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

5th
Jun 2016
Bill Stackpole
@billstackpole
Jun 05 2016 00:00
[device "determine-basal"]
vendor = openaps.vendors.process
extra = determine-basal.ini
pi@billopenaps1:~/billopenaps $ openaps report show
pump://JSON/iter_pump_hours/last_four_pump_hours.json
pump://JSON/model/settings/model.json
pump://JSON/iter_glucose_hours/last_four_cgm_hours.json
pump://JSON/read_settings/settings/settings.json
pump://JSON/read_bg_targets/settings/bg_targets.json
pump://JSON/read_insulin_sensitivies/settings/insulin_sensitivities.json
pump://JSON/read_basal_profile_std/settings/basal_profile.json
get-profile://text/shell/settings/profile.json
pump://JSON/iter_pump_hours/monitor/pumphistory.json
pump://JSON/read_clock/monitor/clock.json
calculate-iob://JSON/shell/monitor/iob.json
pump://JSON/read_temp_basal/monitor/temp_basal.json
determine-basal://text/shell/enact/suggested.json
glucose://JSON/clean/monitor/glucoseclean.json
pump://JSON/model/our-model.json
pump://JSON/iter_glucose/monitor/glucose.json
ns-status://JSON/shell/monitor/upload-status.json
tz://JSON/clock/monitor/clock-zoned.json
pump://JSON/set_temp_basal/enact/enact.json
pump://JSON/reservoir/reservoir.json
pump://JSON/mmtune/monitor/mmtune.json
pump://JSON/set_temp_basal/enact/enacted.json
pump://JSON/mmtune/mmtune.json
pump://JSON/reservoir/monitor/reservoir.json
pump://JSON/read_battery_status/monitor/battery.json
pump://JSON/read_status/monitor/status.json
pebble://text/shell/upload/pebble.json
tz://JSON/rezone/monitor/pumphistory-zoned.json
pump://JSON/read_carb_ratios/settings/carb_ratios.json
ns://JSON/shell/nightscout/recent-treatments.json
ns://JSON/shell/nightscout/uploaded-missing-treatments.json
ns://JSON/shell/nightscout/uploaded-status.json
pi@billopenaps1:~/billopenaps $
should I paste the report configs from cat openaps.ini?
Scott Leibrand
@scottleibrand
Jun 05 2016 00:02
Ah, so the remainder thing fixed it? Good enough then.
Frank Best
@efbest
Jun 05 2016 00:07
That's after calibrating:
openaps use MMT754 iter_glucose 4
[
  {
    "body": "66", 
    "packet_size": 5, 
    "amount": 102, 
    "name": "CalBGForGH", 
    "raw": "42-80-05-10-66", 
    "date": "2016-06-05T02:00:00", 
    "date_type": "minSpecific", 
    "_tell": 4, 
    "op": "0x0e"
  }, 
  {
    "packet_size": 4, 
    "name": "SensorTimestamp", 
    "raw": "4f-97-24-10", 
    "date": "2016-06-04T15:23:00", 
    "date_type": "minSpecific", 
    "_tell": 1, 
    "op": "0x08"
  }, 
  {
    "name": "GlucoseSensorData", 
    "date_type": "prevTimestamp", 
    "_tell": 6, 
    "sgv": 96, 
    "date": "2016-06-04T15:29:00", 
    "packet_size": 0, 
    "op": 48
  }, 
  {
    "name": "GlucoseSensorData", 
    "date_type": "prevTimestamp", 
    "_tell": 7, 
    "sgv": 88, 
    "date": "2016-06-04T15:24:00", 
    "packet_size": 0, 
    "op": 44
  }, 
  {
    "name": "GlucoseSensorData", 
    "date_type": "prevTimestamp", 
    "_tell": 8, 
    "sgv": 86, 
    "date": "2016-06-04T15:19:00", 
    "packet_size": 0, 
    "op": 43
  }
Bill Stackpole
@billstackpole
Jun 05 2016 00:14
just removed the square brackets and re-ran the loop - all fixed (pretty cool! Thanks for the help.)
Jarred Yaw
@jyaw
Jun 05 2016 00:18
@danamlewis saw your tweet with the watch face where you had like 3 quick options like eating soon and 15g carbs. What are you using to do that?
Dana Lewis
@danamlewis
Jun 05 2016 00:21
Need someone to figure out an IFTTt integration with NS for others to be able to do Alexa type stuff too. My Alexa and @scottleibrand 's watchface are pushing buttons on DIYPS to enable the easy entry for me. But I believe someone else built a similar watch face or app that pushed entries into NS care portal , I just don't remember who did it or the name. @jyaw . If no one chimes in here knowing it off hand, you might ask in CGMITC about NS care portal entry app .
Scott Leibrand
@scottleibrand
Jun 05 2016 00:23
I think someone did set up NS IFTTT integration, but it only worked when API_SECRET wasn't enabled for POSTs.
The pebble app we use there is called ThisButton. It works with the IFTTT maker channel.
Frank Best
@efbest
Jun 05 2016 00:26
@scottleibrand Calibrating worked! I think you gave me that advice before - I've got to remember those weird things! And again, like so many times before, thanks for your support!
Dana Lewis
@danamlewis
Jun 05 2016 00:28
@efbest curious about timing - had it been asking for a calibration? When was the last time you had calibrated before?
@jyaw looks like @PazolesBrooke_twitter just asked on CGMITC so guessing she'll report back the findings ;) :)
Frank Best
@efbest
Jun 05 2016 00:30
no, it stopped functioning some 10 hours ago; calibration was due at 2:00, so it was just about the right time! :smile:
Lynne Butts
@flutelynne_twitter
Jun 05 2016 00:31
@flutelynne_twitter jinx!
Lynne Butts
@flutelynne_twitter
Jun 05 2016 00:31
Jinx!
Dana Lewis
@danamlewis
Jun 05 2016 00:31
Hm. Ok. Was wondering if it was related to the funkiness of Medtronic CGM not showing data if you don't calibrate.
Good stuff @flutelynne_twitter @Pazoles ! ^ @jyaw
would be easy to modify to do temp targets and whatnot, too ^
Scott Leibrand
@scottleibrand
Jun 05 2016 00:32
Who are the main developers on the Pebble Careportal app? Wonder if they'd be interested in adding temp targets to it.
Frank Best
@efbest
Jun 05 2016 00:33
@danamlewis I could see the data on the pump screen all the time, but I couldn't find a way to create a 'fresh' glucose.json!
Lynne Butts
@flutelynne_twitter
Jun 05 2016 00:34
an email address pops up: app_dev_568fb97705f633b36200
Dana Lewis
@danamlewis
Jun 05 2016 00:34
If it happens again, maybe make note of the preceding calibration time, there may be some correlations if it is a recurring problem
Frank Best
@efbest
Jun 05 2016 00:35
will do!
Lynne Butts
@flutelynne_twitter
Jun 05 2016 00:35
@scottleibrand This is tapping the developer option on the careportal app in Pebble.
Dana Lewis
@danamlewis
Jun 05 2016 00:35
@efbest :+1:
Jarred Yaw
@jyaw
Jun 05 2016 01:12
@danamlewis @scottleibrand Ok this button sounds cool. Maybe if we can get the IFTTT integration worked out with NS that'll be the ticket. Sounds like you guys have some chain of things that bypasses the need for talking directly to NS from IFTTT. I have the care portal app on my pebble just doesn't do those specific shortcuts. BTW, what exactly is "eating soon mode"?
(this may be another instance where if I was running the dev branch I would be in the know)
Scott Leibrand
@scottleibrand
Jun 05 2016 01:14
It's actually in AMA: temp targets
Dana Lewis
@danamlewis
Jun 05 2016 01:14
It sets targets (temp targets, can be enabled through care portal) to 80 before a meal, idea is to get insulin activity going
Jarred Yaw
@jyaw
Jun 05 2016 01:16
Hmm so that takes some of the potential error out of the pre-bolus then. Nice.
Dana Lewis
@danamlewis
Jun 05 2016 01:16
Totally.
Jarred Yaw
@jyaw
Jun 05 2016 01:19
So it looks like it's time for an upgrade to AMA and autosens then. Thanks for the explanations!
Paul Martin
@Spazholio
Jun 05 2016 01:20
+1 for AMA/temp targets.
Dana Lewis
@danamlewis
Jun 05 2016 01:21
:)
Docs need more love there, so pr anything in that's missing - either setup instructions or English explanations of features
Jarred Yaw
@jyaw
Jun 05 2016 01:36
Just general FYI for anyone setting up the careportal pebble app, the path they reference to treatments in the setup is missing the /api before v1/treatments.
Foot in mouth.... no that was just me. I should read better. Was wondering why it hasn't been working for me.
Jason Calabrese
@jasoncalabrese
Jun 05 2016 01:38
The Pebble Careportal app code is here: https://github.com/CGMintheCloud/PebbleCareportal
the more I think about the more it feels like NS needs to learn to work with multiple CGM sources, also dropping xdrip would mean that I lose reporting on the phone's battery
Scott Leibrand
@scottleibrand
Jun 05 2016 01:44
Why does the phone battery matter if you're not using it as uploader?
Jason Calabrese
@jasoncalabrese
Jun 05 2016 01:44
it would still be needed as a link
running the hotspot
Scott Leibrand
@scottleibrand
Jun 05 2016 01:44
Ah
Jason Calabrese
@jasoncalabrese
Jun 05 2016 01:45
probably other android apps that upload battery and network stats
also finding that multiple receivers is big pain when it comes to pairing
Paul Martin
@Spazholio
Jun 05 2016 01:47
Folks who use an Edison + iOS rig - do you just enable the Personal Hotspot pretty much in perpetuity, or…?
