for the first line since you already stopped ntp but that’s not an issue.
@itsn0tm3 Oh, it’s not even pumphistory-24h.json that’s causing issues. It’s the standard pumphistory.json. Grr.
Garrett Webb
@garetis
Jun 22 2016 01:13
I'm getting this, and it hasn't returned to the prompt yet:
pi@raspberrypi:~/openaps2 $ ./fixtime.sh
[ ok ] Stopping ntp (via systemctl): ntp.service.
Would this work? sudo ntpdate -s time.nist.gov
Brian Bonner
@itsn0tm3
Jun 22 2016 01:15
@Spazholio do you by any chance recall when it started going wonky? I'm almost positive it was after I updated my openaps/oref0 that it went wonky but I did so many things that my brain cells have congealed into one.
Paul Martin
@Spazholio
Jun 22 2016 01:15
I honestly don’t know. The one I shared with you basically forces it to fix itself. I think it does similar to your command. But hey, give yours a shot?
Whatever works for you, you know?
Brian Bonner
@itsn0tm3
Jun 22 2016 01:16
@garetis I use ntpdate but you have to make sure ntpd is stopped first tho since it uses the same port
you will see an error with ntpdate if you try to use it when ntpd is running
Garrett Webb
@garetis
Jun 22 2016 01:17
I found that by googling. But, I feel like the script was toggled and kept stopping it, but I think this may have gotten it...
pi@raspberrypi:~/openaps2 $ sudo /etc/init.d/ntp stop
[ ok ] Stopping ntp (via systemctl): ntp.service.
pi@raspberrypi:~/openaps2 $ nano fixtime.sh
pi@raspberrypi:~/openaps2 $ chmod +x ./fixtime.sh
pi@raspberrypi:~/openaps2 $ ^C
pi@raspberrypi:~/openaps2 $ ./fixtime.sh
[ ok ] Stopping ntp (via systemctl): ntp.service.
[ ok ..] Starting ntp (via systemctl): ntp.service.
pi@raspberrypi:~/openaps2 $ ./fixtime.sh
[ ok ] Stopping ntp (via systemctl): ntp.service.
[ ok ..] Starting ntp (via systemctl): ntp.service.
pi@raspberrypi:~/openaps2 $ ./fixtime.sh
[ ok ] Stopping ntp (via systemctl): ntp.service.
^Z
[1]+  Stopped                 ./fixtime.sh
pi@raspberrypi:~/openaps2 $ ./fixtime.sh
[ ok ] Stopping ntp (via systemctl): ntp.service.
[ ok ] Starting ntp (via systemctl): ntp.service.
pi@raspberrypi:~/openaps2 $ date
Mon 20 Jun 08:26:48 CDT 2016
pi@raspberrypi:~/openaps2 $ ./fixtime.sh
[ ok ] Stopping ntp (via systemctl): ntp.service.
[ ok ] Starting ntp (via systemctl): ntp.service.
pi@raspberrypi:~/openaps2 $ date
Mon 20 Jun 08:26:59 CDT 2016
pi@raspberrypi:~/openaps2 $
Is there a delay for the time to fix? Well, at least I have a path. I'll play with the code and work on updating the system time. Once I get it right I will work on a PR at some point tonight.
Paul Martin
@Spazholio
Jun 22 2016 01:20
@garetis Here’s what mine looks like:
[ ok ] Stopping ntp (via systemctl): ntp.service.
ntpd: time slew +0.012145s
[ ok ] Starting ntp (via systemctl): ntp.service.
Garrett Webb
@garetis
Jun 22 2016 01:21
That's when you run ./fixtime.sh?
Paul Martin
@Spazholio
Jun 22 2016 01:21
Once you have the ntp service stopped, does the command sudo ntpd -q -g return anything?
Yes it is.
Garrett Webb
@garetis
Jun 22 2016 01:22
It just returns to prompt
pi@raspberrypi:~/openaps2 $ sudo /etc/init.d/ntp stop
[ ok ] Stopping ntp (via systemctl): ntp.service.
pi@raspberrypi:~/openaps2 $ sudo ntpd -q -g
pi@raspberrypi:~/openaps2 $
Paul Martin
@Spazholio
Jun 22 2016 01:23
Interesting. Mine always returns a time slew.
I’m running it in rapid-fire succession, and always getting something back.
Brian Bonner
@itsn0tm3
Jun 22 2016 01:24
hmm
I think I know why
pi@livingroomaps:~ $ sudo ntpd -q -g
pi@livingroomaps:~ $ sudo /etc/init.d/ntp stop
[ ok ] Stopping ntp (via systemctl): ntp.service.
pi@livingroomaps:~ $ sudo ntpd -q -g
ntpd: time slew -0.000400s
pi@livingroomaps:~ $
notice something?
if your ntpd is running, you get nothing back, but if you stop it and hten run it, you'll get something back
so that means your ntpd is still running
Paul Martin
@Spazholio
Jun 22 2016 01:25
I dunno, he showed pretty clearly that he was running the commands one right after the other. Hmm.
Garrett Webb
@garetis
Jun 22 2016 01:26
pi@raspberrypi:~/openaps2 $ date
Mon 20 Jun 08:26:59 CDT 2016
pi@raspberrypi:~/openaps2 $ sudo /etc/init.d/ntp stop
[ ok ] Stopping ntp (via systemctl): ntp.service.
pi@raspberrypi:~/openaps2 $ sudo ntpd -q -g
pi@raspberrypi:~/openaps2 $ nano fixtime.sh
pi@raspberrypi:~/openaps2 $ sudo ntpd -q -g
pi@raspberrypi:~/openaps2 $ sudo /etc/init.d/ntp stop
[ ok ] Stopping ntp (via systemctl): ntp.service.
pi@raspberrypi:~/openaps2 $ sudo ntpd -q -g
pi@raspberrypi:~/openaps2 $
Brian Bonner
@itsn0tm3
Jun 22 2016 01:26
@garetis can you stop it, and then check ps ax to make sure nothing is running?
and by nothing I mean ntpd just to clarify
Paul Martin
@Spazholio
Jun 22 2016 01:27
@itsn0tm3 Also, I noticed it going wonly today, after I set up my Pebble watch. I literally can’t fathom how that has ANYTHING to do with it, but I haven’t touched anything else.
Garrett Webb
@garetis
Jun 22 2016 01:27
pi@raspberrypi:~/openaps2 $ sudo /etc/init.d/ntp stop
[ ok ] Stopping ntp (via systemctl): ntp.service.
pi@raspberrypi:~/openaps2 $ ps ax
  PID TTY      STAT   TIME COMMAND
    1 ?        Ss     0:04 /sbin/init
    2 ?        S      0:00 [kthreadd]
    3 ?        S      0:00 [ksoftirqd/0]
    5 ?        S<     0:00 [kworker/0:0H]
    7 ?        S      0:02 [rcu_sched]
    8 ?        S      0:00 [rcu_bh]
    9 ?        S      0:00 [migration/0]
   10 ?        S      0:00 [migration/1]
   11 ?        S      0:00 [ksoftirqd/1]
   13 ?        S<     0:00 [kworker/1:0H]
   14 ?        S      0:00 [migration/2]
   15 ?        S      0:00 [ksoftirqd/2]
   17 ?        S<     0:00 [kworker/2:0H]
   18 ?        S      0:00 [migration/3]
   19 ?        S      0:00 [ksoftirqd/3]
   21 ?        S<     0:00 [kworker/3:0H]
   22 ?        S      0:00 [kdevtmpfs]
   23 ?        S<     0:00 [netns]
   24 ?        S<     0:00 [perf]
   25 ?        S      0:00 [khungtaskd]
   26 ?        S<     0:00 [writeback]
   27 ?        S<     0:00 [crypto]
   28 ?        S<     0:00 [bioset]
   29 ?        S<     0:00 [kblockd]
   31 ?        S<     0:00 [rpciod]
   32 ?        S      0:00 [kswapd0]
   33 ?        S<     0:00 [vmstat]
   34 ?        S      0:00 [fsnotify_mark]
   35 ?        S<     0:00 [nfsiod]
   44 ?        S<     0:00 [kthrotld]
   47 ?        S<     0:00 [bioset]
   48 ?        S<     0:00 [bioset]
   49 ?        S<     0:00 [bioset]
   50 ?        S<     0:00 [bioset]
   51 ?        S<     0:00 [bioset]
   52 ?        S<     0:00 [bioset]
   53 ?        S<     0:00 [bioset]
   54 ?        S<     0:00 [bioset]
   55 ?        S<     0:00 [bioset]
   56 ?        S<     0:00 [bioset]
   57 ?        S<     0:00 [bioset]
   58 ?        S<     0:00 [bioset]
   59 ?        S<     0:00 [bioset]
   60 ?        S<     0:00 [bioset]
   61 ?        S<     0:00 [bioset]
   62 ?        S<     0:00 [bioset]
   63 ?        S<     0:00 [bioset]
   64 ?        S<     0:00 [bioset]
   65 ?        S<     0:00 [bioset]
   66 ?        S<     0:00 [bioset]
   67 ?        S<     0:00 [bioset]
   68 ?        S<     0:00 [bioset]
   69 ?        S<     0:00 [bioset]
   70 ?        S<     0:00 [bioset]
   71 ?        S<     0:00 [VCHIQ-0]
   72 ?        S<     0:00 [VCHIQr-0]
   73 ?        S<     0:00 [VCHIQs-0]
   74 ?        S<     0:00 [iscsi_eh]
   75 ?        S<     0:00 [dwc_otg]
   76 ?        S<     0:00 [DWC Notificatio]
   78 ?        S      0:04 [irq/92-mmc1]
   79 ?        S      0:00 [VCHIQka-0]
   80 ?        S<     0:00 [SMIO]
   81 ?        S<     0:00 [deferwq]
   83 ?        S<     0:00 [bioset]
   84 ?        S      0:00 [mmcqd/0]
   86 ?        S      0:00 [jbd2/mmcblk0p2-]
   87 ?        S<     0:00 [ext4-rsv-conver]
   89 ?        S<     0:00 [ipv6_addrconf]
  135 ?        Ss     0:00 /lib/systemd/systemd-journald
