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

25th
May 2016
Paul Martin
@Spazholio
May 25 2016 00:05
@amoskane I just got my Edison flashed. Arrived in the mail today. =) Are you using the ERF or the TI stick with it?
amoskane
@amoskane
May 25 2016 00:10
Planning on TI stick- I got 2 of those and just got one set up on the pi, so the 2nd is for the edison. How was your edison flash? Hard/easy?
i’ve still been troubleshooting stuff on the pi- mainly getting all the info i want to post to nightscout- i still have spotty results with pump reservoir/battery and IOB and COB… so I haven’t gotten on the edison yet!
Paul Martin
@Spazholio
May 25 2016 00:14
@amoskane The flashing was a piece of cake. Used my Pi to do it (apparently it doesn’t work on OSX). @oskarpearson has a great workup here: https://github.com/oskarpearson/mmeowlink/wiki/Prepare-the-Edison-for-OpenAPS
amoskane
@amoskane
May 25 2016 00:28
yup, that’s my next page! :)
Rob Kresha
@rkresha
May 25 2016 02:36
@danamlewis google form submitted, what's my n #? :thumbsup:
took the pancreas to work with me today, went really well
anyone have an issue with rpi where dns just stops working and it can't resolve hostnames, blows my mind
Paul Martin
@Spazholio
May 25 2016 02:40
Any chance you’re switching networks or anything like that? Or is it just sitting there, then goes all DNS-stupid?
Rob Kresha
@rkresha
May 25 2016 02:43
@Spazholio both
Paul Martin
@Spazholio
May 25 2016 02:44
Anything in the syslog or any other logs that could indicate an issue?
Rob Kresha
@rkresha
May 25 2016 02:44
I've got a script checking for internet connectivity now and if the ping fails, it ifdown wlan0 ifup wlan0 and then releases dhcp lease and renews
Paul Martin
@Spazholio
May 25 2016 02:45
Good call. That’s along the lines of what I would have suggested.
ddaniels1
@ddaniels1
May 25 2016 02:54
Screenshot 2016-05-24 19.51.22.png
Rob Kresha
@rkresha
May 25 2016 02:54
It's just unfortunate, I think part of it is resolvconf jacking with things as well
@ddaniels1 :thumbsup:
ddaniels1
@ddaniels1
May 25 2016 02:55
Wondering what 3U meal bolus remaining means? IOB is 0.66.
Rob Kresha
@rkresha
May 25 2016 02:58
bolus of 5.1 for 56 grams of carbs
3U is left
ddaniels1
@ddaniels1
May 25 2016 02:59
Wouldn't that be reflected in the IOB?
Rob Kresha
@rkresha
May 25 2016 03:00
that's basal IOB
if you hover over it
ddaniels1
@ddaniels1
May 25 2016 03:02
Basal IOB was -0.1, the IOB total should include the bolus. @danamlewis or @scottleibrand or @bewest any ideas?
Rob Kresha
@rkresha
May 25 2016 03:03
@ddaniels1 not sure, sorry for my misinformation
David Cintron
@loudestnoise
May 25 2016 03:26
Neato, able to talk to the pump from my Pi with CareLink plugged in. I haven’t put a reservoir in the pump yet. I only have 3 sets at the moment. Trying to figure out what my next steps are. I’m of course looking at the docs, but it’s not clear where I’m supposed to be with the pump. I’m also reading all the gitter chat for this room dating back to August 2015. Hoping that will help me understand more.
Paul Martin
@Spazholio
May 25 2016 03:39
Any guidance we can give? I’m fairly new myself, but maybe I can point you in the right direction?
Boris and Kayley Raskin
@boryanvbu
May 25 2016 03:41

Hi Everyone, I am setting openaps for the first time. I have the carelink connected to the raspberry pi and the $lsusb command does show that the Carelink [Medtrnic Minimed Carelink] is detected. I already added the medtronic pump with the serial number to the openaps.ini. When I run any commands that start with $ openaps use pump xxxx, I receive the following error

pi@raspberrypi:~/myopenaps $ openaps use pump mytest      
Traceback (most recent call last):
  File "/usr/local/bin/openaps-use", line 4, in <module>
    __import__('pkg_resources').run_script('openaps==0.1.0', 'openaps-use')
  File "build/bdist.linux-armv7l/egg/pkg_resources/__init__.py", line 719, in run_script
  File "build/bdist.linux-armv7l/egg/pkg_resources/__init__.py", line 1504, in run_script
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.1.0-py2.7.egg/EGG-INFO/scripts/openaps-use", line 63, in <module>
    app( )
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.1.0-py2.7.egg/openaps/cli/__init__.py", line 51, in __call__
    self.run(self.args)
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.1.0-py2.7.egg/EGG-INFO/scripts/openaps-use", line 57, in run
    output = app(args, self)
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.1.0-py2.7.egg/openaps/uses/__init__.py", line 92, in __call__
    return self.method.selected(args)(args, app)
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.1.0-py2.7.egg/openaps/uses/__init__.py", line 31, in __call__
    return self.method(args, app)
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.1.0-py2.7.egg/openaps/uses/use.py", line 45, in __call__
    output = self.main(args, app)
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.1.0-py2.7.egg/openaps/vendors/medtronic.py", line 206, in main
    return self.pump.model.my_read_settings( )
AttributeError: 'Pump' object has no attribute 'model';

I have seen a number of users posting the same error, but it turned out to be a problem with their carelink. Any ideas why the error is occurring when the carelink is detected?