Jason Calabrese
@jasoncalabrese
Jun 05 2016 01:51
that's what I'm doing with an android phone
the more efficient way to do it is via bluetooth, sounds like it works with Yocto, not sure about Ubi
Paul Martin
@Spazholio
Jun 05 2016 01:53
As in, using Bluetooth to connect to the hotspot?
Jason Calabrese
@jasoncalabrese
Jun 05 2016 01:53
yeah
then you don't have to remember to turn it on
Paul Martin
@Spazholio
Jun 05 2016 01:54
That sucked on the Pi. With any luck that will be simpler on the Edison. And Ubi, hopefully.
Bill Stackpole
@billstackpole
Jun 05 2016 01:54
I don't have it done yet but yes. It's called BT-NAP (Network Access Point)
Paul Martin
@Spazholio
Jun 05 2016 01:54
And at that point, all of my CGM data is pretty much reliant on being able to download from NS, right? Since connecting the Dexcom to the Edison could prove to be problematic/bulk-adding?
Bill Stackpole
@billstackpole
Jun 05 2016 01:55
I had it turned on in my pi3 via the gui but was unsuccessful scripting it.
Dana Lewis
@danamlewis
Jun 05 2016 01:55
No. Openxshareble to the Edison. (Another reason for 2 receivers - one to the phone for cloud and one to the Edison)
Paul Martin
@Spazholio
Jun 05 2016 01:56
I managed to get it working once. No idea how, and never managed to duplicate the precise series of events...
Oof. So I’d have to carry two receivers?
Eric Luhrs
@eluhrs
Jun 05 2016 01:56
When I started experimenting with Bluetooth hotspot between my iPhone and Edison, I found that I could not also connect the phone to wifi at the same time. Is that normal?
Dana Lewis
@danamlewis
Jun 05 2016 01:57
Yes
Yes to both of you ;)
Paul Martin
@Spazholio
Jun 05 2016 01:58
Rats. I was kinda hoping to have a system small enough to carry around in my pocket or something.
Dana Lewis
@danamlewis
Jun 05 2016 01:58
Receiver can only pair to one thing
Eric Luhrs
@eluhrs
Jun 05 2016 01:58
:)
Is it the same with Android?
Paul Martin
@Spazholio
Jun 05 2016 01:58
Yeah, I figured. =/
Dana Lewis
@danamlewis
Jun 05 2016 01:59
@eluhrs I believe that's probably also likely for android
You might like the idea of a hootoo to bridge the wifi to multiple devices
GeoffWyche
@GeoffWyche
Jun 05 2016 01:59
Outsider/off-the-cuff here, but has there been discussion of a home-device to provide connectivity between a cloud solution (NS) and other home devices such as CGMs and pumps? Something like cloudfleet but with specifically designed software for NS/CGM/pump?
Dana Lewis
@danamlewis
Jun 05 2016 02:00
@GeoffWyche not familiar with that, what does it do?
GeoffWyche
@GeoffWyche
Jun 05 2016 02:00
It's mostly just a cloud device -- connect your home to "some software out there" that takes care of you.
However, it is open source, so it can be adapted to a specific need.
Eric Luhrs
@eluhrs
Jun 05 2016 02:01
@danamlewis I'm not familiar with hootoo. How would it work?
GeoffWyche
@GeoffWyche
Jun 05 2016 02:01
Note, I don't specifically recommend cloudfleet.
It's just an example of a home-hardware-device that provides connectivity of a specific type.
We could build a home-hardware-device that is specific to NS/CGM
Dana Lewis
@danamlewis
Jun 05 2016 02:02
@eluhrs it's a router splitter basically - plug in hootoo, sign it on to a wifi network, then it becomes a network you can have all your devices connect to. Ie call hootoo network openaps, log on to hotel wifi or plane wifi and your rigs and laptop and phone can all connect automatically
Scott Leibrand
@scottleibrand
Jun 05 2016 02:02
@GeoffWyche so what would you want the analogous system to do?
GeoffWyche
@GeoffWyche
Jun 05 2016 02:03
I see lots of problems with people trying to set up home systems to connect to NS. Build a pre-configured hardware device that can talk to NS.
"pre-configured" is loose here...
Dana Lewis
@danamlewis
Jun 05 2016 02:03
Hm. Maybe. You might want to suggest that over in Nightscout/public.
Most people don't do a hardware setup for NS, think the % of people trying to do Pi etc is very small
GeoffWyche
@GeoffWyche
Jun 05 2016 02:04
@danamlewis good point.
Jason Calabrese
@jasoncalabrese
Jun 05 2016 02:04
in some ways a edison/pi running openaps is that
Dana Lewis
@danamlewis
Jun 05 2016 02:04
@mariannesmith77_twitter welcome :)
Eric Luhrs
@eluhrs
Jun 05 2016 02:06
@danamlewis Interesting. But I guess you'd have to carry the hootoo around and always have a plug handy? Would make flying difficult. At least in my travel class. ;)
Dana Lewis
@danamlewis
Jun 05 2016 02:06
Some of them come with a built in battery if that helps you.
GeoffWyche
@GeoffWyche
Jun 05 2016 02:06
Well, yes, I would expect it would have the same mobility as your home refrigerator, even if it was much smaller.
Dana Lewis
@danamlewis
Jun 05 2016 02:07
I always carry batteries to power my phones and rigs, and 50% of flights these days have power, so not a big deal to me. But hootoo I use most often in hotels conferences etc
Paul Martin
@Spazholio
Jun 05 2016 02:07
Has anyone else noticed some slight connection issues if their rig is within 4 feet of them BUT it has to travel through their body? I keep my rig to the left of me while watching TV, and I had to switch my pump to my right side, and I’m noticing more connectivity issues than usual.
Eric Luhrs
@eluhrs
Jun 05 2016 02:07
One of the many things I love about the openaps community. I learn of all kinds of interesting cloud/hardware/app things I never knew about...
Scott Leibrand
@scottleibrand
Jun 05 2016 02:07
@GeoffWyche sounds like what you want is a super-easy setup script for a read-only openaps instance that just uploads to NS.
Dana Lewis
@danamlewis
Jun 05 2016 02:07
@Spazholio 4 feet, I manage even 1/2 foot to block with my body sometimes ;)
@eluhrs yea. Always learning around here, it is awesome! :)
Paul Martin
@Spazholio
Jun 05 2016 02:08
It gets the small queries, but fails quite a bit on the pumphistory-24h.json report.
GeoffWyche
@GeoffWyche
Jun 05 2016 02:08
@scottleibrand not really... more of a "buy (make) this device and it will hook up your CGMs and NS instance"
Scott Leibrand
@scottleibrand
Jun 05 2016 02:09
"hook up" how?
GeoffWyche
@GeoffWyche
Jun 05 2016 02:10
Because the managers of the software distribution (us?) create the software distribution to be user friendly. On install, "what's your pump brand?" etc...
Scott Leibrand
@scottleibrand
Jun 05 2016 02:11
Most people upload CGM data to NS via Dexcom Share bridge, FWIW. That is super easy.
GeoffWyche
@GeoffWyche
Jun 05 2016 02:11
It wouldn't be as clean as a cable-modem, yet, but would be about like installing a new computer in your home.
at least for folks who were already savvy with the CGMs
Scott Leibrand
@scottleibrand
Jun 05 2016 02:12
Still not seeing what the gap is that needs filled if it's not about uploading data the way an openaps rig does.
GeoffWyche
@GeoffWyche
Jun 05 2016 02:13
Only that it doesn't depend on your knowledge of apple, or of windows, or Pi.
Scott Leibrand
@scottleibrand
Jun 05 2016 02:14
So there's Dexcom Share and Minimed Connect...
Eric Luhrs
@eluhrs
Jun 05 2016 02:15
Does minimed connect work with our pumps? I assumed it did not...
Marianne Smith, PMP
@mariannesmith77_twitter
Jun 05 2016 02:15
Hey everyone, my husband and I are just getting started with OpenAPS for our T1D son. So far just completed the general set up steps. Reading the openaps utility belt now, trying to learn what we are doing. The warning about not using openaps commands in the openaps repo is repeated a few times but we don't get it. Could someone explain? we are using ssh to get to our raspberry pi. Will changes we make via raspberry pi change our git files?
Scott Leibrand
@scottleibrand
Jun 05 2016 02:15
only the 723s afaik.
But it sounded like @GeoffWyche was talking about a non-openaps problem and solution.
Eric Luhrs
@eluhrs
Jun 05 2016 02:16
Right. And only some of the 723s work for us, I think.
Marianne Smith, PMP
@mariannesmith77_twitter
Jun 05 2016 02:17
@danamlewis thanks for the warm welcome, you are all an amazing group of people, really glad to be starting this journey - with tons to learn!
GeoffWyche
@GeoffWyche
Jun 05 2016 02:17
@scottleibrand What if OpenAPS had its own computer -- not just software, but a dedicated computer, available for about $50, like a router.
Eric Luhrs
@eluhrs
Jun 05 2016 02:18
I was interested because I'm looking for an automatic way to get my pump data into carelink so I can read it next to my cgm data in tidepool's blip
Dana Lewis
@danamlewis
Jun 05 2016 02:18
@mariannesmith77_twitter which one of the docs are you talking about re: the warning? Is this the sudo warning or re not running in src/openaps?
*which one of the docs = which page are you on? Link?
GeoffWyche
@GeoffWyche
Jun 05 2016 02:18
@mariannesmith77_twitter welcome! there is much to learn here.
Scott Leibrand
@scottleibrand
Jun 05 2016 02:19
@GeoffWyche one thing we've talked about is building custom rigs, maybe using an Edison or Pi Zero, and then selling them, but without any software on them.
Then separately making easy-to-install read-only uploader software to put on it.