  140 ?        Ss     0:00 /lib/systemd/systemd-udevd
  237 ?        S<     0:00 [cfg80211]
  240 ?        S<     0:00 [brcmf_wq/mmc1:0]
  244 ?        S      0:02 [brcmf_wdog/mmc1]
  427 ?        Ss     0:00 /usr/sbin/cron -f
  433 ?        Ss     0:00 /lib/systemd/systemd-logind
  439 ?        Ss     0:07 avahi-daemon: running [raspberrypi.local]
  442 ?        Ss     0:00 /usr/bin/dbus-daemon --system --address=systemd: --nofork --nopidfile --systemd-activation
  447 ?        Ss     0:00 /sbin/dhcpcd -q -b
  456 ?        Ss     0:00 /usr/sbin/thd --daemon --triggers /etc/triggerhappy/triggers.d/ --socket /var/run/thd.socket --pidfile /var/run/thd.pid --user nobody /dev/input/event*
  464 ?        S      0:00 avahi-daemon: chroot helper
  495 ?        S<     0:00 [kworker/0:1H]
  500 ?        Ssl    0:00 /usr/sbin/rsyslogd -n
  510 ?        Ss     0:00 /usr/sbin/sshd -D
  539 ?        Ss     0:02 /sbin/wpa_supplicant -s -B -P /run/wpa_supplicant.wlan0.pid -i wlan0 -D nl80211,wext -c /etc/wpa_supplicant/wpa_supplicant.conf
  546 ?        Ssl    0:00 /usr/sbin/lightdm
  556 tty1     Ss     0:00 /bin/login -f
  645 ?        S<     0:00 [kworker/3:1H]
  649 ?        S<     0:00 [kworker/u9:0]
  650 ?        S<     0:0
Brian Bonner
@itsn0tm3
Jun 22 2016 01:29
try ps ax | grep 'ntpd' instead
if nothing shows then try ntpd --version
Garrett Webb
@garetis
Jun 22 2016 01:29
pi@raspberrypi:~/openaps2 $ ps ax | grep 'ntpd'
 2735 pts/0    T      0:00 sudo ntpd -q -g
 2739 pts/0    T      0:00 ntpd -q -g
 3129 pts/0    S+     0:00 grep --color=auto ntpd
pi@raspberrypi:~/openaps2 $
Brian Bonner
@itsn0tm3
Jun 22 2016 01:30
oh
looks like it's stuck running in the background
try typing fg
then you can ctrl-c it if that works
@Spazholio that is weird
Paul Martin
@Spazholio
Jun 22 2016 01:31
I know, right? =)
Garrett Webb
@garetis
Jun 22 2016 01:31
It's stuck on the prompt:
pi@raspberrypi:~/openaps2 $ fg
./fixtime.sh
^C^C
^[
^X
Paul Martin
@Spazholio
Jun 22 2016 01:32
Maybe do a reboot? Let’s just get back to a clean slate?
Garrett Webb
@garetis
Jun 22 2016 01:33
pi@raspberrypi:~/openaps2 $ fg
./fixtime.sh
^C^C
^[
^X
reboot
sudo shutdown -h now
Brian Bonner
@itsn0tm3
Jun 22 2016 01:33
kick it back into the backgroun then you can do that
ctrl-z
Paul Martin
@Spazholio
Jun 22 2016 01:33
Yeah, what @itsn0tm3 said.
Garrett Webb
@garetis
Jun 22 2016 01:34
It's not resonding...
Same thing in my previous post, but tack on a couple ^Zs
Brian Bonner
@itsn0tm3
Jun 22 2016 01:35
pull the power and plug it back in would be my next idea
Paul Martin
@Spazholio
Jun 22 2016 01:35
Yup.
If it STILL won’t reboot, then back away slowly, because it might be sentient at this point.
Garrett Webb
@garetis
Jun 22 2016 01:35
Ok, thanks, just checking...
Brian Bonner
@itsn0tm3
Jun 22 2016 01:36
LOL
@bewest do you by any chance have anything that can look through pumphistory and show any errors that might trip openaps?
Ben West
@bewest
Jun 22 2016 01:38
it shouldn't trip up
Paul Martin
@Spazholio
Jun 22 2016 01:38
Agreed. =)
Brian Bonner
@itsn0tm3
Jun 22 2016 01:39
then why is it that I can run it manually and then try running the report and it fails, then I try running it manually again and it works fine?
I looked up the report with the --cli option and they're identical
Ben West
@bewest
Jun 22 2016 01:39
what's the error?
Paul Martin
@Spazholio
Jun 22 2016 01:39
This is the pumphistory.json report?
Brian Bonner
@itsn0tm3
Jun 22 2016 01:39
it's driving me loony enough that I've put my edison rig in the corner and told it it to stay there ha
Lynne Butts
@flutelynne_twitter
Jun 22 2016 01:39
@Spazholio For less lag time on your NS numbers, did you try setting your NS settings (in heroku or azure) to BRIDGE_INTERVAL 60000, and BRIDGE_MINUTES 1400?
Brian Bonner
@itsn0tm3
Jun 22 2016 01:40
yea my pumphistory.json is 28 hours tho before I changed it to 14
@bewest invalid packets on transmission for some, zero value for some
Paul Martin
@Spazholio
Jun 22 2016 01:40
@bewest: I can’t speak for @itsn0tm3, but I’m getting this fairly consistently, on two separate rigs:
pump://JSON/iter_pump_hours/settings/pumphistory-24h.json
Traceback (most recent call last):
  File "/usr/local/bin/openaps-report", line 82, in <module>
    app( )
  File "/usr/local/lib/python2.7/dist-packages/openaps/cli/__init__.py", line 51, in __call__
    self.run(self.args)
  File "/usr/local/bin/openaps-report", line 75, in run
    output = app(args, self)
  File "/usr/local/lib/python2.7/dist-packages/openaps/cli/subcommand.py", line 52, in __call__
    return self.method.main(args, app)
  File "/usr/local/lib/python2.7/dist-packages/openaps/reports/invoke.py", line 40, in main
    output = task.method(args, app)
  File "/usr/local/lib/python2.7/dist-packages/openaps/uses/use.py", line 45, in __call__
    output = self.main(args, app)
  File "/usr/local/lib/python2.7/dist-packages/openaps/vendors/medtronic.py", line 639, in main
    for rec in self.range( ):
  File "/usr/local/lib/python2.7/dist-packages/decocare/models/__init__.py", line 67, in __call__
    for record in page:
  File "/usr/local/lib/python2.7/dist-packages/decocare/models/__init__.py", line 46, in download_page
    for record in self.find_records(page):
  File "/usr/local/lib/python2.7/dist-packages/decocare/models/__init__.py", line 157, in find_records
    decoder = history.HistoryPage(response.data, self.inst)
  File "/usr/local/lib/python2.7/dist-packages/decocare/history.py", line 777, in __init__
    if lib.BangInt(crc) != computed:
  File "/usr/local/lib/python2.7/dist-packages/decocare/lib.py", line 235, in BangInt
    ( x, y ) = ints
ValueError: need more than 0 values to unpack
Brian Bonner
@itsn0tm3
Jun 22 2016 01:40
yeah that one
Paul Martin
@Spazholio
Jun 22 2016 01:40
I have no idea what the “BangInt” is, but I’m fairly certain I’ve not seen that term in any of the errors before.
Brian Bonner
@itsn0tm3
Jun 22 2016 01:40
@Spazholio I bet if you look in syslog you'll see invalid packets received error
Ben West
@bewest
Jun 22 2016 01:40
did you have medtronic cgm earlier?
that's not openaps, actually
that's decocare not understanding something that came back from the pump
this isn't that surprising, since I didn't have a manual of all the things it does
Brian Bonner
@itsn0tm3
Jun 22 2016 01:41
hrmm
Ben West
@bewest
Jun 22 2016 01:42
you can get a raw data file using mm-send-comm.py --serial 123456 --init tweak ReadHistoryData --save --page 0
it should decode cleanly with mm-decode-history-page.py
Brian Bonner
@itsn0tm3
Jun 22 2016 01:42
can I increment the page number until I hit an error then I know specifically where it is?
Ben West
@bewest
Jun 22 2016 01:43
yeah
Brian Bonner
@itsn0tm3
Jun 22 2016 01:43
hmm ok let me give that a try
Paul Martin
@Spazholio
Jun 22 2016 01:44
I’ll do the same, see if anything horks.
Garrett Webb
@garetis
Jun 22 2016 01:44
@Spazholio Haha, I think that's one problem I won't have to deal with tonight. I'm not sure how to interpret this...
pi@raspberrypi:~ $ cd openaps2
pi@raspberrypi:~/openaps2 $
pi@raspberrypi:~/openaps2 $ ./fixtime.sh
[ ok ] Stopping ntp (via systemctl): ntp.service.
[ ok ..] Starting ntp (via systemctl): ntp.service.
pi@raspberrypi:~/openaps2 $
Paul Martin
@Spazholio
Jun 22 2016 01:44
(I really shouldn’t have my pump serial memorized. That HAS to be taking up valuable space…)
Ben West
@bewest
Jun 22 2016 01:44
there's a mmeowlink-send.py tool that works with ti stick
Sarah Hodson Grady
@shgrady
Jun 22 2016 01:44
since I now have two pumps, I need to do all the same stuff for the other one? add device and create all the reports? or just update the serial number ?
Ben West
@bewest
Jun 22 2016 01:44
does same thing as mm-send-comm.py
Brian Bonner
@itsn0tm3
Jun 22 2016 01:45
ooh thanks that would have tripped me up