eyim
@eyim
May 25 2016 03:42
@loudestnoise If you want to get the pump working without a reservoir, rather than putting in a reservoir stick your pinky or a pencil into the space and let the piston press against it. It will stop moving up thinking it is a reservoir and then you can use the pump
David Cintron
@loudestnoise
May 25 2016 03:49
@Spazholio I guess I’m just trying to figure out what my next steps are. I made it to https://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-1/visualization.html#configuring-and-uploading-openaps-status in the docs and I feel like it assumes that I have already been using openaps and know what I’m doing.
that may not make sense I realize
I guess I was hoping for a step by step guide. E.g. “Now lets start your openaps! On your Pi via SSH type openaps init to get started!"
so now I’m just going back in Gitter and read other people’s questions hoping to figure it out
@eyim I heard that’s what you can do to trick the pump. Kind of scares me to do with my pump, but maybe I’ll give it a try.
Paul Martin
@Spazholio
May 25 2016 03:54

Do you have NS installed? Do you have a CGM?

And I think the lack of a clear-cut step-by-step encourages you to know before you do. As in, don’t just blindly copy/paste commands and expect to have a working external pancreas “out of the box”. That’s just the impression that I got.

David Cintron
@loudestnoise
May 25 2016 03:55
I do have NS installed. I’m feeding it with BG data via Bridge. I’ve got a G5 transmitter and use my iPhone as a receiver.
Paul Martin
@Spazholio
May 25 2016 03:57
Ok, I’m not going to be able to help there, as I have a G4+Share, and use a directly connected CGM. But so far, that aspect is working - you’ve got NS populated with CGM data?
David Cintron
@loudestnoise
May 25 2016 03:58
I totally understand the don’t just copypasta things expecting it to work. That’s why I guess I’m just trying to keep reading to learn and understand
Yep, I’ve got dev NS running and a profile setup and it’s currently getting populated with CGM data live
Screen Shot 2016-05-24 at 10.59.23 PM.png
Paul Martin
@Spazholio
May 25 2016 04:00
The directions on that link you pasted for setting up an ns device might be slightly outdated. I think @bewest would suggest using oref0 and the nightscout device instead. I mean, they should work still, but I know that he’d really like more people to use oref0. =)
C-Ville
@C-Ville
May 25 2016 04:01
Guten morgen
David Cintron
@loudestnoise
May 25 2016 04:01
Hmm, guess I need to find that info then
C-Ville
@C-Ville
May 25 2016 04:02
ups sorry wrong channel good morning :-)
Paul Martin
@Spazholio
May 25 2016 04:04
https://github.com/openaps/docs/issues/129#issuecomment-216069305 <— that might shed some light on what I’m referring to.
David Cintron
@loudestnoise
May 25 2016 04:05
ah very good, that helps
Paul Martin
@Spazholio
May 25 2016 04:07
It sounds like you’ve got the groundwork laid - you’ve got communications with your pump, and you’ve got a CGM that feeds into NS. Now you just need to start creating reports and learning how they all interact. Have you seen this? https://github.com/scottleibrand/openaps-sh/blob/dev/setup.sh
It’s a shell script that @scottleibrand wrote that creates all the reports that are needed to get a loop up and running. I highly encourage you to read it through, and understand what they’re all doing to get a feel for everything. It really, REALLY helped me when getting set up. Like…a week or so ago. =)
David Cintron
@loudestnoise
May 25 2016 04:10
Hadn’t seen that yet. I’ll start reading over that for sure. Just trying to soak it all up like a sponge :smile:
Paul Martin
@Spazholio
May 25 2016 04:11
If you have questions, just ask. Everyone here a) is crazy helpful, and b) has likely been where you are now. =)
David Cintron
@loudestnoise
May 25 2016 04:14
That’s great to know. I feel like I’m at least making progress with communicating to the pump!
Paul Martin
@Spazholio
May 25 2016 04:15
Big time. Personally, I got fed up with the limited range of the Carelink stick, and got a TI stick very quickly, but the basics are the same.
David Cintron
@loudestnoise
May 25 2016 04:17
Oh I full expect if I get it working with the “basic setup” I’ll want to move to Edison/TI :laughing:
ddaniels1
@ddaniels1
May 25 2016 04:17
That took me about a week:)
David Cintron
@loudestnoise
May 25 2016 04:17
I also just have to figure out how to get a consistent set of medtronic pump supplies
I’ve been using OmniPod since 2013
ddaniels1
@ddaniels1
May 25 2016 04:17
The initial set up can be frusturating. Stick with it, plenty of people here to help
David Cintron
@loudestnoise
May 25 2016 04:18
and I don’t think my insurance will go for the supplies for two different insulin pumps
Paul Martin
@Spazholio
May 25 2016 04:18
I would check the usual places - CL, eBay, Medwow, etc.
David Cintron
@loudestnoise
May 25 2016 04:19
Yeah, I’m sure I’ll be able to get it somehow
eszcloud
@eszcloud
May 25 2016 06:06
Likewise, @mhaeberli!
Scott Leibrand
@scottleibrand
May 25 2016 06:24
@rkresha meal assist is a little odd it doesn't use COB, but rather calculates whether the carbs over the last few hours have been "covered" by boluses or high temps. If not, it tries to administer the remaining insulin gradually over the next DIA hours by adjusting the normal basal (and then further adjusting that for high or low BG as normal). It works, but I switched to a completely different method that actually estimates COB and uses that directly in the advanced meal assist branch that we'll be merging to dev soon.
@loudestnoise if the docs are assuming you've done something you haven't, keep reading. Some of them are out of order a bit since the reorg. We need someone (you?) to note which things need moved up, and then actually move them and PR in the changes.