JaysonEwer
@JaysonEwer
Jun 05 2016 02:20
@mariannesmith77_twitter Welcome!
Marianne Smith, PMP
@mariannesmith77_twitter
Jun 05 2016 02:20
@danamlewis I'm studying this document: https://github.com/openaps/openaps/blob/master/README.md
Scott Leibrand
@scottleibrand
Jun 05 2016 02:20
Not sure who would want to do the hardware side of that. One issue is the best hardware is a moving target.
Marianne Smith, PMP
@mariannesmith77_twitter
Jun 05 2016 02:21
@danamlewis here is an example of what is confusing but seems important to understand since it is repeated a few times: "Do not use openaps commands in the the openaps repo. Only use the openaps directory for hacking on the core library, or for managing upgrades through git. Running openaps inside of the openaps source directory will error in the best case, and mess up your openaps install in the worst case."
Dana Lewis
@danamlewis
Jun 05 2016 02:21
@mariannesmith77_twitter you won't be installing from source. You'll be installing from package, so jump over instead to the openaps docs.
Scott Leibrand
@scottleibrand
Jun 05 2016 02:22
Another question is whether it's worth doing that sort of effort on a full hardware solution or to go with a RileyLink and a phone app based on something like LoopKit
GeoffWyche
@GeoffWyche
Jun 05 2016 02:22
@scottleibrand I agree -- moving target... but I usually think of OpenAPS as being purely software. What if it had its own hardware platform? with an embedded RileyLink, etc.
Dana Lewis
@danamlewis
Jun 05 2016 02:22
That's a easier to navigate version @mariannesmith77_twitter
Ben West
@bewest
Jun 05 2016 02:23
I'm not sure a phone app is what people actually want/need
as far as having a ready-made kit/rig it's something I've been working toward
Scott Leibrand
@scottleibrand
Jun 05 2016 02:23
lots of talk about things like custom sparkfun boards...
Ben West
@bewest
Jun 05 2016 02:23
check out web-loop-config and oacids
together they provide simple 1,2,3 gui panel to set up openaps
GeoffWyche
@GeoffWyche
Jun 05 2016 02:23
@scottleibrand Ah -- that was my original question -- are people thinking about it?
Ben West
@bewest
Jun 05 2016 02:24
yeah, we have meeting with intel/spakfun coming up, and I've got some other contacts as well
Marianne Smith, PMP
@mariannesmith77_twitter
Jun 05 2016 02:24
@danamlewis ok, great. We are working on the openaps docs but got redirected there... We are on "Configuring and Uploading APS status" on this page: https://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-1/visualization.html#setup-script
Ben West
@bewest
Jun 05 2016 02:24
would be great to get more organized :-)
GeoffWyche
@GeoffWyche
Jun 05 2016 02:24
@bewest :supercool:
Ben West
@bewest
Jun 05 2016 02:24
we met ceo and founder from seeed studios at maker faire, good contacts from there
Marianne Smith, PMP
@mariannesmith77_twitter
Jun 05 2016 02:26
@danamlewis first step under "Configuring and Uploading APS status" and we got stuck, need to learn more about what to do inside open aps before we move further
Jason Calabrese
@jasoncalabrese
Jun 05 2016 02:26
@bewest what's the new way to select the latest file
Ben West
@bewest
Jun 05 2016 02:26
openaps use tz select file1 file2 file3
be to sure to specify the --date field to use
Jason Calabrese
@jasoncalabrese
Jun 05 2016 02:27
it's on tz?
guess that makes sense
Dana Lewis
@danamlewis
Jun 05 2016 02:28
@mariannesmith77_twitter stuck how? What did you run on the command line and what was the output?
Jason Calabrese
@jasoncalabrese
Jun 05 2016 02:28
there's nothing to do a recency check right? if older than X run report
I know we can do it with a bash alias, but would be nice to clean that up
GeoffWyche
@GeoffWyche
Jun 05 2016 02:29
@bewest If you hear of any contacts that is starting up and needs developers, I'd be happy to talk to them.
Marianne Smith, PMP
@mariannesmith77_twitter
Jun 05 2016 02:31
@danamlewis we are trying to generate a monitor/upload-status.json report. Run on command line "openaps report add monitor/upload-status.json" and got error "...... too few arguments". Husband is on the raspberry pi on another computer while I study on this one or I would cut/copy here.
Dana Lewis
@danamlewis
Jun 05 2016 02:33
Throw some -h in there
GeoffWyche
@GeoffWyche
Jun 05 2016 02:35
-h usually gives help.
Marianne Smith, PMP
@mariannesmith77_twitter
Jun 05 2016 02:40
@danamlewis ok, command line "openaps report add monitor/upload-status.json -h" and we get what appears to be a help file.
we don't understand the help file, not programmers (yet!)
Dana Lewis
@danamlewis
Jun 05 2016 02:43
It's a great place to learn! @bewest put in a lot of stuff to show you various tips for how to run each command
Since you're getting too few arg, it means you're missing something it is expedting
The -h output should show you what it is expecting to get
Ben West
@bewest
Jun 05 2016 02:46
oref0 copy-fresher --since
@mariannesmith77_twitter oref0 was updated about a week ago to prevent that
muffindad
@muffindad
Jun 05 2016 02:47
COMPLETED - step 0
Matt Pazoles
@Pazoles
Jun 05 2016 02:47
Some unexpected stuff happening with the pump change - the openaps pill on nightscout now is always showing up as updated 1 minute ago, the only information in the pill is IOB and basal IOB
Jason Calabrese
@jasoncalabrese
Jun 05 2016 02:48
copy-fresh is half of what I want
Matt Pazoles
@Pazoles
Jun 05 2016 02:48
I can run the loop and get the output, and can post treatments to NS
but for some reason, can't send in the status
Jason Calabrese
@jasoncalabrese
Jun 05 2016 02:48
I want to avoid trying ble, if I have network and get get bg from NS
live4sw
@live4sw
Jun 05 2016 02:49
@Pazoles I had a similar issue for a few days, are you using a TI stick?
Matt Pazoles
@Pazoles
Jun 05 2016 02:49
Yep
live4sw
@live4sw
Jun 05 2016 02:49
w/Edison?
Matt Pazoles
@Pazoles
Jun 05 2016 02:49
RPi
Jason Calabrese
@jasoncalabrese
Jun 05 2016 02:49
@Pazoles I fixed an issue with that, and just merged it to NS dev
Dana Lewis
@danamlewis
Jun 05 2016 02:49
Kudos @muffindad!
Jason Calabrese
@jasoncalabrese
Jun 05 2016 02:49
the new error messages needed to be handled some
the contents of enacted/suggested become an array with an error
Paul Martin
@Spazholio
Jun 05 2016 02:51
@jasoncalabrese What is the most recent version of NS?
Jason Calabrese
@jasoncalabrese
Jun 05 2016 02:51
the fix was only to consider enacted/suggested if theres a timestamp field, but it's a little ugly
the dev branch
Ben West
@bewest
Jun 05 2016 02:52
@jasoncalabrese I just put --gaps, --count, --microseconds --seconds, --minutes and --hours for several uses in dexcom driver
--gaps can read output from lsgaps as well
Paul Martin
@Spazholio
Jun 05 2016 02:53

Ok, I’m running the dev version. And once I git pull, are the proper commands to run:

sudo npm install                                                                                                                                                                                                                                                                                                      
node node_modules/bower/bin/bower --allow-root install

Or is there something else/different that needs to be done?

Ben West
@bewest
Jun 05 2016 02:53
actually for that, it's just npm install
no sudo
Paul Martin
@Spazholio
Jun 05 2016 02:54
Ok. And the second line?
Ben West
@bewest
Jun 05 2016 02:54
shouldn't be needed
Paul Martin
@Spazholio
Jun 05 2016 02:54
We’ll find out in a moment...
Jason Calabrese
@jasoncalabrese
Jun 05 2016 02:55
the gaps seem great when you already know the source, but it you want to pick the source we need something else
Paul Martin
@Spazholio
Jun 05 2016 02:56
npm http GET https://registry.npmjs.org/crossfilter
npm http GET https://registry.npmjs.org/sugar
npm http GET https://registry.npmjs.org/mfb
npm ERR! git fetch -a origin (git://github.com/bewest/git-rev.git) error: cannot open FETCH_HEAD: Permission denied
npm ERR! Error: Command failed: error: cannot open FETCH_HEAD: Permission denied
npm ERR! 
npm ERR! 
npm ERR!     at ChildProcess.exithandler (child_process.js:637:15)
npm ERR!     at ChildProcess.EventEmitter.emit (events.js:98:17)
npm ERR!     at maybeClose (child_process.js:743:16)
npm ERR!     at Socket.<anonymous> (child_process.js:956:11)
npm ERR!     at Socket.EventEmitter.emit (events.js:95:17)
npm ERR!     at Pipe.close (net.js:466:12)
npm ERR! If you need help, you may report this log at:
npm ERR!     <http://github.com/isaacs/npm/issues>
npm ERR! or email it to:
npm ERR!     <npm-@googlegroups.com>

npm ERR! System Linux 2.6.32-042stab113.21
npm ERR! command "/usr/bin/nodejs" "/usr/bin/npm" "install"
npm ERR! cwd /home/spaz/nightscout
npm ERR! node -v v0.10.25
npm ERR! npm -v 1.3.10
npm ERR! code 1
npm ERR! git fetch -a origin (git://github.com/ktind/sgvdata.git) error: cannot open FETCH_HEAD: Permission denied
npm ERR! git fetch -a origin (git://github.com/mddub/minimed-connect-to-nightscout) error: cannot open FETCH_HEAD: Permission denied
npm ERR! git fetch -a origin (git://github.com/bewest/share2nightscout-bridge.git) error: cannot open FETCH_HEAD: Permission denied
That’s running npm install without sudo
Ben West
@bewest
Jun 05 2016 02:56
think sudo was run earlier probably
sudo rm -Rf node_modules
Paul Martin
@Spazholio
Jun 05 2016 02:57
It was. Should I chown -R username:username .?