Ben West
@bewest
Jun 22 2016 01:45
you can just change the serial number using openaps use pump config --serial 123456
you can also back up entire loop and import it somewhere else
oref0 export-loop exports everything
openaps import imports it again
Sarah Hodson Grady
@shgrady
Jun 22 2016 01:46
but is that the best, to change the serial number? I was hoping to do quicker switches between the models, or is that not possible?
Ben West
@bewest
Jun 22 2016 01:47
what do you mean by quicker switches?
you can have have two pumps: openaps add bluefish medtronic openaps add redfish medtronic would add two pumps to same instance
Sarah Hodson Grady
@shgrady
Jun 22 2016 01:48
like week 1, I am doing a lot of biking so I want the small one with the clip my512, week 2 I have an interview, want to wear the pink 722 for more insulin
Ben West
@bewest
Jun 22 2016 01:49
you can simply change the serial number then: openaps use pump config --serial 123456
Paul Martin
@Spazholio
Jun 22 2016 01:49
@bewest So, I’m erroring out when trying that command you suggested. Does it need anything other than: mmeowlink-send.py --serial [MYSERIAL] --init tweak ReadHistoryData --save --page 0?
Ben West
@bewest
Jun 22 2016 01:49
it'd be interesting to think about sending openaps a signal from Nightscout or pebble or something to do that
Brian Bonner
@itsn0tm3
Jun 22 2016 01:49
@Spazholio yes it needs --port [porthere]
Paul Martin
@Spazholio
Jun 22 2016 01:50
Yep, just realized.
Brian Bonner
@itsn0tm3
Jun 22 2016 01:50
--port /dev/mmeowlink
Sarah Hodson Grady
@shgrady
Jun 22 2016 01:50
right, but if both were in the instance, couldn't I just tuck the one out of CareLink range?
Ben West
@bewest
Jun 22 2016 01:50
it needs --subg_rfspy as well
Brian Bonner
@itsn0tm3
Jun 22 2016 01:50
oh? it runs fine for me without that
Paul Martin
@Spazholio
Jun 22 2016 01:50
Yeah, same.
And how will I know if it encounters the “bad” data (if it exists)? Will the Python just choke on trying to parse it?
Ben West
@bewest
Jun 22 2016 01:52
@shgrady openaps needs to know via the serial number which one to talk to
it'd be possible to come up with an openaps feature to switch them in the software somehow
but it sounds like the easiest way for now is to log in and do openaps use pump --serial 123456 when you switch pumps at beginning of week 2
it just needs to be done once, when you switch the pump
@Spazholio this workflow separates the fetching of the data from the parsing of the data
Sarah Hodson Grady
@shgrady
Jun 22 2016 01:54
yep, sounds good, just when I have been do the manual commands to get a feel for the reports the 512 for instance has an error in the read status
*doing
Ben West
@bewest
Jun 22 2016 01:54
that's true, 512 doesn't have as many features
including "status"
Paul Martin
@Spazholio
Jun 22 2016 01:55
Ah, ok. Well, the commands are running. A couple I’ve had to re-try, but they’re all returning data. Am I looking for anything specific? Something to indicate what might be causing the issue?
Ben West
@bewest
Jun 22 2016 01:56
here's an example: bewest/decoding-carelink#126
idea is that the data should be fetched fine, it's always 1024 bytes
Sarah Hodson Grady
@shgrady
Jun 22 2016 01:57
thats why I was thinking I needed them as separate instances, not just swaps, but maybe I am getting ahead of myself again, earlier, since my pump is full of saline, was trying to figure out the bolus command for a good flush.
Ben West
@bewest
Jun 22 2016 01:57
mm-decode-history-page.py should be able to decode it without any errors
Sarah Hodson Grady
@shgrady
Jun 22 2016 01:57
separate devices maybe a better word
Ben West
@bewest
Jun 22 2016 01:57
you could use separate devices, sure
Paul Martin
@Spazholio
Jun 22 2016 01:58
So save each page to a file, then run mm-decode-history-page.py against those files to see if anythings screwy?
Sarah Hodson Grady
@shgrady
Jun 22 2016 01:58
k, thanks.
Ben West
@bewest
Jun 22 2016 01:58
yeah, it should have a name for every record as well as a legit timestamp
Paul Martin
@Spazholio
Jun 22 2016 01:58
k
Well, the first two pages have errors.
Ben West
@bewest
Jun 22 2016 02:00
do you use medtronic cgm at all?
Paul Martin
@Spazholio
Jun 22 2016 02:00
No, never.
Ben West
@bewest
Jun 22 2016 02:00
ok, paste it into gist/issue
the output of that tool is markdown, natively, it's probably quite long, so not here
make sure to specify your --model
Paul Martin
@Spazholio
Jun 22 2016 02:01
The output of the mmeowlink-send.py or the other one?
Ben West
@bewest
Jun 22 2016 02:01
-send just fetches data, I don't anticipate any errors there
Paul Martin
@Spazholio
Jun 22 2016 02:02
Oh, I wasn’t specifying —model.
No changes though.
Ok, so to be clear - I’ll gist (is that a verb?) whatever - do you want the output of the mm-decode-history-page.py or the mmeowlink-send.py?
Ben West
@bewest
Jun 22 2016 02:04
decoding
Paul Martin
@Spazholio
Jun 22 2016 02:04
Gotcha. Standby.
Ben West
@bewest
Jun 22 2016 02:04
was there a problem with -send?
Paul Martin
@Spazholio
Jun 22 2016 02:04
No. But I saved them to a file, then ran mm-decode-history-page.py against that file.
Was that…incorrect?
Ben West
@bewest
Jun 22 2016 02:05
no, using -send.py --save should create a ReadHistoryData-page-0.data file
Paul Martin
@Spazholio
Jun 22 2016 02:05
Oh hell.
Brian Bonner
@itsn0tm3
Jun 22 2016 02:05
(nods) I got that
Ben West
@bewest
Jun 22 2016 02:05
that file can be fed to mm-decode-history-page.py --model
Paul Martin
@Spazholio
Jun 22 2016 02:05
It did. I just didn’t check the dir afterward. =)
Ben West
@bewest
Jun 22 2016 02:06
the output of both of these tools is markdown
Brian Bonner
@itsn0tm3
Jun 22 2016 02:06
interestingly 2 pages have errors at the top about catching a null
Paul Martin
@Spazholio
Jun 22 2016 02:06
Ok, NOW I’m getting some stuff.
Brian Bonner
@itsn0tm3
Jun 22 2016 02:06
and 2 pages have unknown at the bottom
Ben West
@bewest
Jun 22 2016 02:06
ah, might not be an error if you add it up
Brian Bonner
@itsn0tm3
Jun 22 2016 02:07
any way to get ALL the pages at once?
(how many does it contain anyway?)
Paul Martin
@Spazholio
Jun 22 2016 02:07
Ok, the output of each mm-decode-history-page.py —model 722 is huge. Can I grep for a particular keyword, like ERROR, that might help me go through it faster? Or will I get a Python error if it encounters issues?
Ben West
@bewest
Jun 22 2016 02:07
for page in $(seq 0 32); do .... --page $page done
you can | grep RECORD for a list of one liners
Paul Martin
@Spazholio
Jun 22 2016 02:11
#### RECORD 134 Base unknown head[2], body[0] op[0x00]
Is that something interesting?
Everything else has things I recognize.
Ben West
@bewest
Jun 22 2016 02:11
yeah, can't be right
the one before that is wrong also
Brian Bonner
@itsn0tm3
Jun 22 2016 02:11
I have some of those too
Ben West
@bewest
Jun 22 2016 02:11
the one before it is the cause/problem
Brian Bonner
@itsn0tm3
Jun 22 2016 02:12
its always the last record on the page it seems
Paul Martin
@Spazholio
Jun 22 2016 02:12
#### RECORD 134 TempBasal 2016-06-21T10:14:30 head[2], body[1] op[0x33]
#### RECORD 135 TempBasalDuration 2016-06-21T10:14:30 head[2], body[0] op[0x16]
#### RECORD 136 Base unknown head[2], body[0] op[0x8f]
@itsn0tm3 Mine too.
Brian Bonner
@itsn0tm3
Jun 22 2016 02:12
#### RECORD 122 TempBasalDuration 2016-06-21T04:49:48 head[2], body[0] op[0x16]
#### RECORD 123 TempBasal 2016-06-21T05:08:06 head[2], body[1] op[0x33]
#### RECORD 124 Base unknown head[2], body[0] op[0x00]
@Spazholio is yours by any chance on page 2 too?
Ben West
@bewest
Jun 22 2016 02:13
the head is 0x00 0x00?
Paul Martin
@Spazholio
Jun 22 2016 02:13
@itsn0tm3 And page 1. But yes.
@bewest I don’t know how to answer your question.
Brian Bonner
@itsn0tm3
Jun 22 2016 02:14
#### RECORD 124 Base unknown head[2], body[0] op[0x00]