Dana Lewis
@danamlewis
May 25 2016 06:49
@rkresha just got off an 8 hour plane ride, give me some time to catch up ;) will ping you when I catch up in a day or two :)
Ben West
@bewest
May 25 2016 07:14
oy try oref0 template :-)
Ben West
@bewest
May 25 2016 08:05
@loudestnoise try oref0 template -h and playing with the options and openaps import: openaps/docs#132
C-Ville
@C-Ville
May 25 2016 08:12
@C-Ville
talking about battery hungry, currently I pull my pump settings in each loop cycle. As already mentioned in the Document that is not needed and I assume that drains the battery relatively fast. I am reluctant to put the request for pump settings in the cron job with a lower frequency than the monitor/suggest/enact reports, because I am afraid that after a ccertain period of time this might lead to collision when both cron jobs put a pump request to the pump at the same time. Am I just paranoid or how do you handle that?
sorry wrong channel, will move that to intend to bolus
Scott Leibrand
@scottleibrand
May 25 2016 08:14
you are correct: multiple crons will likely collide at some point
what we do is refresh the settings at the end of every loop. if the refresh fails due to poor comms, that's no big deal, and it will try again next loop.
in my experience, "doing less" in the loop doesn't help much with pump battery: once the radio is turned on with an init, it doesn't matter how many things you do while it's on.
C-Ville
@C-Ville
May 25 2016 08:36
@scottleibrand thanks for clarifying, so I will just keep it as it is :-)
Scott Leibrand
@scottleibrand
May 25 2016 08:38
and by "at the end of the loop" I mean after the enact
C-Ville
@C-Ville
May 25 2016 08:39
okay, so I will make some changes, I have it sqeezed in somwehere between preflight and monitoring things
Nathan Morse
@morsen
May 25 2016 09:44
My son’s been sitting exactly at target for a couple of hours now. You all are awesome! Thank you!
Dana Lewis
@danamlewis
May 25 2016 09:44
:) :sparkles:
you just thought we were all crazy at the first meetup. welcome to the land of crazy-but-real ;)
Nathan Morse
@morsen
May 25 2016 09:46
actually i’ve been too busy (read “lazy”) to get involved since i first heard about you in early 2015, but I’m very glad to be here where you are. after i get things settled into a groove i can start contributing. thx again :)
Dana Lewis
@danamlewis
May 25 2016 09:47
:) :)
Chris Oattes
@cjo20
May 25 2016 10:24
danamlewis: you're in the UK now?
Dana Lewis
@danamlewis
May 25 2016 10:25
correct ;) in Shoreditch as we speak!
Chris Oattes
@cjo20
May 25 2016 10:25
welcome :) Sorry it's so grey today :P
Dana Lewis
@danamlewis
May 25 2016 10:25
what gave it away, finally having timezone appropriate tweets? lol.
no worries, just like i’m back in seattle ;)
Chris Oattes
@cjo20
May 25 2016 10:25
yeah :P
had nice blue skies yesterday
Dana Lewis
@danamlewis
May 25 2016 10:26
holding out hope for tomorrow!
Chris Oattes
@cjo20
May 25 2016 10:27
do you get to rest today after flying overnight? Or do you have to work?
Dana Lewis
@danamlewis
May 25 2016 10:28
I’m working today :D
Dana Lewis
@danamlewis
May 25 2016 10:35
@cjo20 are you going to be able to make it tomorrow?
Chris Oattes
@cjo20
May 25 2016 10:43
yes, I should be able to
@danamlewis I think I updated the whoozin to say 'yes'
Dana Lewis
@danamlewis
May 25 2016 10:45
:+1:
and you saw the lunch one, too?
Chris Oattes
@cjo20
May 25 2016 10:46
I saw it, but I'm unlikely to be able to make it. I've got work, and the train to get in to London takes 30 minutes
so I'd get to Euston and then have to turn around again :P
Dana Lewis
@danamlewis
May 25 2016 10:48
fair enough
Chris Oattes
@cjo20
May 25 2016 11:03
I've decided to take tomorrow afternoon off work, so if I can think of something to fill in the time between 2pm and 7pm I might come along
Scott Leibrand
@scottleibrand
May 25 2016 11:31
Some of the other attendees might have some free time as well. Maybe y'all could do a mini-hackathon. :-)
mm22dl
@mm22dl
May 25 2016 12:17
Hey there! New question of the day! :-D Since I don't own a CGM yet (thinking about getting the Dexcom G4 by the end of the summer), I was wondering if there is a way to simulate CGM data income with OpenAPS and see if everything works fine? Are there maxbe datasets somewhere that I could use and compare my results to? thanks
!
:)
Dana Lewis
@danamlewis
May 25 2016 12:19
^ FYI, people keep asking for that. If anyone wants to pay it forward, writing up a quick way to access someone’s already public data ;) would be a good one.
mm22dl
@mm22dl
May 25 2016 12:34
I'll see what I can do! Still got plenty of time anyway! ^^ Thanks!
Dana Lewis
@danamlewis
May 25 2016 12:56
yep! that wasn’t at you, that was more for someone else looking for something to do as a volunteer project, unless you end up doing it…that’s great too :D
there’s not an existing known, documented way to do what you’re asking, but it’s possible and i’ve heard it enough times to think it would be great to get written up
mm22dl
@mm22dl
May 25 2016 13:02
I'm not promising anything either! :-P But if it's in my range of capabilities, I'll try coming up with something for sure. :)
Scott Leibrand
@scottleibrand
May 25 2016 13:16
If you know someone's NS URL, you can download their BG data to a test rig the same way you'd do your own.
mattmholliday
@mattmholliday
May 25 2016 14:27
I'm relativity new to all of this. I have NS setup and am now looking at Open aps to start making an open loop.
Does OpenAps replace Nightscout? or does it just send extra data to it?
Scott Leibrand
@scottleibrand
May 25 2016 14:29
OpenAPS can download BG data from NS and upload information on insulin delivery, etc. to NS.
mattmholliday
@mattmholliday
May 25 2016 14:30
ok. cool
Scott Leibrand
@scottleibrand
May 25 2016 14:30
Once they're doing that, many people use Nightscout as the main User Interface for OpenAPS
Dana Lewis
@danamlewis
May 25 2016 14:30
@mattmholliday you’ll want to keep NS, it’s great for visualizing what the loop does
mattmholliday
@mattmholliday
May 25 2016 14:32
cool. is oref0 currently the best place to start?