Or that...
Same errors.
Pete Schwamb
@ps2
Jun 05 2016 03:00
Just realised: maybe this room should be called intend-to-temp-basal. ;)
Dana Lewis
@danamlewis
Jun 05 2016 03:00
;)
Ben West
@bewest
Jun 05 2016 03:01
yeah ,do the chown
there might some stuff in ~/.npm as well
Paul Martin
@Spazholio
Jun 05 2016 03:03
So I should run the chown on the ~/.npm dir as well?
Ben West
@bewest
Jun 05 2016 03:03
yup, if it's owned by root
Paul Martin
@Spazholio
Jun 05 2016 03:03
There are lots of root-owned files in there. Ok, running.
Ben West
@bewest
Jun 05 2016 03:05
@jasoncalabrese was thinking always run lsgaps on the cached version, and always try to fill it with each source
then use lsgaps and select again to get an active version
that way if one of the sources fails, it's trying to fill against the others
Paul Martin
@Spazholio
Jun 05 2016 03:06
@bewest Much appreciated, sir. Worked fine.
Jason Calabrese
@jasoncalabrese
Jun 05 2016 03:10
mixing sources might be bad
even just pulling mixed sources from NS
I think we'll need to compare the device when calculating deltas
Paul Martin
@Spazholio
Jun 05 2016 03:11
Going back to a really basic question - if my receiver is disconnected from my rig, my loop should be able to run just fine by working with NS data, correct?
Jason Calabrese
@jasoncalabrese
Jun 05 2016 03:12
yeah
Ben West
@bewest
Jun 05 2016 03:12
there's no features to dynamically select between devices, no
Jason Calabrese
@jasoncalabrese
Jun 05 2016 03:12
just hacks
Ben West
@bewest
Jun 05 2016 03:12
if it does it's because you did something extra :-)
Paul Martin
@Spazholio
Jun 05 2016 03:12
Well, right now it pulls from both, then merges them, a la @bewest magic.
So assuming that CGM data is unavailable, my NS data should merge with nothing, and just use that…right?
Ben West
@bewest
Jun 05 2016 03:13
yeah, basically
or merge with old data
and only the latest is used
Jason Calabrese
@jasoncalabrese
Jun 05 2016 03:13
unless it causes an error
Ben West
@bewest
Jun 05 2016 03:14
yeah, it'll likely crash/cause an error
Paul Martin
@Spazholio
Jun 05 2016 03:14
@bewest I have stopped doing “extra” things, and trust that you know what the hell I should be doing seeing as you have a much better grip on this than I do. =)
Ben West
@bewest
Jun 05 2016 03:14
you already found several with oref0 tools choking on cal and mbg entries
there will be other issues with non-json data getting created when devices fail currently
but assuming you have 2 devices that basically produce same data
Paul Martin
@Spazholio
Jun 05 2016 03:15
I’m thinking for when I move to the Edison, and I’ll likely just be using NS data since connecting the Dexcom would be problematic (and I use Share, so all that data should be present anyway). And then I’ll plug in the Dexcom at night to fill in any gaps that might have happened throughout the day.
Ben West
@bewest
Jun 05 2016 03:15
and they can always at least produce 0 data
0 valid data
if they both support --gaps, should be able to keep a cached copy of data for each, and then use lsgaps to figure out how much each needs to get to update the cache, then run lsgaps and select again against the newly updated cache files to run oref0 stuff against
in that way, they'd each try to update only the last bit, and if one fails, the other will still provide data
Paul Martin
@Spazholio
Jun 05 2016 03:17
Well, I’m currently selecting both the raw-cgm/glucose-zoned.json and the raw-ns/glucose-zoned.json files. I haven’t implemented the lsgaps part yet, but it’s waiting in the wings.
Ben West
@bewest
Jun 05 2016 03:18
yeah, that's the way to start
for now that's the "new updated cache" to select against, I'm not sure what happens if one of the files is missing or empty
Paul Martin
@Spazholio
Jun 05 2016 03:19
Now if the raw-cgm/glucose-zoned.json file doesn’t exist...
Ok, that was my next question.
Ben West
@bewest
Jun 05 2016 03:19
open issues for that :-)
Paul Martin
@Spazholio
Jun 05 2016 03:19
I’m thinking if that errors out, part of my loop could just create a zero length file, or a valid JSON file that has nothing in it to select against.
Ben West
@bewest
Jun 05 2016 03:19
yeah []
Paul Martin
@Spazholio
Jun 05 2016 03:19
Wait, are you telling me to open issues, or are you telling me there ARE open issues? =)
Ben West
@bewest
Jun 05 2016 03:20
no, this is new ground
so features/issues like this it's helpful to open an issue
Paul Martin
@Spazholio
Jun 05 2016 03:20
Who are we kidding, this is ALL new ground. =)
Ben West
@bewest
Jun 05 2016 03:20
tends to get lost in gitter
Paul Martin
@Spazholio
Jun 05 2016 03:21
Yeah, once I start to get my Edison set up from scratch, I may hit you up for broad strokes, because a) I’ve forgotten how I set up my rig initially in the first place, 2) I was using old methods back then, and III) I remember you just stepping 1-2 people through doing this exact same thing, but Gitter is hard to sift through. =)
Eric Luhrs
@eluhrs
Jun 05 2016 03:22
@Spazholio surely you can muster a iv) ;)
Paul Martin
@Spazholio
Jun 05 2016 03:23
iv) I really should have taken notes when @bewest was stepping the others through it a day or so ago? =)
Eric Luhrs
@eluhrs
Jun 05 2016 03:23
Maybe a bullet point next?
Paul Martin
@Spazholio
Jun 05 2016 03:25
  • Beat the horse until it’s pulp?
(gimme another) =)
Eric Luhrs
@eluhrs
Jun 05 2016 03:25
very good.
:)
Dana Lewis
@danamlewis
Jun 05 2016 03:25
lol
Jason Calabrese
@jasoncalabrese
Jun 05 2016 03:26
I guess just with the copy-fresher I could get rid of the delete
Dana Lewis
@danamlewis
Jun 05 2016 03:26
@bewest beautiful notes in progress :)
Lynne Butts
@flutelynne_twitter
Jun 05 2016 03:27
@Spazholio I write down dates and times to find things. Some major help from Ben on June 1 15:24 and beyond.
Nightscout help June 2 15:47 and on.
Jason Calabrese
@jasoncalabrese
Jun 05 2016 03:31
you can get links from gitter
Paul Martin
@Spazholio
Jun 05 2016 03:31
@flutelynne_twitter Thank you muchly. I’ll keep that in mind. =)
Jason Calabrese
@jasoncalabrese
Jun 05 2016 03:32
Screenshot 2016-06-04 20.31.47.png
Screenshot 2016-06-04 20.32.13.png
Paul Martin
@Spazholio
Jun 05 2016 03:32
@bewest Thanks for those links. I have….far too many tabs open right now. I really need to act on some of them soon.
(56 at current count)
Ben West
@bewest
Jun 05 2016 03:34
very nice :)
the archive is easier to browse, actually https://gitter.im/nightscout/intend-to-bolus/archives/2016/06/01
Paul Martin
@Spazholio
Jun 05 2016 03:38
57.
Lynne Butts
@flutelynne_twitter
Jun 05 2016 03:38
@Spazholio I didn't know that was even possible! I get nervous when I have 5.
Paul Martin
@Spazholio
Jun 05 2016 03:38
I scoff at your “5”. =)
Marianne Smith, PMP
@mariannesmith77_twitter
Jun 05 2016 03:39
@bewest @danamlewis @GeoffWyche Thanks for your help, we figured out what we were missing. Had to follow some of the instructions here: https://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-2/using-openaps-tools.html before we could do the step we were on. Gotta go to sleep now, almost midnight here. Thanks for your help tonight and in the future!
Paul Martin
@Spazholio
Jun 05 2016 03:39
I had the executive director at my college actually swear at the number of tabs I had open. Not at ME, but at the quantity of tabs. And that man doesn’t generally swear. It was a proud moment for me.
Dana Lewis
@danamlewis
Jun 05 2016 03:39
@mariannesmith77_twitter that would be great to do a PR to fix it for the next person! I think multiple people have run into that.
Ben West
@bewest
Jun 05 2016 03:41
hehe
I like lots of tabs also
Marianne Smith, PMP
@mariannesmith77_twitter
Jun 05 2016 03:41
@danamlewis will work on the PR tomorrow
Dana Lewis
@danamlewis
Jun 05 2016 03:41
awesome thank you! @mariannesmith77_twitter
Paul Martin
@Spazholio
Jun 05 2016 03:42
I’ve never gotten the hang of bookmarks. I have lots of OpenAPS pages bookmarked, but never look at them. It’s just easier to keep ‘em open. I know I have too many when I can’t see the favicons anymore. =)
Ben West
@bewest
Jun 05 2016 03:43
bookmarks are just for autocomplete to work better
it's same problem, but worse with 10k bookmarks
Dana Lewis
@danamlewis
Jun 05 2016 03:43
heh
Paul Martin
@Spazholio
Jun 05 2016 03:43
That’s….too many bookmarks.
Ben West
@bewest
Jun 05 2016 03:43
at least with 80 tabs I can tab through them
is there a tab counter, I figure there's plenty of room for more if I can still see all the logos and make out any letters at all
Dana Lewis
@danamlewis
Jun 05 2016 03:45
@bewest your computer always stresses me out ;)
Paul Martin
@Spazholio
Jun 05 2016 03:45
I use Session Buddy: https://chrome.google.com/webstore/detail/session-buddy/edacconmaakjimmfgnblocblbcdcpbko. So when I have to reboot, I get ‘em all back. =)
GeoffWyche
@GeoffWyche
Jun 05 2016 03:45
wow, i'm behind. Only 0.1k bookmarks.