    op hex (2)
    0000   0x00 0x58                                  .X
    decimal
              0   88
    datetime (unknown)

    body (0)

`end ReadHistoryData-page-2.data: 125 records`
Ben West
@bewest
Jun 22 2016 02:14
it's not a property of page number
ah, and that's the end
Paul Martin
@Spazholio
Jun 22 2016 02:14
#### RECORD 135 TempBasalDuration 2016-06-21T10:14:30 head[2], body[0] op[0x16]
###### DECODED
```python
{'duration (min)': 30}
op hex (2)
0000   0x16 0x01                                  ..
decimal
         22    1
datetime (2016-06-21T10:14:30)
0000   0x5e 0x8e 0x0a 0x55 0x10                   ^..U.
body (0)

RECORD 136 Base unknown head[2], body[0] op[0x8f]

op hex (2)
0000   0x8f 0x29                                  .)
decimal
        143   41
datetime (unknown)

body (0)
end ReadHistoryData-page-2.data: 137 records
```
…dammit. Stupid Gitter.
Ben West
@bewest
Jun 22 2016 02:14
it's already markdown formatted
Paul Martin
@Spazholio
Jun 22 2016 02:15
Well Gitter should know what I meant to do, not what I told it to do. Yeesh.
Brian Bonner
@itsn0tm3
Jun 22 2016 02:15
hehe
Paul Martin
@Spazholio
Jun 22 2016 02:15
I can gist those sections if it will help.
Ben West
@bewest
Jun 22 2016 02:16
yeah
what's happening here, is it's actually kind of hard to tell we're at the end of the page
So it’s possible the “page” is cutting off the the middle of an entry or something like that?
Ben West
@bewest
Jun 22 2016 02:18
the decoder is rather greedy, but it can be easily adjusted so that if there's not enough room for a whole record near an expected page boundary it'll just look for crc
no, it's not middle, that looks like the crc
I need the whole output
or alternatively the .data file
if you name the output file ending in .md it'll get highlighted nicely
it's already formatted markdown
Paul Martin
@Spazholio
Jun 22 2016 02:20
How can I get you the entire file? That’ll ensure completeness. I can’t seem to upload a file to a gist...
Ben West
@bewest
Jun 22 2016 02:20
you can drag/drop it here I think
or in "edit" mode in gist
or you can git clone the gist and manually add/push the file
Wasn’t sure if I could do that with just any old file.
Brian Bonner
@itsn0tm3
Jun 22 2016 02:24
Hmmmm intersting
Ben West
@bewest
Jun 22 2016 02:24
or you can bewest/decoding-carelink#196
Brian Bonner
@itsn0tm3
Jun 22 2016 02:24
I started from page 32 and worked my way back
Ben West
@bewest
Jun 22 2016 02:24
you can attach things to the issue I believe
Brian Bonner
@itsn0tm3
Jun 22 2016 02:24
and I don't see those unknowns on those pages
Paul Martin
@Spazholio
Jun 22 2016 02:25
Oh, I started at page 0 and went up. =)
Ben West
@bewest
Jun 22 2016 02:25
the amount of data per page will vary dramatically
if you use openaps, there will be many many more records of temp basal, causing faster page rotation
a page only fits 1024 bytes, and 2 of those are for crc
Brian Bonner
@itsn0tm3
Jun 22 2016 02:29
I don't see anything wonky yet other than unknown at the very last record
interestingly it doesn't happen as often as it was happening at the first few pages
Paul Martin
@Spazholio
Jun 22 2016 02:32
Admittedly, I stopped after the first 2, which is where I reported ‘em.
Brian Bonner
@itsn0tm3
Jun 22 2016 02:33
heh mine starting giving me fits when I went to SLC
so I'm wonderin if there was something in there from back then
which was why I'm going down from 32
and just curious too
Paul Martin
@Spazholio
Jun 22 2016 02:34
Oddly enough, my rig seems to be working just fine now. headdesk
Brian Bonner
@itsn0tm3
Jun 22 2016 02:34
hahahahaha same here
a watched pot never errors out does it ;)
Brian Bonner
@itsn0tm3
Jun 22 2016 02:39
welp I'm down to page 15 when I started killing the batteries in the pump
@bewest btw I noticed that these don't have a timestamp: #### RECORD 105 UnabsorbedInsulinBolus unknown head[5], body[0] op[0x5c]
oooh
found a new one
#### RECORD 121 Base unknown head[2], body[0] op[0x00]
that's teh last record for page 15
Paul Martin
@Spazholio
Jun 22 2016 02:44
It’s like the worst treasure hunt ever.
Brian Bonner
@itsn0tm3
Jun 22 2016 02:44
no kidding, still waiting to find the pot of gold
Paul Martin
@Spazholio
Jun 22 2016 02:46
It’s probably buried underneath the used hypodermics and the old tire.
But after that - GOLD, JERRY!
Brian Bonner
@itsn0tm3
Jun 22 2016 02:48
down to page 10 now, I don't think there's no gold to be found here :(
Ben West
@bewest
Jun 22 2016 02:49
hmm
what was the error again?
invalid CRC?
seems like...
Brian Bonner
@itsn0tm3
Jun 22 2016 02:50
@bewest me or @Spazholio ?
Ben West
@bewest
Jun 22 2016 02:52
both: this error, correct? bewest/decoding-carelink#196
Brian Bonner
@itsn0tm3
Jun 22 2016 02:53
yes
the really interesting thing for me is that unknown isn't showing up for a LOT of pages now
seems to be just the first few pages and the last few where it shows up
Paul Martin
@Spazholio
Jun 22 2016 02:54
Yes, that’s the exact error.
Brian Bonner
@itsn0tm3
Jun 22 2016 02:57
alrighty, done with ALL of the pages
no pot of gold :(
did see that Base again tho on page 4
it's the last record too again
#### RECORD 127 Base unknown head[2], body[0] op[0x36]
#### RECORD 127 Base unknown head[2], body[0] op[0x36]

    op hex (2)
    0000   0x36 0xf9                                  6.
    decimal
             54  249
    datetime (unknown)

    body (0)