JaysonEwer
@JaysonEwer
May 25 2016 14:33
Welcome, @mattmholliday have you gone through the following:https://openaps.org/reference-design/
mattmholliday
@mattmholliday
May 25 2016 14:35
Yes. Thanks @JaysonEwer have had a read through it. As i have a 640g and no way to set a temp basil programmatically i will only be able to leverage On open loop
have been reading the docs here https://github.com/openaps/docs
JaysonEwer
@JaysonEwer
May 25 2016 14:37
Ah, very good.
mattmholliday
@mattmholliday
May 25 2016 14:37
At this stage I am aiming to get everything up to Phase 3 setup. And have it suggest to a pebble watch what temp basal i should set for the next 30mins
So you are communicating with your pump?
Are you using Enlite or Dexcom CGM?
mattmholliday
@mattmholliday
May 25 2016 14:42
Yes currently using https://github.com/pazaan/640gAndroidUploader to send the data to ns
Enlite
I'm looking at setting openaps up on a server and just have it feed me recomendations, instead of carrying it around with me on a PI
JaysonEwer
@JaysonEwer
May 25 2016 14:50
Ah, nice!
Are you using connect to send data to your phone and then on to NS?
Sorry, silly question
That's what the uploader does you mentioned above.
Did you get Openaps setup on the server yet?
mattmholliday
@mattmholliday
May 25 2016 14:53
https://github.com/pazaan/640gAndroidUploader Uses the Bayer Contour next link to get SG and Pump info from the pump and posts it to NS
haven't installed OpenAPS as of yet was just trying to see how it all linked back to NS. and if running an Open Loop works
everything looked like it required openaps to talk directly to the pump and cgm. I'm glad @scottleibrand was able to clarify that Openaps can get its data from NS aswell.
JaysonEwer
@JaysonEwer
May 25 2016 14:57
this is a very interesting setup that you are proposing. As far as I know no one is doing exactly this.
Yes, I would suggest working through the RTD docs to setup Openaps and as you go take care to note all the steps that you need to modify with your cloud instance of Openaps. And when you run into hurdles that you need assistance with ask in here and we'll all do our best to get you over those hurdles. 😎
mattmholliday
@mattmholliday
May 25 2016 15:00
Well hopefully it works. :) would love to have a closed loop but don't have a pump that would enable me to.
JaysonEwer
@JaysonEwer
May 25 2016 15:02
The interesting thing is making sure you have communications the whole time. I'm assuming you will always have good wifi or mobile connection? In other words you don't, at the moment, see yourself in an off-line mode.
mattmholliday
@mattmholliday
May 25 2016 15:03
no worries there i sit at a desk all day.
Andrew Baugh
@baughaw
May 25 2016 15:03
Will you then enter the basal rate by hand?
mattmholliday
@mattmholliday
May 25 2016 15:04
@baughaw Do you mean the Temp basal for corrections?
Andrew Baugh
@baughaw
May 25 2016 15:04
temp basal
yeah
mattmholliday
@mattmholliday
May 25 2016 15:05
yes
Andrew Baugh
@baughaw
May 25 2016 15:05
my loop runs every 5 mins and the temp is constantly changing.. might be difficult to keep up
JaysonEwer
@JaysonEwer
May 25 2016 15:07
Yes, it will be a bit challenging, but probably better than not having it.
mattmholliday
@mattmholliday
May 25 2016 15:07
i use pomodoro to manage my time at work so hoping i can slow the loop down and only act every 20-30mins
the 640g has an automatic low glucose suspend so i will mainly be chasing highs and bad carb guesses
Andrew Baugh
@baughaw
May 25 2016 15:08
@JaysonEwer I agree it will be better than nothing.
JaysonEwer
@JaysonEwer
May 25 2016 15:09
@baughaw are you from Pittsburgh? I saw your note the other day.
I grew up in North East PA, but some of my family lives near Pittsburgh, now.
Scott Leibrand
@scottleibrand
May 25 2016 15:18
@mattmholliday you should also check out https://gitter.im/pazaan/decoding-contour-next-link and https://twitter.com/tims_pants/status/735423181607993346 re: looping with a 640G.
They haven't finished the CNL decoding yet, but it looks like they're getting close, so we all need to be thinking about how to do so as safely as possible.
In particular, as I noted in my comment on that document, we need to start exploring how the predictive low glucose management works when BG starts to drop after doing microboluses to correct a high. that might be something you could start testing manually as you start open looping
and, as noted in that article, HAPP is an excellent tool for open looping while you start figuring out how to build an OpenAPS rig that will do more of what you want.
Andrew Baugh
@baughaw
May 25 2016 15:21
@JaysonEwer Yeah I am. I saw you have family here
mattmholliday
@mattmholliday
May 25 2016 15:23
Thanks @scottleibrand will go read that now. From what I have experienced the lgs seems to work quite.
Scott Leibrand
@scottleibrand
May 25 2016 15:24
yeah, the question is how much we can rely on a black-box algorithm like that to low-temp when needed, or whether we need to be a lot more conservative in high-temping since it may not take into account IOB and may only low-temp when BG actually starts to fall quite a bit.
I've never used it myself, so I would love to see some good analyses of exactly how it seems to work in different situations, so we can think about how much to trust it as part of a closed loop.
mattmholliday
@mattmholliday
May 25 2016 15:29
Will watch it more closely but it seems to kick in when sugars are falling rapidly. The rate of fall required to triggerseems to be less the closer the the low limit you get. I've also never seen it kick on with a sg above 8 mmol
Once I have a weeks worth of data happy to share some reports with times it tripped.