Dana Lewis
@danamlewis
Jun 05 2016 03:45
re: the archive tip for searching…that would be great to PR into the docs, too ;) :D #brokenrecord #PRallthethings
Lynne Butts
@flutelynne_twitter
Jun 05 2016 04:00
I have lots of bookmarks. Especially for OpenAPS. And a lot on my reading list. Some things printed out, about 2 inches of paper with my notes and struggles. Just need a better brain to sort it all out.
Only people on this page would actually count each one @Spazholio : )
Disappointing that the battery I am using would not let me have it on charge and run the loop. Wifi would not hook back up. Ended up plugging in my pi with its cord, and separately charging the battery. Got about 10 hours out of it, though.
Paul Martin
@Spazholio
Jun 05 2016 04:11
There are some recommendations floating around out there for batteries that do pass-through charging. Oddly enough, that’s not one of the tabs I have open, though.
@danamlewis What is the temp targets a feature of? Meaning, I have it showing in NS, but is that all I need, or do I need to be running AMA in order for it to work properly?
Dana Lewis
@danamlewis
Jun 05 2016 04:14
More AMA goodness
Brian Bonner
@itsn0tm3
Jun 05 2016 04:15
is AMA in dev now or is it still on it's own branch?
Dana Lewis
@danamlewis
Jun 05 2016 04:18
Still separate branch
Dragan Maksimovic
@dm61
Jun 05 2016 04:18
@danamlewis I've just looked at the current version of the docs. The visualization section comes in too early - no wonder people are confused. It should be moved after using-openaps-tools.md. I do not know how to move doc sections around or delete/insert new sections, so I do not know how to PR that. Might be a few clicks for someone who knows how to do that? Once it's moved to where it should be, the visualization section needs to be re-done based on the new staff and notes by @bewest (which look a lot simpler than what we had before). I think I could PR that some time next week.
Brian Bonner
@itsn0tm3
Jun 05 2016 04:18
ah ok I saw some tickets about merging it into dev but couldn't tell if it actually did get merged or not
Dana Lewis
@danamlewis
Jun 05 2016 04:21
@dm61 we moved it forward bc the visualization needs to get setup up early on, and not at end. (It was moved forward with the knowledge that someone needed to update that section.) A Pr to adjust the language of what parts come before/after would be great, and yes it all needs to be redone also with the new tools in mind. Will look forward to a PR :)
Eric Luhrs
@eluhrs
Jun 05 2016 04:21
@flutelynne_twitter I've had great luck with this battery:
You can use it and charge at tne same time. Lasts about 11 hours.
Paul Martin
@Spazholio
Jun 05 2016 04:29
That was the one! I have one that never seems to actually top off, but as long as you keep it plugged in when you’re not out and about, it seems to work well. It lasted me roughly 30 hours last time I ran it down completely.
Paul Martin
@Spazholio
Jun 05 2016 04:47
@scottleibrand Can you point me to the docs that show what the auto-sens symbols mean?
Scott Leibrand
@scottleibrand
Jun 05 2016 04:51
don't think it's documented, just in the code
- means negative deviation, + means positive
> means positive enough it's being ignored, as likely being caused by carb absorption
! is some sort of error I think
probably one more I can't remember
Paul Martin
@Spazholio
Jun 05 2016 04:52
Huh. That’s a lot of errors yesterday. =)
Scott Leibrand
@scottleibrand
Jun 05 2016 04:53
you might not be passing all the history it needs?
Paul Martin
@Spazholio
Jun 05 2016 04:54
--+++++--+++++++++++++++++++++++++++>+!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!+---+!!!!!!!>>+----++-->++-----++++-+-+++--++---+++---+++------+--++++----++--++--++++-+--+--->>>-----------+++----+++++------+------------>>+---+++++++----+----+--+-++!!+!!+--------!!!!!!!!!++------+++-++-+->++--+- 52% of non-meal deviations negative (target 45%-50%)
Just looks like I had a bad chunk o’ data, maybe?
Scott Leibrand
@scottleibrand
Jun 05 2016 04:55
could've been your sensor restart
Paul Martin
@Spazholio
Jun 05 2016 04:55
I don’t think I usually have that many !s.
I’ll keep an eye on it through the day tomorrow.
Scott Leibrand
@scottleibrand
Jun 05 2016 04:55
the first round of !!! looks like about the time you were stopping and restarting your sensor
Paul Martin
@Spazholio
Jun 05 2016 04:55
Yeah, between the sensor screwing up, and me forgetting to “Start Sensor” today, I have about 4 hours of no data.
Oh, is it working….backwards, I suppose? The first two - are the most recent, or roughly 24 hours ago?
Scott Leibrand
@scottleibrand
Jun 05 2016 04:56
it starts with most recent, yeah
Paul Martin
@Spazholio
Jun 05 2016 04:56
Oh! Well…that makes a whole lot more sense.
Scott Leibrand
@scottleibrand
Jun 05 2016 04:58
the threshold for > might not be very good for you with the low-carb diet
does it detect resistance very often?
Paul Martin
@Spazholio
Jun 05 2016 05:00
Whenever I tail my openaps.log file (which is quite a bit, honestly) it seems to almost always adjust my ISF from 20 to 22. Nothing major, but fairly consistently.
Scott Leibrand
@scottleibrand
Jun 05 2016 05:03
Ok. If it's not lowering it, then that must not be an issue.
Unless it's missing sensitivity.
Paul Martin
@Spazholio
Jun 05 2016 05:05
I’d love to be able to see what it sets my ISF to on an hour-to-hour basis. Is that information stored in Mongo?
Scott Leibrand
@scottleibrand
Jun 05 2016 05:12
it's in the reason field in the openaps pill
Paul Martin
@Spazholio
Jun 05 2016 05:13
So there’d be a bit of Mongo learning plus parsing to be able to pull only that information out.
Scott Leibrand
@scottleibrand
Jun 05 2016 05:13
you should be able to hit the NS API
/api/v1/devicestatus.json
ericmtx
@ericmtx
Jun 05 2016 05:18
My daughter's git repo is corrupted on her edison. Tried oref0-reset-git and get "fatal: index file smaller than expected" when running in her openaps instance. Am I doing it wrong?
Paul Martin
@Spazholio
Jun 05 2016 05:18
I see them in there, yes. And while I could probably figure this out - how do I specify a date range? Is there a way to just say “Give me the last 5 days of data” or “Just stuff from the pump”?
Scott Leibrand
@scottleibrand
Jun 05 2016 05:19
/api/v1/treatments.json?find[created_at][$gte]=2016-04-16T12:57-0700&find[eventType][$eq]=Temporary Target
something like that
ericmtx
@ericmtx
Jun 05 2016 05:24
hmm - now I ran oref0-reset-git in the /openaps directory and the /maraloop instance now says it is "not a git repository". Did I kill her setup?
ericmtx
@ericmtx
Jun 05 2016 05:30
just did openaps init in her instance directory and that may have solved it
success
John Males
@johnmales
Jun 05 2016 05:52
Hi @efbest - sorry no fix for the out of date enlite CGM readings as yet, I’m waiting for it to happen again to capture some data. If it does happen to you, it would be good to capture the current and previous glucose pages and send to @bewest, as it is thought that it is something to do with changing glucose pages on the pump. I have found that fiddling with the pump e.g.. pressing esc-esc-up-down-esc or similar allows the cgm readings to be read again, so as a work around I’m thinking of implementing something to send a string of commands like this if the cgm data is out of date.
Lynne Butts
@flutelynne_twitter
Jun 05 2016 05:55
@eluhrs Thanks!
Frank Best
@efbest
Jun 05 2016 10:37
@johnmales thanks for your comments! Calibrating the pump (Scott L.'s proposal) did the job for me!
live4sw
@live4sw
Jun 05 2016 13:10
Kind of high-level question, but have people done work to analyze actual insulin delivered over the course of longer time periods (e.g., 30 minutes, 1 hour, 2 hours) when there are very frequent temp basal changes? I have two rigs running now and sometimes there are changes as often as every 2-3 minutes or so.
not super common but it definitely happens
Andy Probolus
@AndyProbolus_twitter
Jun 05 2016 15:23
good morning everyone. @mariannesmith77_twitter and I are working on OpenAPS
easy question for you. Can I use terminal to indicate where I want the script files saved?
Scott Leibrand
@scottleibrand
Jun 05 2016 15:33
@live4sw don't think anyone has measured it empirically, but from what we know about how the pump works, frequent basal changes shouldn't matter unless your basals are super low, and even then should mostly average out.
Matt Pazoles
@Pazoles
Jun 05 2016 15:33
Thanks @jasoncalabrese for that NS fix last night - finally got everything updated and deployed!