`end ReadHistoryData-page-4.data: 128 records`
Ben West
@bewest
Jun 22 2016 02:59
yeah, it's the crc :-)
Paul Martin
@Spazholio
Jun 22 2016 02:59
He’s smiling. That’s either really good, or he’s snapped. =)
Brian Bonner
@itsn0tm3
Jun 22 2016 02:59
heh
Ben West
@bewest
Jun 22 2016 02:59
it's actually explained in bewest/decoding-carelink#196
Brian Bonner
@itsn0tm3
Jun 22 2016 03:01
when you download 28 hours worth, does it also download it as pages and the CRC is what throws it?
Ben West
@bewest
Jun 22 2016 03:01
it always downloads whole pages
there's not a way to download half a page
Brian Bonner
@itsn0tm3
Jun 22 2016 03:01
ah ok
Paul Martin
@Spazholio
Jun 22 2016 03:02
So do we know what’s causing the BangInt error? Or how to detect/trap/avoid it?
Ben West
@bewest
Jun 22 2016 03:03
it's not a BangInt error
Paul Martin
@Spazholio
Jun 22 2016 03:03
Well, I was just calling it that since that’s what sticks out in the Python error dump.
Tim Mecklem
@tmecklem
Jun 22 2016 03:04
hello all. My wife has decided we need to look into this for her, so I'm getting up to speed on the requirements and learning about the various projects involved.
Ben West
@bewest
Jun 22 2016 03:05
openaps/decocare#1 @Spazholio , @itsn0tm3
Tim Mecklem
@tmecklem
Jun 22 2016 03:06
We are going to start out with a pi3 and the carelink stick and hopefully graduate to something smaller and easier on the battery (maybe a pi zero and slice of pi?). Anyway, just introducing myself for now.
Ben West
@bewest
Jun 22 2016 03:07
howdy @tmecklem :-)
Paul Martin
@Spazholio
Jun 22 2016 03:07
@bewest So, pull the dev branch of…dexcom-reader?
Ben West
@bewest
Jun 22 2016 03:08
no, that's decocare, and it's moved locations
Brian Bonner
@itsn0tm3
Jun 22 2016 03:08
decocare
Tim Mecklem
@tmecklem
Jun 22 2016 03:08
hey @bewest
Brian Bonner
@itsn0tm3
Jun 22 2016 03:08
welcome @tmecklem !
Ben West
@bewest
Jun 22 2016 03:08
sudo pip install git+git://github.com/openaps/decocare.git@dev
Tim Mecklem
@tmecklem
Jun 22 2016 03:08
hey @itsn0tm3
Paul Martin
@Spazholio
Jun 22 2016 03:08
Oddly enough, that’s what I meant. dexcom-reader would make no sense here. =)
Tim Mecklem
@tmecklem
Jun 22 2016 03:09
I'm a java dev turned ruby dev turned elixir (nerves-project) fan. It'll take me a while to get up to speed on the python code, but I hope to be able to assist here some too.
If nothing else, I have a small arsenal of tiny computers to put to various tasks.
Ben West
@bewest
Jun 22 2016 03:10
:-) tiny computers are fun
maybe good repo name, tcaf
Brian Bonner
@itsn0tm3
Jun 22 2016 03:11
@bewest how can we test the changes? can I run mm-decode again and see if it tries to parse the last record incorrectly again?
or do I need to download the page again?
Tim Mecklem
@tmecklem
Jun 22 2016 03:12
I'm kind of addicted to buying a pi zero whenever I pass Microcenter. I've even started giving them out here and there. I think that makes me a pusher.
Ben West
@bewest
Jun 22 2016 03:12
oh, I was wrong
nevermind
Brian Bonner
@itsn0tm3
Jun 22 2016 03:13
@tmecklem heh thats funny
Paul Martin
@Spazholio
Jun 22 2016 03:14
@bewest Wrong about what? Tiny computers ARE fun!
Garrett Webb
@garetis
Jun 22 2016 03:35
This happened, I'm not sure what it means... some of the same commands didn't help. I ended up setting the date manually, and it seems to workaround the issue for now until I can figure out how to get the system to do it itself. Either way, looks like I got everything I need to open loop tonight...
pi@raspberrypi:~/openaps2 $ sudo apt-get install ntpdate
Reading package lists... Done
Building dependency tree
Reading state information... Done
The following extra packages will be installed:
  lockfile-progs
The following NEW packages will be installed:
  lockfile-progs ntpdate
0 upgraded, 2 newly installed, 0 to remove and 0 not upgraded.
Need to get 79.4 kB of archives.
After this operation, 203 kB of additional disk space will be used.
Do you want to continue? [Y/n] y
Get:1 http://mirrordirector.raspbian.org/raspbian/ jessie/main ntpdate armhf 1:4.2.6.p5+dfsg-7+deb8u1 [68.8 kB]
Get:2 http://mirrordirector.raspbian.org/raspbian/ jessie/main lockfile-progs armhf 0.1.17 [10.6 kB]
Fetched 79.4 kB in 2s (26.8 kB/s)
Selecting previously unselected package ntpdate.
(Reading database ... 121098 files and directories currently installed.)
Preparing to unpack .../ntpdate_1%3a4.2.6.p5+dfsg-7+deb8u1_armhf.deb ...
Unpacking ntpdate (1:4.2.6.p5+dfsg-7+deb8u1) ...
Selecting previously unselected package lockfile-progs.
Preparing to unpack .../lockfile-progs_0.1.17_armhf.deb ...
Unpacking lockfile-progs (0.1.17) ...
Processing triggers for man-db (2.7.0.2-5) ...
Setting up ntpdate (1:4.2.6.p5+dfsg-7+deb8u1) ...
Setting up lockfile-progs (0.1.17) ...
Brian Bonner
@itsn0tm3
Jun 22 2016 03:36
looks OK
one thing you will need to remember is that you need to stop ntpd first before using ntpdate
Garrett Webb
@garetis
Jun 22 2016 03:42
Yeah, I'm not sure what's going on. I originally set the date with raspi-config, so wherever it is getting the date from is off by 25 hours or so, but thanks for the help. When I figure out a reliable fix I'll drop in a PR about it
Brian Bonner
@itsn0tm3
Jun 22 2016 03:43
25 hours?
did you also set the timezone?
(in the internationalization options)
Garrett Webb
@garetis
Jun 22 2016 03:44
Yeah. It was reporting the date as yesterday and off by +/- an hour. It could've been 23 hours
Brian Bonner
@itsn0tm3
Jun 22 2016 03:44
optionally you can just leave ntpd running overnight ( just reboot it then leave it running )
and it should in theory fix itself by morning
Garrett Webb
@garetis
Jun 22 2016 03:47
I'm going to need to reboot to take the RPi next to my bed for tonight, so we'll see what happens when that happens. Til then I may stay up a little late to keep the open loop running
What are the treatments in openaps upload-recent-treatments? I'm still unable to upload to NS
I get no treatments to upload when I run that
iananders
@iananders
Jun 22 2016 04:19
trying to invoke my suggested.json report and it appears i have a timing mismatch, but don't quite understand what's going on "BG data is too old, or clock set incorrectly Tue Jun 21 2016 21:09:14 GMT+0000 (UTC) vs Wed Jun 22 2016 04:17:05 GMT+0000 (UTC)"
Garrett Webb
@garetis
Jun 22 2016 04:20
I was dealing with that earlier today
what happens when you do date?
Wed Jun 22...?
iananders
@iananders
Jun 22 2016 04:22
how do i do that? (apologize for my ignorance - and also seeing you were just talking about this same thing, or it appears that way)
Garrett Webb
@garetis
Jun 22 2016 04:23
just type "date" in the command line
iananders
@iananders
Jun 22 2016 04:23
yeah weird - i get June 22
Garrett Webb
@garetis
Jun 22 2016 04:24
do you remember going through sudo raspi-config and setting up the date?
iananders
@iananders
Jun 22 2016 04:25
i don't think i did - i am following pretty closely to the docs and that seemed like an optional step - thats my bad
can i do that now?
Garrett Webb
@garetis
Jun 22 2016 04:26
yeah, go ahead and do it now
iananders
@iananders
Jun 22 2016 04:26
oh yeah i did this - but didn't change anything
Garrett Webb
@garetis
Jun 22 2016 04:28
ok, let me find something real quick
try putting this in the command line and let me know what happens
date ; sudo service ntp stop ; sudo ntpd -q -g ; sudo service ntp start ; date
iananders
@iananders
Jun 22 2016 04:29
Tue 21 Jun 21:29:18 PDT 2016
ntpd: time slew +0.026520s
Tue 21 Jun 21:29:27 PDT 2016
Garrett Webb
@garetis
Jun 22 2016 04:30
so I think it's up to date now? Try your report again
iananders
@iananders
Jun 22 2016 04:32
still a mismatch, just closer now
BG data is too old, or clock set incorrectly Tue Jun 21 2016 21:09:14 GMT-0700 (PDT) vs Tue Jun 21 2016 21:31:22 GMT-0700 (PDT)