Scott Leibrand
@scottleibrand
May 25 2016 15:40
check out the comments on that blog post: looks like the 640G may actually be IOB-aware. would love to see some hands-on testing of that.
from what I've heard it uses something like a 45m-in-the-future BG prediction and triggers if that is below threshold. not sure how IOB would get used in making that prediction though
mattmholliday
@mattmholliday
May 25 2016 15:46
Hmm might put some manual blouses through tomorrow and see if it trips it
Disconnected of course
rbrenner87
@rbrenner87
May 25 2016 15:52
I am unable to get the openaps pill to show up on my nightscout azure website. Confirmed running the dev branch with github as my deployment source. The plugin checkbox shows up in the settings menu on the site but no openaps pill is ever displayed when checked. Other plugins like IOB COB CAGE all show up fine on the right hand side when enabled. Didn't want to bother anyone here with such a simple question but I've redeployed, deleted the whole application and recreated from scratch, tried different browsers and am stumped. Any advice would be greatly appreciated.
Andrew Baugh
@baughaw
May 25 2016 15:52
Does anyone know if it is possible to replace the pump housing? I have a lead on a 523 but it is cracked at the battery. I have another 523 (firmware 2.5) that could be switched out if it is possible.
David Cintron
@loudestnoise
May 25 2016 16:12
@rbrenner87 just a guess, but do you have the DEVICESTATUS_ADVANCED=true in addition to ENABLE=openaps I don’t actually know what the DEVICESTATUS_ADVANCED does, but just came to mind
rbrenner87
@rbrenner87
May 25 2016 16:18
@loudestnoise thanks for the suggestion. I do have both DEVICESTATUS_ADVANCED=true and openaps added under ENABLE in azure application settings. The openaps plugin checkbox is displayed under settings > show plugins on my nightscout webpage but the actual pill never shows up when the box is checked.
Lynne Butts
@flutelynne_twitter
May 25 2016 16:30
@loudestnoise I am at a similar phase of setup (though waiting for my carelink usb). I have been using NS with G5 bridge, and NS for 2 yrs, set up openaps to read from NS. How do I verify that openaps is reading NS? Tried invoking reports, etc., but not getting anything.
ericmtx
@ericmtx
May 25 2016 16:34
@flutelynne_twitter do you get data from the site when you go to https://yourwebsite/api/v1/entries.json ? That would be the location that provides the data that openaps would pull from NS
Lynne Butts
@flutelynne_twitter
May 25 2016 16:41
@ericmtx Yes! Very cool! Thank you. This is quite a process, as well it should be. Fingers crossed.
Jason Calabrese
@jasoncalabrese
May 25 2016 16:45
@loudestnoise @rbrenner87 DEVICESTATUS_ADVANCED=true is used to return all the device statuses from the last 24 hours, without that we're in simple cgm only mode and only return the last device status and assume it's for uploader battery
all the info for the Pump and OpenAPS pills comes from devicestatus
David Cintron
@loudestnoise
May 25 2016 16:46
Good to know!
Jason Calabrese
@jasoncalabrese
May 25 2016 16:47
we still need something to support pulling 48-hours ago to 24 hours ago for retro mode when scrolling back, since it was too heavy to always load
David Cintron
@loudestnoise
May 25 2016 16:48
@flutelynne_twitter so from what I’ve gathered you can pull data from NS in openaps by using ns-status, although @Spazholio I think was telling me last night that might be less preferred method and to instead use nightscout
I could be saying that wrong though
I am able to pull in CGM values from NS by doing this use command
(pi @ raspberrypi): ~/daveaps
⇢  openaps use ns shell get entries.json 'count=1'
[
  {
    "direction": "Flat", 
    "trend": 4, 
    "dateString": "2016-05-25T15:58:44.000Z", 
    "sgv": 153, 
    "device": "share2", 
    "date": 1464191924000, 
    "_id": "5745cc2221122306055c45d4", 
    "type": "sgv"
  }
]
but that’s as far as I’ve gotten. I’m not yet doing reports, but hope to soon be getting there
rbrenner87
@rbrenner87
May 25 2016 16:55
@jasoncalabrese ever had the OpenAPS pill refuse to show up even though the checkbox is present and checked under Show Plugins under settings on nightscout site? Azure settings all correct as well including 'DEVICESTATUS_ADVANCED=true'
Jason Calabrese
@jasoncalabrese
May 25 2016 17:00
could be data related, check YOUR_SITE/api/v1/treatments.json?find[enteredBy]=/openaps/
Lynne Butts
@flutelynne_twitter
May 25 2016 17:01
@loudestnoise Thank you.
Scott Leibrand
@scottleibrand
May 25 2016 17:03
we desperately need a volunteer to write up the nightscout docs. We're wasting way too much time here covering the same ground repeatedly.
Lynne Butts
@flutelynne_twitter
May 25 2016 17:04
@loudestnoise So wonderful to know I am on the right spot.
Scott Leibrand
@scottleibrand
May 25 2016 17:04
Should be a great pay it forward opportunity for anyone who's just finished setting it up that way or wants to do it right the first time.
David Cintron
@loudestnoise
May 25 2016 17:04
@scottleibrand :+1:
I would love to be that person, but still eh trying to figure it all out
Scott Leibrand
@scottleibrand
May 25 2016 17:06
Lots of folks here happy to help, especially if we know you're documenting as you go. :-)
rbrenner87
@rbrenner87
May 25 2016 17:06
@scottleibrand Sorry, wont waste anyone else's time asking. Searched quite a bit through the gitter chat archives and couldn't find a solution.
Scott Leibrand
@scottleibrand
May 25 2016 17:07
Please do ask, but write up what works as you go along.
Too many people seem to assume if they can't figure it out they must be doing something wrong. More likely the docs need updates.
I'm half tempted to go vandalize the docs with "this part sucks: please PR improvements" all over the place. ;-)
Dana Lewis
@danamlewis
May 25 2016 17:09
I've asked for that to be done ;)
Scott Leibrand
@scottleibrand
May 25 2016 17:10
maybe that's the first step: ask people with questions to PR those questions into the docs for someone to answer them there as well?