Brooke Pazoles
@PazolesBrooke_twitter
Jun 05 2016 15:43
This message was deleted
muffindad
@muffindad
Jun 05 2016 16:34
High guys, what are my options (if any) for using Dexcom G5 w/share to set up NightScout? On Phase 1 of OpenAPS
Matt Pazoles
@Pazoles
Jun 05 2016 16:36
Hi @muffindad - @PazolesBrooke_twitter and I use nightscout with a DexcomG5 w/share, you can get all the functionality except raw_bg (the readings that dexcom hides when there's noise in the data, and uses to create the smoothed line that's ultimately displayed)
muffindad
@muffindad
Jun 05 2016 16:37
Thanks Matt. Do you use xDrip or a simpler implementation w/ the cable and Uploader on android? I'm trying to get pointed the right direction before I buy anything else / spend anymore time
Matt Pazoles
@Pazoles
Jun 05 2016 16:38
We use the dexcom app on Brooke's iPhone via bluetooth, so the data already goes to the cloud. Nightscout can connect to the dexcom site and scrape the data from there
I think that xDrip is currently in a closed beta test for the G5
muffindad
@muffindad
Jun 05 2016 16:39
ah, makes sense. That seems pretty simple! Especially since I have an old iPhone I could use
Scott Leibrand
@scottleibrand
Jun 05 2016 16:39
you can also plug a G5 receiver directly into a Pi now with the latest openaps code
someone needs to document that
Dana Lewis
@danamlewis
Jun 05 2016 16:39
^ that needs to be reflected in the docs
Matt Pazoles
@Pazoles
Jun 05 2016 16:40
^ with raw_bg?
muffindad
@muffindad
Jun 05 2016 16:40
Good to know @scottleibrand Does my RPI impelemtation then read the NightScout website to do the closed loop? Provided it has an internect connection
Dana Lewis
@danamlewis
Jun 05 2016 16:43
@Pazoles no, re how to do G5 receiver plug in. I believe current language says something like "G5 is possible but with extra workarounds". Thanks to @bewest work, no extra hoop jumping needed for G5, but docs need to be updated.
Scott Leibrand
@scottleibrand
Jun 05 2016 16:43
@Pazoles no, the g5 receiver doesn't report raw data
Dana Lewis
@danamlewis
Jun 05 2016 16:44
@muffindad you can design your loop to use both local data (plugged in receiver if it's available) or from NS if Internet
Scott Leibrand
@scottleibrand
Jun 05 2016 16:44
only way to get raw from a g5 is to BLE pair directly to the transmitter like Nate did with the xDripG5 iOS app
and I'm not sure if he ever actually implemented those opcodes or not
Paul Martin
@Spazholio
Jun 05 2016 16:45
I still haven’t quite figured out how to get raw data from a connected G4. I’m not even sure if I need/want that data, but it’d be nice to know how. =)
Dana Lewis
@danamlewis
Jun 05 2016 16:46
@Spazholio it helps when you're waiting for 2 hour startup period. *however still requires you to remember to hit start on the sensor ;)
Also gives a clue of trends during ???
Paul Martin
@Spazholio
Jun 05 2016 16:47
Well, the last time I tried, all I managed to do was upload 24h of duplicate data. So I’m quite certain I did something wrong. =)
(I still can’t believe I forgot to start the sensor facepalm)
Matt Pazoles
@Pazoles
Jun 05 2016 16:54
Has anyone hooked up their TI stick to a raspberry pi via serial? Is uart0_alt1_SRF_ERF_US_STDLOC.hex the right firmware?
Dana Lewis
@danamlewis
Jun 05 2016 16:54
@AndyProbolus_twitter morning! Which script are you talking about?
Scott Leibrand
@scottleibrand
Jun 05 2016 16:55
@Pazoles also checkout the oskarpearson/mmeowlink channel for that kind of stuff
ryanchen26
@ryanchen26
Jun 05 2016 16:57
Is it covered in the docs what openaps does when the CGM data is disrupted by: 1. ???, 2. Share server is down(we are pulling from NS), 3. Out of range (data is particularly old), 4. No data due to calibration. Does it just snooze and run the preset basal schedule until CGM data is restored?
Matt Pazoles
@Pazoles
Jun 05 2016 16:58
@scottleibrand - Good call, thanks!
Lynne Butts
@flutelynne_twitter
Jun 05 2016 17:07
@danamlewis I was going to put that in my PR the other day, but when I posted that for someone else, ben corrected with another way. We were working on that around June 1 16:11 or so. Had started with the idea of using "openaps use <mypumpname> config --G5" So there probably is a better way, but that is around the G5 receiver conversation.
Scott Leibrand
@scottleibrand
Jun 05 2016 17:08
@ryanchen26 depending on the nature of the failure, it sometimes lets the last 30m temp basal run to completion, and sometimes cancels it. then it just lets your normally schedule basal run (doesn't set any temps) until it can loop again
ryanchen26
@ryanchen26
Jun 05 2016 17:11
Sounds reasonable, thanks.
Marianne Smith, PMP
@mariannesmith77_twitter
Jun 05 2016 18:20
@danamlewis @AndyProbolus_twitter is asking about
One helpful thing to do before starting any software work is to log your terminal session. This will allow you to go back and see what you did at a later date. This will also be immensely helpful if you request help from other OpenAPS contributors as you will be able to provide an entire history of the commands you used. To enable this, just run $ script <filename> at the beginning of your session. It will inform you that Script started, file is <filename>. When you are done, simply $ exit and it will announce Script done, file is <filename>. At that point, you can review the file as necessary.
Dana Lewis
@danamlewis
Jun 05 2016 18:23
Ah. Yes. Just make sure the file name includes whatever path you want, i.e. Where to save it to. @mariannesmith77_twitter @AndyProbolus_twitter
jaylagorio
@jaylagorio
Jun 05 2016 18:24
Dexcom users - I'm a bit new to this. Eventually, like all rechargeable batteries, I expect the receiver will stop holding a charge as well as it did when it was new and the battery will start to die sooner. What are your experiences with replacing it? Do you have to send the receiver in when they send you a new one? Or do they send you a new one and you throw the old one out?
Matt Pazoles
@Pazoles
Jun 05 2016 18:25
Working on AMA right now - what feeds into mealCOB?
Marianne Smith, PMP
@mariannesmith77_twitter
Jun 05 2016 18:25
Thanks @danamlewis
jaylagorio
@jaylagorio
Jun 05 2016 18:25
I have my first receiver and transmitter and I know when the transmitter dies they send a new one, but I don't know what to expect when the receiver gets old
Dana Lewis
@danamlewis
Jun 05 2016 18:25
It depends on how it dies @jaylagorio
And whether it's in warranty or not, etc
jaylagorio
@jaylagorio
Jun 05 2016 18:26
I think the warranty is a year or something
So if I want to replace it outside of the warranty period I assume I'd buy a new one - but in that case do they want it back? Or do they discount the new one for a returned old one?
Dana Lewis
@danamlewis
Jun 05 2016 18:30
I don't think I've heard of them doing that (buy back)
That's usually how ppl end up with multiple receivers
muffindad
@muffindad
Jun 05 2016 18:31
@jaylagorio I've been told recently that you can purchase another receiver every 12 months
jaylagorio
@jaylagorio
Jun 05 2016 18:31
Good to know. I'm in the "market" for a failing G5 receiver to continue adding capabilities to my Dexcom library. I certainly don't need one to try adding those capabilities given the source code already available online but I don't want to claim compatibility for anything I haven't tested myself.
muffindad
@muffindad
Jun 05 2016 18:31
@jaylagorio you can get new transmitter every 6
jaylagorio
@jaylagorio
Jun 05 2016 18:32
@muffindad I think that lines up with what I had heard, thanks for the confimation
Dana Lewis
@danamlewis
Jun 05 2016 18:33
@Pazoles mealCOB in AMA comes from entered carbs: it's measure as how many absorbed evidenced by BG rise aka deviation, and assumes a minimum of 3mg/dl/5min
Matt Pazoles
@Pazoles
Jun 05 2016 18:41
Hmm, I'm getting mealCOB: Null despite still having nightscout displaying COB
cat monitor/meal.json
{"carbs":56,"boluses":2.1,"mealCOB":null}
Screen Shot 2016-06-05 at 2.41.22 PM.png
Does this just need more data before it'll start producing values?
Dana Lewis
@danamlewis
Jun 05 2016 18:44
Did you set up meal assist en route to AMA? What have you done so far?
Lots of enabling and setup required
Dana Lewis
@danamlewis
Jun 05 2016 18:53
@PazolesBrooke_twitter for your carb test...did you have any net iob action? Ie had you done any temps in the hours preceding the test?
Milos Kozak
@MilosKozak
Jun 05 2016 19:03
@scottleibrand what is exact meaning of calculated iob activity?
Matt Pazoles
@Pazoles
Jun 05 2016 19:14
@danamlewis - Found the issue, was a -/_ mixup on my settings/profile.json report!
Screen Shot 2016-06-05 at 3.11.30 PM.png
and good call on the carb test. That definitely explains it.
Dana Lewis
@danamlewis
Jun 05 2016 19:16
Yea. Wrote that post way before we closed the loop, so need to make sure no open loop activity will impact it :)
Paul Martin
@Spazholio
Jun 05 2016 19:44
@bewest In your new oref0 templating tools, what does mint refer to?
Lynne Butts
@flutelynne_twitter
Jun 05 2016 19:54
@jaylagorio Usually you can get a one-time replacement for $199. Worth the investment.
Ben West
@bewest
Jun 05 2016 20:11
@Spazholio mint is a verb, meaning to stamp/imprint/create
printing presses take a template/mask and a medium, and press the medium into the template, which is how you mint a new thing
Paul Martin
@Spazholio
Jun 05 2016 20:12
Ok, I thought it might have been that. Just wasn’t 100%.
Ben West
@bewest
Jun 05 2016 20:12
has to do with coinage traditionally
openaps "trades" in config/devices/reports/uses, so this creates the blobs needed to represent them
thinking is we can have lots of template suites to cover all the different use cases
could probaby have used render instead of mint but mint is only 4 characters and easy to type
Scott Leibrand
@scottleibrand
Jun 05 2016 20:40
@MilosKozak insulin activity is the derivative of IOB: how much insulin is taking effect every 5 minutes. Multiply by ISF to get how much BG should be falling/rising from insulin activity.