is this comparing the pi time with the dexcom time? or the pi with pump time?
Garrett Webb
@garetis
Jun 22 2016 04:33
My understanding is that it's comparing the system time (which I think we have made up-to-date) to the CGM/NS time.
So, how are you doing CGM?
And do you have NS up and running?
iananders
@iananders
Jun 22 2016 04:33
got it - dexcom g4 share via usb to pi
i do
Garrett Webb
@garetis
Jun 22 2016 04:34
what happens when you do openaps use cgm DescribeClocks
iananders
@iananders
Jun 22 2016 04:35
"display": {
"clock": "2016-06-21T21:34:48",
"offset": 2539968.0
},
"rtc": {
"epoch": "2010-07-19T11:46:44"
},
"system": {
"utc": "2016-05-23T12:02:00",
"offset": 184464916.0
Garrett Webb
@garetis
Jun 22 2016 04:37
What about NS, do you have that up and running?
iananders
@iananders
Jun 22 2016 04:39
i did - but it looks like my phone de-paired and now i reached max login attempts somehow on the share app - i know comedy of errors)
Garrett Webb
@garetis
Jun 22 2016 04:40
I'm not sure about NS because mine is up, but not running, but I know there's a setting in there where you can go in and set the time. You might want to check that, and see if it is close to actual time, or not
iananders
@iananders
Jun 22 2016 04:41
my share app is going crazy now - cycling through error messages that don't make sense
ok thanks
so even though the dexcom is hard connected, it's still comparing the system time to the NS time
Garrett Webb
@garetis
Jun 22 2016 04:42
I got my dexcom today and have only been using it for 6 hours, and I have Android, so I'm a little of an outsider when it comes to that... Yeah, my dexcom is hard connected and I was having that error until I did sudo date -s "21 JUN 2016 22:15:00" to set the time manually -- that's what I had to resort to when those other codes you did reset you time -- they didn't work for me
iananders
@iananders
Jun 22 2016 04:45
thanks so much for the info and the hand holding
seems like my NS time zone was incorrect
Garrett Webb
@garetis
Jun 22 2016 04:45
Sweeet
That's about all I got for troubleshooting, if that doesn't fix it then we'll need to consult the higher-ups
iananders
@iananders
Jun 22 2016 04:46
strangely still getting the same error
cheers for your help, ill fly it up the flagpole
Garrett Webb
@garetis
Jun 22 2016 04:46
It might take time to update, and it might take a reboot
just spit-balling
iananders
@iananders
Jun 22 2016 04:47
that would make sense
is it even possible to reboot the nightscout? rebooting my pi now
Scott Leibrand
@scottleibrand
Jun 22 2016 05:00
Have you identified which file has the wrong timestamp?
Garrett Webb
@garetis
Jun 22 2016 05:02
Not sure about NS, that was all pretty foreign for me
iananders
@iananders
Jun 22 2016 05:02
seems like the NS timestamp is frozen now, yet the site seems to be running and getting fresh data
Sarah Hodson Grady
@shgrady
Jun 22 2016 05:26
@iananders not exactly, but I have severely broken my Nightscout before, amazing what a redeploy fixes, still don't understand that one.
iananders
@iananders
Jun 22 2016 05:36
and by redeploy you mean completely from github, not just stopping and restarting what i already have?
C-Ville
@C-Ville
Jun 22 2016 12:29
I do have a question regarding "neutral temp basal", I thought neutral temp basal is supposed to be the basal rate that is currently programmed in the pump. My highest programmed basal rate is 0.8 u/h, nevertheless the OpenAPS sets "neutral temp basal" of 1 u/h, even when it seems like I am well in target and the temp basal should be the (default) programmed 0.8. Do I have a wrong understanding of neutral temp basal?
Paul Martin
@Spazholio
Jun 22 2016 12:43
@scottleibrand I thought I submitted a patch to that file to indicate which timestamp was which. Is that only in dev?
Sarah Hodson Grady
@shgrady
Jun 22 2016 13:15
@iananders yeah, the first three times I tried to use the new CGM Careportal 5.4, my site froze and displayed nothing new. Going into Azure, disconnecting, then redeploying got it running again all 3 times, like I said, not a reboot, and weird, but as a last resort.
@iananders also, not related to your current problem, which appears to be known, but you had asked if NS could be rebooted.
Sarah Hodson Grady
@shgrady
Jun 22 2016 13:47
I have had NS for over a year, in which time the Azure portal has changed, I did not know their was a stop/start, that's why I did disconnect and redeploy.
Scott Leibrand
@scottleibrand
Jun 22 2016 13:54
@C-Ville autosens (and meal assist, pre-AMA) adjust what is considered neutral.
@Spazholio likely. If you think it'd be useful to release that to master before we get AMA merged to dev, might be worth opening / commenting on the dev to master PR to that effect.
C-Ville
@C-Ville
Jun 22 2016 13:55
@scottleibrand I think I have neither autosens nor meal assist enabled
Chris Oattes
@cjo20
Jun 22 2016 13:57
what does "AMA" stand for?
advanced meal assist?
Scott Leibrand
@scottleibrand
Jun 22 2016 13:57
Advanced meal assist.
What does profile.json say your basal is?
and what does the suggested.json look like?
C-Ville
@C-Ville
Jun 22 2016 14:33
@scottleibrand I feel so stupid now and sorry for bugging you apparently I used an old profile.json that had 1u/h in there. after reinvoking that report it is now updated. I might have omitted that report at one point in time in my loop alias. I will check once I am back and stable in range what happens. Looks like it is working now
Colin Lennon
@colinlennon
Jun 22 2016 15:02
Hi everyone. I'm working on an idea I've had to allow offline looping when using xDrip (I'm a Dexcom G4 user but don't have a receiver)
I know of the option to connect a Wixel directly to a Pi/Edison, but I like using the xDrip android app - alerts and a nice UI!
How about a simple Android app which listens to the local broadcasts which xDrip can send, then echoes these out to the Pi/Edison as an HTTP PUT request
On the Pi/Edison side, a simple Python web server which receives the request, parses the parameters and creates or appends to a glucose.json file, which OpenAPS can read (if the same JSON format as Nightscout is used)
Android phone and Pi/Edison need to be BLE tethered or connected to same Wifi network/personal hotspot, but eliminates the need for an internet connection to pull glucose values from Nightscout
I'm only just starting on this now, but interested to hear any thoughts anyone else has on this topic
C-Ville
@C-Ville
Jun 22 2016 15:12
@colinlennon looking at your description it seems that you would not be able to monitor the basal rate changes the loop is doing, am I correct?
Colin Lennon
@colinlennon
Jun 22 2016 15:15
@C-Ville Yeah, that's correct. I'm not sure if there's already an offline Pebble solution
Need to look into that
C-Ville
@C-Ville
Jun 22 2016 15:34
@colinlennon pushing your idea further I am wondering if you could have the edison with the wixel work as an xdrip box or a wifi wixel, that way you would not need to carry around an extra wixel
Colin Lennon
@colinlennon
Jun 22 2016 15:36
@C-Ville That is possible, check out https://github.com/oschumac/python-uart-pi-xbridge
although I haven't tried that yet personally, so I'm not sure if that works as a wifi wixel yet
Jason Wittmer
@jbwittmer
Jun 22 2016 17:04