Dana Lewis
@danamlewis
May 25 2016 17:10
:+1:
Scott Leibrand
@scottleibrand
May 25 2016 17:16
any volunteers to do that right now? Only qualification required is you have to be stuck because the docs are unclear. ;-)
garykidd
@garykidd
May 25 2016 17:16
Was just going to print full .pdf version of openaps from rtd. .pdf version seems to cut off certain text--otherwise being able to create .pdf is extremely handy.
Paul Martin
@Spazholio
May 25 2016 17:17
@scottleibrand Do you know if the nightscout device got merged into the AMA branch? I know that @bewest was working on it and might have had a mini-breakthrough, but I didn't follow up with him afterward...
Scott Leibrand
@scottleibrand
May 25 2016 17:17
Pretty sure I merged that this week.
Yep: openaps/oref0#121
@garykidd if you can figure out how to "fix" the docs to make them compatible with RTD PDFs, feel free to PR the fix to dev.
David Cintron
@loudestnoise
May 25 2016 17:23
@scottleibrand So I definitely meet that qualification of being stuck. I want to help with the docs as much as I can, so let’s do this!
I forked the docs repo to my github and now I would just edit the doc and do a PR?
but I’m supposed to do it back to dev? or master
Scott Leibrand
@scottleibrand
May 25 2016 17:24
For now the instructions say dev, so let's do that until we change it officially.
and yea on the rest.
David Cintron
@loudestnoise
May 25 2016 17:27
and when you say “PR those questions into the docs” are you suggesting I try and put what I’m trying to do into the docs and then use the comments on the PR as the discussion on what I’m doing wrong or what needs to be changed?
garykidd
@garykidd
May 25 2016 17:28
blob
Scott Leibrand
@scottleibrand
May 25 2016 17:28
Or just put in a little blurb into the docs that says "this part is confusing" or "doesn't work" and ask for someone to fix it and remove your comment.
garykidd
@garykidd
May 25 2016 17:29
@scottleibrand here's an example of the text getting cut off in the .pdf version... I dont really know how to fix though. I realize this is a small thing...but still could be helpful to fix
David Cintron
@loudestnoise
May 25 2016 17:29
@scottleibrand Ok, I think I can handle that!
Scott Leibrand
@scottleibrand
May 25 2016 17:29
Putting a \ at the end of the line allows you to wrap commands like that.
Not sure if that would make it less understandable though.
garykidd
@garykidd
May 25 2016 17:30
@scottleibrand can PR the issue though (maybe not before this wekeend though)
@scottleibrand some sections already wrap though and appear completely normal---wonder what's driving the difference.
Scott Leibrand
@scottleibrand
May 25 2016 17:32
We can also use GitHub issues for more meta things like this.
The inline stuff is to prompt newbies who are going through the docs to realize where they can help.
Maybe it's single backticks vs. multi line triple backticks?
Paul Martin
@Spazholio
May 25 2016 17:35
Currently, I get my CGM data directly from my Dexcom receiver. When I switch over to the Edison, I won't be able to do that, will I?
Scott Leibrand
@scottleibrand
May 25 2016 17:37
Only if you use openxshareble with a G4 receiver that isn't paired to a phone.
Otherwise you can get it from NS.
garykidd
@garykidd
May 25 2016 17:38
@scottleibrand do you use azure for NS for your implementation?
Paul Martin
@Spazholio
May 25 2016 17:38
But that implies network connectivity. At least now, if I don't have network connectivity, the loop will still function. If I have to get everything from NS, and I can't, doesn't that mean the loop can't really function properly?
Scott Leibrand
@scottleibrand
May 25 2016 17:39
Correct. Hence openxshareble.
Paul Martin
@Spazholio
May 25 2016 17:39
But my Dexcom is paired to my phone. Hence, time to start being more aggressive with the iPhone hotspot. =)
Rob Kresha
@rkresha
May 25 2016 17:40
@Spazholio iphone?
Paul Martin
@Spazholio
May 25 2016 17:41
Correct.
Scott Leibrand
@scottleibrand
May 25 2016 17:43
Or figure out BT tethering.
Paul Martin
@Spazholio
May 25 2016 17:45
That was a special kind of hell with the Pi. I'll just stick to the WiFi hotspot. It's not like I'm away from WiFi for all that long anyway. Usually it's just when in the car driving 30 mins to and from work.
David Cintron
@loudestnoise
May 25 2016 17:46
I just want one of those Google hot air balloons with WiFi to follow me around everywhere I go
Scott Leibrand
@scottleibrand
May 25 2016 18:00
Or just get a mifi type device. :-)
ericmtx
@ericmtx
May 25 2016 18:00
a mifi and some helium balloons. You're all set
David Cintron
@loudestnoise
May 25 2016 18:04
@ericmtx LOL
Paul Martin
@Spazholio
May 25 2016 18:06
Don't suppose anyone has a spare ERF they'd like to get some money for? =) I was going to order them online, but stuff from the UK takes forever to arrive. And shipping costs suck.
David Cintron
@loudestnoise
May 25 2016 18:08
Is the ERF option better than a TI stick? Is one physically smaller?
Paul Martin
@Spazholio
May 25 2016 18:08
ERF's smaller, but with reduced range compared with the TI stick from what I've heard.
eyim
@eyim
May 25 2016 18:09
@Spazholio Paul - send me your address
David Cintron
@loudestnoise
May 25 2016 18:10
do you also not need one of the flasher things for ERF like you do for TI stick? Who makes this ERF device anyway?
Paul Martin
@Spazholio
May 25 2016 18:10
Well, I have a TI stick, so I got the CC debugger at the same time. And yes, I think you need one for it. Pretty sure, actually.
eyim
@eyim
May 25 2016 18:11
Flashing ERF is different than using the CC debugger for TI
Paul Martin
@Spazholio
May 25 2016 18:11
I thought it still used the device, but differently. No? My mistake.