(We call that BG impact, or BGI)
jaylagorio
@jaylagorio
Jun 05 2016 20:42
@flutelynne_twitter Thanks for the tip! I'm looking for one someone doesn't want anymore and would rather give it to me than throw it out. I don't need it to be fully functional, just functional enough to verify that my app works with it. I have a G4 receiver but some of the structures with the G5 receivers are different.
Milos Kozak
@MilosKozak
Jun 05 2016 20:43
that's what i'm thinking. i ported calculation from oref0 to android and it seems it calculates it wrong :( the number must be very small in comparation to iob
Ben West
@bewest
Jun 05 2016 20:44
for G5, do openaps use cgm config --G5 and then use openaps use cgm oref0_glucose --no-raw
Scott Leibrand
@scottleibrand
Jun 05 2016 20:44
@MilosKozak Check the time units. I don't remember if I do it per minute or per 5m or what.
Milos Kozak
@MilosKozak
Jun 05 2016 20:46
it looks like you split it not by time but to 0.05U chunks
Milos Kozak
@MilosKozak
Jun 05 2016 21:01
not sure if this is not too big because shorter temps could be less than 0.05U
ryanchen26
@ryanchen26
Jun 05 2016 21:15
For the medtronic pumps the manufacturer recommends only using alkaline batteries. Anyone using lithium-ion AAA?
ericmtx
@ericmtx
Jun 05 2016 21:31
trying to enable AMA. have created the carb_ratios.json report and data appears in the settings directory. But when running loop the call to settings/profile.json results in 'Namespace' object has no attribute 'carb_ratios'
oh, I think I need to add that to the profile.json report definition in the openaps.ini
Paul Martin
@Spazholio
Jun 05 2016 21:34
@bewest So, I’ve unplugged my Dexcom, and am attempting to rely solely on my NS data, but having some issues. The merged monitor/glucose.json seems to be intact, but I’m still having issues. Got a few seconds to help me debug?
@scottleibrand I’m noticing that auto-sens.json seems to run twice when my CGM is unplugged as well. Wouldn’t mind a second set of eyes on my workflow. I think I’ve got a “can’t see the forest for the trees” situation going on...
Also, when the CGM is plugged in, the auto-sens.json report shows twice the “symbols” as it does when it’s unplugged. That just seems weird.
Dana Lewis
@danamlewis
Jun 05 2016 21:40
@ryanchen26 you'll definitely want lithium for looping. Alkaline only gives 2-3 days of usage. Lithium gives a week+. No issues using the pump with them.
Paul Martin
@Spazholio
Jun 05 2016 21:42
And it’s changing my ISF from 20 to 30 with the CGM unplugged. That’s a far cry from the 22 I’m acheiving with it plugged in.
Scott Leibrand
@scottleibrand
Jun 05 2016 21:43
How many BG data points are you downloading from NS? 288?
it only appears to run twice because the same library is used by oref0-meal and autosens.
Paul Martin
@Spazholio
Jun 05 2016 21:44
Yes.
Scott Leibrand
@scottleibrand
Jun 05 2016 21:44
And how long is your pumphistory?
Paul Martin
@Spazholio
Jun 05 2016 21:44
(yes to the 288)
Hang on...
3 hours. But the 24h variant is 28.
Here’s the ENACT section of my loop:
STARTING ENACT
=++++---+=-+-----++-+-+++------+------+--++---+++-+++++++++++-+-----+-------++-------++---+---------------++-++++------+----+++-+ 60% of non-meal deviations negative (target 45%-50%)
Excess insulin sensitivity detected: ISF adjusted from 20 to 30
13 0 '2016-06-05T15:24:16-04:00'
=++++---+=-+-----++-+-+++------+------+--++---+++-+++++++++++-+-----+-------++-------++---+---------------++-++++------+----+++-+ 60% of non-meal deviations negative (target 45%-50%)
Excess insulin sensitivity detected: ISF adjusted from 20 to 30
0
{"carbs":13,"boluses":2.6,"mealCOB":0}
{"ratio":0.67}
BG data is too old, or clock set incorrectly.  Your CGM time is Sun Jun 05 2016 21:37:28 GMT-0400 (EDT) but your system time is Sun Jun 05 2016 17:41:42 GMT-0400 (EDT)
pump://JSON/iter_pump_hours/settings/pumphistory-24h.json
reporting settings/pumphistory-24h.json
tz://JSON/rezone/settings/pumphistory-24h-zoned.json
reporting settings/pumphistory-24h-zoned.json
clean-nightscout-glucose://JSON/shell/monitor/auto-sens-glucose.json
reporting monitor/auto-sens-glucose.json
auto-sens://text/shell/settings/auto-sens.json
reporting settings/auto-sens.json
meal://text/shell/monitor/meal.json
reporting monitor/meal.json
oref0-determine-basal://text/shell/enact/suggested.json
reporting enact/suggested.json
[{"msg":"BG data is too old, or clock set incorrectly.  Your CGM time is Sun Jun 05 2016 21:37:28 GMT-0400 (EDT) but your system time is Sun Jun 05 2016 17:41:42 GMT-0400 (EDT)"}]
No action required
ENACT COMPLETE
The times in my monitor/glucose.json look just fine though. Wondering how that error is getting invoked, since there’s no CGM.
Scott Leibrand
@scottleibrand
Jun 05 2016 21:59
What is the most recent glucose time?
Milos Kozak
@MilosKozak
Jun 05 2016 22:01
@scottleibrand i'm pretty sure activity is calculated wrong (debugged now in NS too). activity value is usually around 10% of iob. which is too much for 5 min effect
Paul Martin
@Spazholio
Jun 05 2016 22:02
"dateString": "2016-06-05T21:57:29-04:00”
That’s the top entry in monitor/glucose.json.
Milos Kozak
@MilosKozak
Jun 05 2016 22:04
i'm getting the same results from code ported from oref0
this is 2 hours old treatment
blob
Paul Martin
@Spazholio
Jun 05 2016 22:09
Your CGM time is Sun Jun 05 2016 22:02:28 GMT-0400 (EDT) but your system time is Sun Jun 05 2016 18:06:52 GMT-0400 (EDT)
If you subtract 4 hours from my CGM time, then it’s accurate. If you subtract 4 hours from my system time, then it’s 4 hours in the past. Or however that’s suppopsed to read - it’s 6:02pm my time. I feel like this is key to what’s going on, but I think everything is set correctly. Where can I verify that?
Scott Leibrand
@scottleibrand
Jun 05 2016 22:10
date?
Paul Martin
@Spazholio
Jun 05 2016 22:10
Your CGM time is Sun Jun 05 2016 22:02:28 GMT-0400 (EDT) but your system time is Sun Jun 05 2016 18:06:52 GMT-0400 (EDT)
Grrrrr
Sun Jun 5 18:10:21 EDT 2016
Scott Leibrand
@scottleibrand
Jun 05 2016 22:11
looks like you're zoning the glucose data wrong.
this is from NS?
Paul Martin
@Spazholio
Jun 05 2016 22:11
It is.
[report "raw-ns/ns-glucose-zoned.json"]
use = rezone
reporter = JSON
astimezone = True
date = timestamp dateString start_at end_at created_at
adjust = replace
input = raw-ns/ns-glucose.json                                                                                                                                                                                                                                                                                        
device = tz
timezone =
Do I need to have EDT in the timezone field? Some of them have it, some of them don’t.
Scott Leibrand
@scottleibrand
Jun 05 2016 22:12
What does the entry in ns-glucose look like?
Paul Martin
@Spazholio
Jun 05 2016 22:12
Zoned or non?
Scott Leibrand
@scottleibrand
Jun 05 2016 22:12
no, you should remove any explicit TZ so it'll use the current rig tz.
All of them?
Paul Martin
@Spazholio
Jun 05 2016 22:13

Non-zoned

[                                                                                                                                                                                                                                                                                                                     
  {
    "_id": "5754a2d94adc61af4af889b2",
    "sgv": 119, 
    "date": 1465164448000,
    "dateString": "2016-06-05T22:07:28.000Z",
    "trend": 4,
    "direction": "Flat",
    "device": "share2",
    "type": "sgv",
    "glucose": 119
  },

Zoned

[                                                                                                                                                                                                                                                                                                                     
  {
    "direction": "Flat",.
    "trend": 4,.
    "dateString": "2016-06-05T22:07:28-04:00",.
    "sgv": 119,.
    "device": "share2",.
    "date": 1465164448000,.
    "_id": "5754a2d94adc61af4af889b2",.
    "type": "sgv",.
    "glucose": 119
  },
Scott Leibrand
@scottleibrand
Jun 05 2016 22:14
That is zoning it wrong.
The input is already zoned (Z) so it shouldn't add a -0400 there.
If that input file is always from NS, you can use it directly instead of zoning it.
Paul Martin
@Spazholio
Jun 05 2016 22:15
Ahhhh, ok. Well, I posted how I’m zoning it up above - am I doing something incorrectly?
Oh. Hmm. Ok.
Scott Leibrand
@scottleibrand
Jun 05 2016 22:16
Because NS CGM data is always zoned.
unlike pump/CGM data.
Paul Martin
@Spazholio
Jun 05 2016 22:17
Ok, I’ve switched my monitor/glucose.json to merge with my raw-ns/ns-glucose.json instead of my raw-ns/ns-glucose-zoned.json. Let’s see what happens...
Matt Pazoles
@Pazoles
Jun 05 2016 22:19
Is there a way to sync up the COB value from openaps with the COB pill on nightscout?
Paul Martin
@Spazholio
Jun 05 2016 22:20
I wonder if there could be some sort of idiot-checking in the code that would recognize already-zoned data and not attempt to re-zone it?
Scott Leibrand
@scottleibrand
Jun 05 2016 22:21
@Spazholio I thought it already did that. Q for @bewest I guess.