So I accomplished second night in a row looping last night. Technically things went well with the equipment, but the loop never got sugars down to desired range (set 135-145 currently to be conservative). In fairness, horrible meal at dinner - pizza. Loop continued to give extra basal insulin all night but sugar just didn't drop. Really stable around 210 the entire night. Interestingly the prior night with a much lower carb meal got to about 170 and stable, but also didn't get down to range. His first night on loop (about a week ago) the system did very well at keeping in range. For additional info, son (11) has been running a little high during the day as well despite correction, and my suspicion is that he needs higher basal insulin rate. We're about "due" for a increase based on his recent growth spurts. Am I correct that that would explain the behavior seen? If his basal (or correction factor) isn't set high enough the algorithm will never be able to get him down to desired range? I plan to increase basal rates on pump and retry tonight but wanted to make sure I'm not missing anything. Hope to be able to submit my form to @danamlewis after tonight and get the count to (n=1)*86
Scott Leibrand
@scottleibrand
Jun 22 2016 17:23
@jbwittmer it sounds like your evaluation is probably correct. were you bumping up against maxSafeBasal, or was it able to give as much as it wanted, and just thought eventualBG would be in range (but it wasn't)? FWIW, if you had autosens configured, it might help, so you might consider enabling that once you get your basals tuned properly.
Jason Wittmer
@jbwittmer
Jun 22 2016 17:25
This message was deleted
@scottleibrand No I was well below his maxSafeBasal. It just thought eventualBG would be in range but it never got there. I've looked at the AMA and autosens options but wanted to get the loop working solidly before tackling that. Thanks for your thoughts.
Scott Leibrand
@scottleibrand
Jun 22 2016 17:29
yeah, that's definitely the best approach. sounds like you're entirely on the right track.
Dana Lewis
@danamlewis
Jun 22 2016 17:32
@jbwittmer yes, if basals are way off, the loop can work really hard but still may not be able to get down in range depending on all the variables.
Paul Martin
@Spazholio
Jun 22 2016 17:33
@scottleibrand Is there an easy way for me to see what auto-sens reported for the last 24 h? Possibly longer? I want to see how it trends throughout the day/week.
Dana Lewis
@danamlewis
Jun 22 2016 17:33
@Spazholio nope. need to find a way to grep the logs and stick it in a file somewhere
it’s on my wishlist. along with extra hours in the day to actually do it, and many other things ;)
Paul Martin
@Spazholio
Jun 22 2016 17:34
Does it get uploaded to NS at all? If not, can we?
Jason Wittmer
@jbwittmer
Jun 22 2016 17:34
@scottleibrand By the way, wanted to say thanks for all that you and @danamlewis have done to put the resources out there for people like me to work with. It really feels good to be able to do something (for once) to improve control for my son in a very meaningful way, and to likely improve his quality of life and sleep (and for that matter my sleep) as well. I'll never be able to say thank you enough ways.
Dana Lewis
@danamlewis
Jun 22 2016 17:35
@jbwittmer :) <3 and @bewest and many others ^ :)
Jason Wittmer
@jbwittmer
Jun 22 2016 17:36
@danamlewis Yes absolutely, many others to thank and also to all of the people that have answered my questions here. Don't think any of you are off the hook, however, as I'm sure I'll have more as I try to refine things.
Dana Lewis
@danamlewis
Jun 22 2016 17:36
:) :D
Scott Leibrand
@scottleibrand
Jun 22 2016 17:37
@Spazholio the ISF should be in the reason field uploaded to NS.
Paul Martin
@Spazholio
Jun 22 2016 17:38
Right, but there's no report or way to just pull that particular number so I can see how it changes hour to hour as it stands, right? I know I can pull the data out, just don't want to reinvent the wheel.
Dana Lewis
@danamlewis
Jun 22 2016 17:38
correct, no report. you should build one, I’ll help you test it :)
Brooke Armstrong Pazoles
@bpazoles
Jun 22 2016 17:50
@jbwittmer I was having similar overnight issues and found that decreasing my ISF a bit (changed it from 60 to 55) helped a lot. sounds like you've got a good plan with adjusting basals but just wanted to share something else that may help. We also increased my max safe basal and that seemed to help as well
Scott Leibrand
@scottleibrand
Jun 22 2016 17:51
:+1: to trying out a lower ISF if you want the loop to be a bit more aggressive.
Brooke Armstrong Pazoles
@bpazoles
Jun 22 2016 17:52
I was still in range, just on the higher end instead of middle or lower end. Tweaking ISF fixed that right up.
Jason Wittmer
@jbwittmer
Jun 22 2016 17:52
@bpazoles Thanks. Changing ISF was going to be my next step if basal rate adjustments didn't work. I think that both are likely necessary based on what I'm seeing during the day, but trying to change one variable at a time. I think his overall needs have just changed fairly abruptly. He seems to do this from time to time as he grows.
Brooke Armstrong Pazoles
@bpazoles
Jun 22 2016 17:53
That sounds like a solid plan to me. Definitely agree about changing one variable (as much as you can with T1D!!) at a time.
Growth hormones are no joke - rapid and aggressive !
Jason Wittmer
@jbwittmer
Jun 22 2016 17:54
@bpazoles :)
Jarred Yaw
@jyaw
Jun 22 2016 18:12
@Spazholio I'd test this report as well, would love to know adjusted ISF over time
Paul Martin
@Spazholio
Jun 22 2016 18:36
@scottleibrand When running the "Treatments" report, I see very few (if any) items under the "Blood Glucose" entry. Might it be a good idea to add a "Sensor Glucose" column to that report? That way we can get a "feel" for how things were going when treatments are applied? I know the Notes field shows some data, but I only seem to have ~4 entries that even have anything in the "Notes" field per day. I'd like to see something that tells me why it made each decision that it did. Am I missing something?
Scott Leibrand
@scottleibrand
Jun 22 2016 18:40
reports are completely out of scope for me. I would ping @jasoncalabrese and @MilosKozak in nightscout/public
Dragan Maksimovic
@dm61
Jun 22 2016 18:50
@Spazholio , "Treatments" report includes only what's on the pump, including any fingerstick bg values - if you have a meter that transmits to the pump, or if you enter bg manually on the pump. This all gets uploaded to NS where you already have your CGM included by some other means. So, I do not think it would add any further value to include sensor glucose in the "treatments."
Paul Martin
@Spazholio
Jun 22 2016 18:53
Hmmm, ok. The way I was looking at it was, yes it's all information from the pump, but being able to correlate it with what it's acting on (sensor data, notes field, etc) and being able to see that all on one screen might prove useful. Is there a report that just lists data from NS? I don't want the visualizations all the time. For example, I'd love to see the notes field for each entry for a full 24 hours. It doesn't sound like it would be too difficult to query the DB if that's something I need to do, I was just curious if such a thing existed.
Dragan Maksimovic
@dm61
Jun 22 2016 19:00
You could just go to NS Reports and see everything in various formats, including a complete list of treatments, etc. You could also get anything you like back via NS API. One can think of all kinds of stuff that could be done further. For example, I use an NS Announcement to activate (run closed loop) or deactivate the loop (run everything except enact).
Garrett Webb
@garetis
Jun 22 2016 19:54
Where could I get more information on the fields of enact/suggested.json? Specifically, "Boluses", SnoozeBG", and "BGI"?
Paul Martin
@Spazholio
Jun 22 2016 20:10
@dm61 Which report are you referring to? Or did you just mean in general?
Bill Stackpole
@billstackpole
Jun 22 2016 20:39
@garetis - is there an analogous command for the Enlight / Medtronic Connect system to the openaps use cgm DescribeClocks command you noted above? As an enlight user I have no cgm device specified (no such device noted in output of openaps use -h). Is this command intended to poll the clock in the DexCom?
Didie
@elodaille01_twitter
Jun 22 2016 20:39
hello