Scott Leibrand
@scottleibrand
May 25 2016 18:15
@loudestnoise Awesome, thanks. In addition to the call out of what needs to be fixed, you linked people to a workaround. Perfect. :-)
Everyone see how easy that was? openaps/docs#192
David Cintron
@loudestnoise
May 25 2016 18:16
Yeah, well the best part is I hadn’t kept on reading that far to get to Phase 2
Scott Leibrand
@scottleibrand
May 25 2016 18:17
And now everyone else will know to do so, and not get stuck. :-)
David Cintron
@loudestnoise
May 25 2016 18:17
so my working inside openaps use was just reading from gitter history and @Spazholio’s links he gave me last night to the nightscout that @bewest wrote up
Hichame Yessou
@hichameyessou
May 25 2016 18:46

Someone have an idea of why when I'm trying sudo openaps use pump model I get back nothing? Just a ""

The sensor is not linked to the pump, it can be a problem?

Scott Leibrand
@scottleibrand
May 25 2016 18:47
Don't use sudo there.
But the problem is likely with the stick. Have you tried a power cycle?
Hichame Yessou
@hichameyessou
May 25 2016 18:48
Ok, power cycle?
ericmtx
@ericmtx
May 25 2016 18:51
there's a way to reset the USB/carelink command in the docs, but you can also unplug and replug it manually for the same effect
and you have your pump serial number defined, I think in pump.ini?
Hichame Yessou
@hichameyessou
May 25 2016 18:56
I tried to plug and unplug the carelink but nothing changed, yes I already setup it
Hichame Yessou
@hichameyessou
May 25 2016 19:03
It looks like that I'm putting the wrong serial number because I get the same result with a real wrong one
There is any option to check the serial number of a pump from the pump? (Maybe the serial number sticked on the pump is wrong)
David Cintron
@loudestnoise
May 25 2016 19:07
@hichameyessou Yeah, if you hit Esc and button down to the bottom it has the serial listed
Hichame Yessou
@hichameyessou
May 25 2016 19:08
Thank you! The SN is correct :s
Dan Piessens
@dpiessens
May 25 2016 19:11
@Spazholio I'm still working on paring the G5 to the edison directly, making progress using xDrip code as a reference for the python. The only other workaround is to wire the TI directly to the edison with a GPIO board and then use usb for the Dexcom receiver
Hichame Yessou
@hichameyessou
May 25 2016 19:23
I got another idea, is possible that the European Carelink cannot work with an US pump?
Scott Leibrand
@scottleibrand
May 25 2016 19:28
Possible. They do use different frequencies.
Hichame Yessou
@hichameyessou
May 25 2016 19:29
So I'll try with mmeowlink and my CC1111, maybe I can resolve it without buying a new carelink
Ben West
@bewest
May 25 2016 19:50
@dpiessens any sources/wip? :-)
regarrdifng g5, I've been planning a daemon to work in conjunction with oacids
Dan Piessens
@dpiessens
May 25 2016 20:21
@bewest Yup, right here: https://github.com/dpiessens/openxshareble forgive the fact it's hacked off your original, it was a good starting point.
What does oacids do?
Paul Martin
@Spazholio
May 25 2016 20:24

So, I updated to the latest in the AMA branch, and now I’m getting a few errors. Unsure why.

[{"msg":"Optional feature Meal Assist enabled, but could not read required meal data.","file":"monitor/meal.json","error":{}}]. And when I try to invoke that report manually, I get Could not parse input data: [SyntaxError: Unexpected end of input].

Scott Leibrand
@scottleibrand
May 25 2016 20:24
which report?
is meal.json empty and you're trying to invoke that report?
if so, try checking the inputs and running oref0-meal manually
Paul Martin
@Spazholio
May 25 2016 20:27
Running oref0-meal monitor/pumphistory.json settings/profile.json monitor/clock.json monitor/glucose.json settings/basal_profile.json give me {"carbs":0,"boluses":0,"mealCOB":0}. And yes, it was empty. I didn’t bother with the carbhistory.json file, since while mine exists, it’s empty.
Scott Leibrand
@scottleibrand
May 25 2016 20:31
and are those the same inputs used by your meal.json report? does it do the same?
need more details on what exactly is failing
Paul Martin
@Spazholio
May 25 2016 20:32
openaps use meal shell monitor/pumphistory-zoned.json settings/profile.json monitor/clock-zoned.json settings/carbhistory.json monitor/glucose.json settings/basal_profile.json [] is how I generate my meal.json. I didn’t use the -zoned.json in the manual command above, but other than that, it’s the same.
Scott Leibrand
@scottleibrand
May 25 2016 20:33
order of arguments matters
Paul Martin
@Spazholio
May 25 2016 20:33
blink
How…did that happen?
Scott Leibrand
@scottleibrand
May 25 2016 20:33
I changed them
Paul Martin
@Spazholio
May 25 2016 20:33
And why didn’t it die horribly….oh.
Scott Leibrand
@scottleibrand
May 25 2016 20:33
and added backwards compatibility, but...
it could be that, or it could be your [] remainder
Paul Martin
@Spazholio
May 25 2016 20:34
I tried removing the [] remainder, and it didn’t change anything.
Scott Leibrand
@scottleibrand
May 25 2016 20:34
you know how to narrow this stuff down: LMK what works and what doesn't. :)
Paul Martin
@Spazholio
May 25 2016 20:35
Yeah, I didn’t realize the order changed. Fixing now...
Scott Leibrand
@scottleibrand
May 25 2016 20:35
it should still work. if it doesn't work out of order, it's a bug in my backwards compatibility workaround
Paul Martin
@Spazholio
May 25 2016 20:36
Well, I changed the order, and it worked.
Scott Leibrand
@scottleibrand
May 25 2016 20:36
so need to figure out what you're doing differently than me. may be the empty carbhistory or something
Paul Martin
@Spazholio
May 25 2016 20:36
So I get to be the bearer of bad news. =)
Hang on...
Scott Leibrand
@scottleibrand
May 25 2016 20:36
some people are really good at finding bugs. :)
Paul Martin
@Spazholio
May 25 2016 20:37
fields = pumphistory profile clock carbhistory glucose basal_profile <— that’s how it was in my meal.ini
Scott Leibrand
@scottleibrand
May 25 2016 20:37
I like those kind of people: makes it much easier to fix them, because they have learned from experience how to report the bugs. :)
that looks correct for the old order
(without actually looking at code)
garykidd
@garykidd
May 25 2016 20:39
This message was deleted
Paul Martin
@Spazholio
May 25 2016 20:39
Well, there’s nothing else special in my meal.ini. Let me check the report in openaps.ini
garykidd
@garykidd
May 25 2016 20:39
This message was deleted
This message was deleted
Paul Martin
@Spazholio
May 25 2016 20:40
[report "monitor/meal.json"]                                                                                                                                                                                                                                                                                         
profile = settings/profile.json
use = shell
reporter = text
clock = monitor/clock-zoned.json
json_default = True
pumphistory = monitor/pumphistory-zoned.json
device = meal
carbhistory = settings/carbhistory.json
glucose = monitor/glucose.json
basal_profile = settings/basal_profile.json
remainder = []
Scott Leibrand
@scottleibrand
May 25 2016 20:40
single backtick test successful. twice.
Paul Martin
@Spazholio
May 25 2016 20:40
hehe
Scott Leibrand
@scottleibrand
May 25 2016 20:40
so to narrow down the argument order thing, can you run oref0-meal manually with the old order and make it fail for me?
need to know what exact inputs being absent/empty makes it fail
Ben West
@bewest
May 25 2016 20:41
report show --cli monitor/meal.json to get matching use
Paul Martin
@Spazholio
May 25 2016 20:41
One moment.
I pasted my report show —cli up above. I love that command so very very much. =)
Ben West
@bewest
May 25 2016 20:42
:-D made with :heart:
garykidd
@garykidd
May 25 2016 20:42
@scottleibrand thanks. just was geting back to your earlier comment on the .pdf file. the second test was actually triple backtick--which I thought should've shown up differently.
Paul Martin
@Spazholio
May 25 2016 20:43
oref0-meal monitor/pumphistory.json settings/profile.json monitor/clock.json settings/carbhistory.json monitor/glucose.json settings/basal_profile.json this is the manual version, but run with my previous order. Gives me: Could not parse input data: [SyntaxError: Unexpected end of input]
Same thing happens if I add the remainder on the end.
Scott Leibrand
@scottleibrand
May 25 2016 20:51
@garykidd you were probably missing the newlines
@Spazholio and that was with the empty settings/carbhistory.json right?
Paul Martin
@Spazholio
May 25 2016 20:52
That’s correct.
Scott Leibrand
@scottleibrand
May 25 2016 20:52
ok. can you do me one more favor and open an oref0 issue on that?
Paul Martin
@Spazholio
May 25 2016 20:52
Now that the ns device is working again, I’ll try generating that file in a bit.
Ok.
Scott Leibrand
@scottleibrand
May 25 2016 20:54
and then if you get bored later you can look at https://github.com/openaps/oref0/blob/advanced-meal-assist/bin/oref0-meal.js#L49-L52 and figure out how to make it error more gracefully if that file is empty
maybe we should go back to doing it this way: https://github.com/openaps/oref0/blob/dev/bin/oref0-meal.js#L54-L60
it's 11pm here, so I'll look at it later if someone else doesn't figure it out first
Paul Martin
@Spazholio
May 25 2016 20:56
I can take a look in a bit.
Scott Leibrand
@scottleibrand
May 25 2016 21:27
has anyone made a list of which pumps we do and don’t think have radio communication capabilities and might be usable for looping? If not, I'm starting one, and would love contributions: https://docs.google.com/document/d/1FGYdqILhf3zXgGJYBRbpfC1Fz0ZgmRsjrcVFN3FLUj8/edit?usp=sharing (will probably post this to openaps/docs or somewhere like that once we have it finalized, but I also need it for something else)
David Cintron
@loudestnoise
May 25 2016 21:28
@bewest I’ve watched that talk before, but now that I’m started working on #OpenAPS it is more interesting. Especially from the slide he showed at one point of an OmniPod. Means they can of course interact with it, given they know how...
Ben West
@bewest
May 25 2016 21:39
also, BigFoot is in there
and the shot at the end is of the Mazlish family
David Cintron
@loudestnoise
May 25 2016 21:43
if that presentation was redone a few years later it would have been like check out this guy @bewest and @danamlewis. They decided to just make their own artificial panceras system :shipit:
Ben West
@bewest
May 25 2016 21:45
well, it turns out Mazlish was enabled by me posting carelink online
David Cintron
@loudestnoise
May 25 2016 21:46
oh wow that’s cool
Tom Boudreau
@TomBoudr_twitter
May 25 2016 21:51
Q for the group. I have noticed my SSH is very slow, like sometimes the characters take seconds to appear... may have been after I did this command from the docs 'sudo apt-get update && sudo apt-get -y upgrade' I this normal RPi behavior or is something likely wrong. Wifi and such should not be the issue.
Scott Leibrand
@scottleibrand
May 25 2016 21:53
I notice that sometimes, and it generally seems to be wifi or IP address related. You might try just rebooting, or if you actually want to troubleshoot, see if you have a valid IP address and/or try to renew it with dhclient.
Paul Martin
@Spazholio
May 25 2016 21:54
I get that ALL the time on my Pi3.