Check that you can communicate with your Dexcom receiver

Now let's try communicating with the Dexcom receiver.

Hint: Your Dexcom should be nearly fully charged before plugging it in to your Raspberry Pi. If, when you plug in your Dexcom, it causes your WiFi dongle to stop blinking, or if the charging icon on the Dexcom keeps cycling on and off, that is a sign that it is drawing too much power and needs to be charged.

Workaround: If you continue to have problems, try increasing the mA output to the USB ports, you can do this by running the following command $ sudo bash -c "echo -e \"#Enable Max USB power\nmax_usb_current=1\" >> /boot/config.txt".

Reboot via $ sudo shutdown -r now to pick up the changes.
$ openaps use <my_dexcom_name> iter_glucose 1

should return something like:

[
{
"trend_arrow": "FLAT",
"system_time": "2015-08-23T21:45:29",
"display_time": "2015-08-23T13:46:21",
"glucose": 137
}
]

it does not work
someone can help me, please ?
Sarah Hodson Grady
@shgrady
Jun 22 2016 20:42
@elodaille01_twitter what name did you use when you added the cgm as a device?
Didie
@elodaille01_twitter
Jun 22 2016 20:43
LucileDexcom
Sarah Hodson Grady
@shgrady
Jun 22 2016 20:43
and what happens when you enter openaps use LucileDexcom iter_glucose 1
Didie
@elodaille01_twitter
Jun 22 2016 20:45
AttributeError : "None type" object has no attribute "iter_records"
Sarah Hodson Grady
@shgrady
Jun 22 2016 20:46
looking for the command to see devices, I don't think you added the device correctly
Didie
@elodaille01_twitter
Jun 22 2016 20:47
My Dexcom receiver is plugged into your Raspberry Pi
Sarah Hodson Grady
@shgrady
Jun 22 2016 20:47
$ openaps device show
Didie
@elodaille01_twitter
Jun 22 2016 20:47
Yes, with "openaps device show" is good
Sarah Hodson Grady
@shgrady
Jun 22 2016 20:49
right, but its saying LucileDexcom cannot iter_records when we want it to iter_glucose
Didie
@elodaille01_twitter
Jun 22 2016 20:50
yes, i don't understand
too bad, i'll try again tomorrow
Sarah Hodson Grady
@shgrady
Jun 22 2016 20:54
if you accidentally added your CGM as a pump, you might get attribute errors, I think, sorry I don't know, just trying to get enough info out of you so somebody smart can help
Didie
@elodaille01_twitter
Jun 22 2016 20:55
Openaps device show :
dexcom://LucileDexcom
medtronic://LucilePump
that's ok
Sarah Hodson Grady
@shgrady
Jun 22 2016 20:56
try $openaps use LucileDexcom -h
do you get a list of commands or another error?
Didie
@elodaille01_twitter
Jun 22 2016 20:59
thank you but i don't see why
see you tomorrow, perhaps
bye
my Dexcom plugged or not, i have the same error
Garrett Webb
@garetis
Jun 22 2016 21:02
@billstackpole So as you get through the docs you'll get to a point where it talks about the glucose device, that converts the medtronic cgm data into a format that is usable. The pump I'm using is the 712, so it doesn't work with enlite and I had to jump on with dexcom
Didie
@elodaille01_twitter
Jun 22 2016 21:02
still, it will start charging
Bill Stackpole
@billstackpole
Jun 22 2016 21:13
So @garetis - the command is dex-specific? (723 here btw)
Understand the glucosecleaner alias. <smile>
Jarred Yaw
@jyaw
Jun 22 2016 21:18
@Spazholio couldn't you just correlate between entries.json and treatments.json? That seems pretty easy to do given all the timestamps are alligned
Paul Martin
@Spazholio
Jun 22 2016 21:24
@jyaw Certainly possible. I’m considering writing an adhoc report for myself (unsure how to integrate it into NS properly so far) just to get a feel for it. And again, if something like it exists already, I’d hate to reinvent the wheel. =)
Scott Leibrand
@scottleibrand
Jun 22 2016 21:24
@billstackpole look at clockset.sh to see the two ways to read/set the clocks on MDT and Dex
Milos Kozak
@MilosKozak
Jun 22 2016 21:28
@Spazholio reports shows what you put there. if you want bg there include it in treatments
Dragan Maksimovic
@dm61
Jun 22 2016 22:01
@Spazholio sorry for the confusion, too many "reports" :smile: ; by "NS Reports" I meant the tab "Reports" under "Settings" on the NS site. Just as many of us do with NS CGM data, you can download and correlate anything you like from the db associated with NS.
Paul Martin
@Spazholio
Jun 22 2016 22:05
@dm61 Yep, that’s what I meant too. But I can probably make my own pretty easily. I just didn’t want to if something else already existed. My main idea now is to get an ISF report.
@MilosKozak Can you elaborate what you mean?
Milos Kozak
@MilosKozak
Jun 22 2016 22:10
if you want to see bg in treatments you must fill it when you create the teatment
Paul Martin
@Spazholio
Jun 22 2016 22:23
Meaning, my enact/suggested.json?
Milos Kozak
@MilosKozak
Jun 22 2016 22:24
for oref0 like this. where you generate temp basal treatment for upload
Paul Martin
@Spazholio
Jun 22 2016 22:26
Ok, I think that’s my enact/suggested.json file. And that has a bg field in there already.
Ben West
@bewest
Jun 22 2016 22:40
pump-history -> treatments
if you use the wizard on the pump, it should include the bgs when you upload treatments from the pump's history
enact/suggested.json has nothing to do with what goes into NS actually except for the openaps status pill
most of the time, oref0-suggest/oref0.json AKA enact/suggested.json is the result of determine-basal and will get passed to set_temp_basal
keep in mind that none of that is actually pump history
set_temp_basal will create a new record in the pump's history if it's successful, but it doesn't output any historical records from the pump. the output from set_temp_basal is simply the result of the command that was requested, not the historical record
Ben West
@bewest
Jun 22 2016 22:45
the stuff going into treatments is the output of iter_pump_history or similar and reformatted by ns shell format-recent-history-treatments monitor/pump-history-zoned.json model.json
it's uploaded via ns shell upload-non-empty-treatments - where - is the filename of the previous output or being piped in via stdin
Paul Martin
@Spazholio
Jun 22 2016 22:49
Well, what I want - and I’ll check my reports - is to have a page that has my ISF from auto-sens next to my sensor glucose. And maybe some treatment data for good measure. =)
Ben West
@bewest
Jun 22 2016 22:50
@Spazholio regarding bewest/decoding-carelink#196 are you still having this? I was wrong yesterday about finding/solving it
does it happen if you use the carelink stick?
what version of mmeowlink and subg_rfspy?
what's happening in that traceback is the CRC was actually missing
I'm not sure that certain pages will aggravate it
Paul Martin
@Spazholio
Jun 22 2016 22:52
@bewest I haven’t seen the issue resurface, but I haven’t had cause to. I can look again though.
Ben West
@bewest
Jun 22 2016 22:52
well, it looks like throwing error there is right thing to do
in order to do something better/fix it I need to know if it happens with carelink also and what versions of mmeowlink/subg_rfspy
Paul Martin
@Spazholio
Jun 22 2016 22:53
0.8.5 and dev version, respectively.
jordan-berger
@jordan-berger
Jun 22 2016 22:54
i am working on building an OpenAPS and am running into an issue. I need some help troubleshooting the connection between OpenAPS and Nightscout.
Ben West
@bewest
Jun 22 2016 22:54
there's a version of docs still under dev meant to help with that, @jordan-berger http://openaps.readthedocs.io/en/dev/docs/walkthrough/phase-1/index.html
jordan-berger
@jordan-berger
Jun 22 2016 22:57
i have been through that and am stuck. I have nightscout reading CGM data. a Raspberry PI with OpenAPS installed. OpenAPS communicates with the pump but not with Nightscout.
Ben West
@bewest
Jun 22 2016 22:57
there's some additional discussion you can comment on here openaps/docs#129
this one is updated, you have an ns device?
created from nightscout autoconfigure-device-crud?
with the device created this way, openaps use ns shell preflight should print true
jordan-berger
@jordan-berger
Jun 22 2016 23:05

pi@raspberrypi:~$ nightscout autoconfigure-device-crud
nightscout autoconfigure-device-crud <NIGHTSCOUT_HOST> <API_SECRET>

sets up:
openaps use ns shell get entries.json 'count=10'
openaps use ns shell upload treatments.json recently/combined-treatments.json

Ben West
@bewest
Jun 22 2016 23:08
what happens with openaps use ns shell preflight?
jordan-berger
@jordan-berger
Jun 22 2016 23:08
returns true
Ben West
@bewest
Jun 22 2016 23:08
ah great, it can talk to nightscout
jordan-berger
@jordan-berger
Jun 22 2016 23:09
Nightscout indicates OpenAPS Unknown
i have the pump and openaps plugins enabled
Ben West
@bewest
Jun 22 2016 23:12
do you have the pill already prepared via openaps use ns shell status somehow?
it needs bunch of files: monitor/clock.json oref0-monitor/iob.json oref0-predict/oref0.json oref0-enacted/enacted-temp-basal.json monitor/battery.json monitor/reservoir.json monitor/status.json are what mine are named
jordan-berger
@jordan-berger
Jun 22 2016 23:13
probably not. What step in the documentation is that?
Ben West
@bewest
Jun 22 2016 23:14
I'd do something like openaps report add nightscout/openaps-pill.json JSON ns shell status monitor/clock.json oref0-monitor/iob.json oref0-predict/oref0.json oref0-enacted/enacted-temp-basal.json monitor/battery.json monitor/reservoir.json monitor/status.json
https://github.com/openaps/docs/issues/129#issuecomment-223459583, it was recently updated actually so that it uses the same ns device
jordan-berger
@jordan-berger
Jun 22 2016 23:17
i ran your command, then ran openaps use ns shell status. It returned a long reply. What am i looking for to determine if it suceeded?
Ben West
@bewest
Jun 22 2016 23:21
or the other option would be to fix the template a bit and simply do oref0 template mint reports common-nightscout based off of recipe
you need to give the status command all those files
it's documented by ns-status, it's calling ns-status under the hood to prepare the pill's data
jordan-berger
@jordan-berger
Jun 22 2016 23:23
I used the regular documentation . Should i use the walkthrough? "This walkthrough outlines an easier method than the below. We need volunteers to run through this method and replace the sections below as appropriate."
Ben West
@bewest
Jun 22 2016 23:23
so for one thing, openaps pill is optional
feel free to skip this step
or disable plugin until later
to get the pill working, the ns shell status has to be given all the same files as documented by ns-status -h
and it's two steps: first step is prepare and compile the data
the second step is uploading the prepared data to nightscout
so ns shell status monitor/clock.json oref0-monitor/iob.json oref0-predict/oref0.json oref0-enacted/enacted-temp-basal.json monitor/battery.json monitor/reservoir.json monitor/status.json for me prepares all the data because I have those files, but I think people commonly using different names and we're still cooking up re-usable recipes
iananders
@iananders
Jun 22 2016 23:28
i am trying to invoke my suggested.json and it appears my NS time is frozen and not advancing at all - i tried to re-deploy NS but the time hasn't updated - anyone have any ideas?