Paul Martin
@Spazholio
Jun 05 2016 22:21
It could even output something like, “Sigh, we talked about this already Paul - you don’t need to zone this file…"
I’m really curious why the auto-sens.json seems to change my ISF more drastically than with data from the CGM.
Drat.
BG data is too old, or clock set incorrectly.  Your CGM time is Sun Jun 05 2016 22:17:28 GMT-0400 (EDT) but your system time is Sun Jun 05 2016 18:21:54 GMT-0400 (EDT)
Dana Lewis
@danamlewis
Jun 05 2016 22:23
@Pazoles wouldn't that be great if someone did that? :) ;)
Matt Pazoles
@Pazoles
Jun 05 2016 22:24
haha, hint taken
Dana Lewis
@danamlewis
Jun 05 2016 22:24
:) but yes, they are different and that's good to note
Scott Leibrand
@scottleibrand
Jun 05 2016 22:24
We do it already for the IOB pill, so that should show you how.
And @jasoncalabrese would be the one to answer questions if you get stuck.
Paul Martin
@Spazholio
Jun 05 2016 22:25
I think my date is still set wrong…somehow. Is there anything else I can check?
Matt Pazoles
@Pazoles
Jun 05 2016 22:27
:+1:
Scott Leibrand
@scottleibrand
Jun 05 2016 22:28
@Spazholio what does glucose.json have now?
you might need to delete old future-dated glucose entries.
Paul Martin
@Spazholio
Jun 05 2016 22:29
"dateString": "2016-06-05T22:22:28-04:00”
From NS?
The Admin Tools show that I don’t have any future dated entries (if that’s what you meant).
Scott Leibrand
@scottleibrand
Jun 05 2016 22:29
that (22:22) is still busted.
walk back through the process and see how it's getting future-zoned.
May need to delete all the files with mis-zoned data.
Paul Martin
@Spazholio
Jun 05 2016 22:31
The process is (as I recall) - I pull all the CGM data, zone it, then pull all the NS data, zone it (but don’t use that file any more), then use the select function to merge those two files, and that’s my monitor/glucose.json report.
Scott Leibrand
@scottleibrand
Jun 05 2016 22:33
Yeah, so look at all the intermediate files to see where that 22:22 is coming from.
Paul Martin
@Spazholio
Jun 05 2016 22:33
Here’s how I merge them:
[report "monitor/glucose.json"]
use = select
reporter = JSON
no_timezone = False
current = now 
date = dateString
gaps =
input = raw-ns/ns-glucose.json raw-cgm/glucose-zoned.json                                                                                                                                                                                                                                                             
device = tz
timezone =
prev =
Well, it’s in the raw-ns/ns-glucose.json file. I don’t have any further back to go from there.
However, the wife and I are going out for dinner, so I’ll have to resume this at a later time. Maybe @bewest would have an idea, as he was instrumental in helping me get this aspect of it set up? =)
ryanchen26
@ryanchen26
Jun 05 2016 22:38
In the case of meals that usually requires an extended-bolus (Pizza) does the OpenAPS system eliminate the need for an extended-bolus? Generally speaking? Of course everyone is different, and we'll do plenty of testing, I'm just looking for some insight into this aspect.
Dana Lewis
@danamlewis
Jun 05 2016 22:38
Yes, especially true with advanced meal assist (which requires regular entry of meal carbs in order to set it up)
Eating soon mode will also make a big difference for meals
Regardless of meal makeup
ryanchen26
@ryanchen26
Jun 05 2016 22:46
Any thoughts regarding how the OpenAPS system deals with (if it does) higher insulin resistance in the mornings? In the case of toddlers the recommendation is to have higher basal rate 1hr prior to the meal and continuing until 2hrs after. Is this method necessary (again, generally speaking) when using OpenAPS?
Dana Lewis
@danamlewis
Jun 05 2016 22:47
You can change your settings by time of day (isf or ratios or targets)
But in addition once you get set up, if those things are off and the BG rises, the loop will respond to that change
Scott Leibrand
@scottleibrand
Jun 05 2016 22:48
OpenAPS will use whatever basals and ratios you have set up, and then adjust
Dana Lewis
@danamlewis
Jun 05 2016 22:48
Loop essentially fine tunes on top of your settings, so you want baseline basals and ratios to be generally right
Ben West
@bewest
Jun 05 2016 22:48
there's nothing wrong with the zoning above
I think the oref0 scripts need some love to parse dates correctly
maybe bring in sugarjs to help with dates
Scott Leibrand
@scottleibrand
Jun 05 2016 22:48
there's also autosens, for days when you're running sensitive or resistant.
Ben West
@bewest
Jun 05 2016 22:49
Z is a valid zoned, the parsing needs to take it into account, same way the python tools do
Scott Leibrand
@scottleibrand
Jun 05 2016 22:49
@bewest it's not 22:00 on the east coast yet.
Ben West
@bewest
Jun 05 2016 22:49
it's Zulu time
UTC
perfectly valid
Scott Leibrand
@scottleibrand
Jun 05 2016 22:49
The problem is he zones a Z entry and it changes it to -0400
It's rezoning it that gets wonky.
Ben West
@bewest
Jun 05 2016 22:50
ah, it's using replace mechanism?
hmm, I use --astimezone on mine and it comes out fine
ryanchen26
@ryanchen26
Jun 05 2016 22:51
Sounds good. My wife and I are making a check list for this week and throwing out questions to the group. So far we are manually entering the openaps suggestions for basal rates and it's having a fantastic result. Autosens sounds interesting. Thanks for the feedback.
Ben West
@bewest
Jun 05 2016 22:52
ah ok
so I need a shift operation for --adjust
Dana Lewis
@danamlewis
Jun 05 2016 22:53
@ryanchen26 glad to hear!
Ben West
@bewest
Jun 05 2016 22:56
actually, no --astimezone works for me
Ben West
@bewest
Jun 05 2016 23:04
bewest@bewest-MacBookPro:~/Documents/foobar$ openaps use ns shell get entries.json count=1 | openaps use tz rezone --timezone EST -a missing       -
[ 
  { 
    "trend_arrow": "FLAT",
    "display_time": "2016-06-05T15:56:14-07:00",
    "_id": "5754af1ff28c031d44d213af",
    "system_time": "2016-06-05T21:53:24-07:00",
    "direction": "Flat",
    "dateString": "2016-06-05T15:56:14-07:00",
    "device": "openaps://cgm",
    "unfiltered": 150048,
    "rssi": 192,
    "date": 1465167374000,
    "filtered": 153248,
    "sgv": 123,
    "type": "sgv",
    "glucose": 123
  }
]bewest@bewest-MacBookPro:~/Documents/foobar$ openaps use ns shell get entries.json cunt=1 | openaps use tz rezone --timezone EST -a replace      -
replacing first tzoffset(None, -25200)
[ 
  { 
    "trend_arrow": "FLAT",
    "display_time": "2016-06-05T15:56:14-07:00",
    "_id": "5754af1ff28c031d44d213af",
    "system_time": "2016-06-05T21:53:24-07:00",
    "direction": "Flat",
    "dateString": "2016-06-05T15:56:14-05:00",
    "device": "openaps://cgm",
    "unfiltered": 150048,
    "rssi": 192,
    "date": 1465167374000,
    "filtered": 153248,
    "sgv": 123,
    "type": "sgv",
    "glucose": 123
  }
]bewest@bewest-MacBookPro:~/Documents/foobar$ openaps use ns shell get entries.json cunt=1 | openaps use tz rezone --timezone EST -a missing --astimezone      -
[ 
  { 
    "trend_arrow": "FLAT",
    "display_time": "2016-06-05T15:56:14-07:00",
    "_id": "5754af1ff28c031d44d213af",
    "system_time": "2016-06-05T21:53:24-07:00",
    "direction": "Flat",
    "dateString": "2016-06-05T17:56:14-05:00",
    "device": "openaps://cgm",
    "unfiltered": 150048,
    "rssi": 192,
    "date": 1465167374000,
    "filtered": 153248,
    "sgv": 123,
    "type": "sgv",
    "glucose": 123
  }
]bewest@bewest-MacBookPro:~/Documents/foobar$
make sure adjust = missing and astimezone = True
also you probably want no --timezone I'm providing one above to show the affect the other flags have (and because it's already in my system's zone)
if you use -a missing --astimezone it should work as expected
looks like you have -a replace
Paul Martin
@Spazholio
Jun 05 2016 23:14
I just want you to know that I am out at a restaurant right now, cannot implement these fixes, and it is KILLING me. :)
Ben West
@bewest
Jun 05 2016 23:14
:-) enjoy dinner
Jason Calabrese
@jasoncalabrese
Jun 05 2016 23:14
@Pazoles @danamlewis @scottleibrand I would like to use the ama COB in NS, but we need to add a field to suggested/enacted or include meal.json in the ns-status upload
Scott Leibrand
@scottleibrand
Jun 05 2016 23:15
would be easy enough to add it to the top level of suggest.json
@Spazholio careful, soon you'll be pulling out your phone and doing stuff via your ssh app. ;-)
Paul Martin
@Spazholio
Jun 05 2016 23:16
I can't. My reverse SSH script doesn't terminate properly when I switch wireless networks, so it's busted. :)
Jason Calabrese
@jasoncalabrese
Jun 05 2016 23:17
adding it to suggest makes sense, wouldn't be too hard to expose then
Brooke Pazoles
@PazolesBrooke_twitter
Jun 05 2016 23:49
Has anyone experienced the problem where a meal bolus is given through the pump, but the treatment entry that comes up on NightScout only shows the carb - no insulin? The treatment type is listed as carb correction. It usually says Meal Bolus when I do this.
Scott Leibrand
@scottleibrand
Jun 05 2016 23:50
@PazolesBrooke_twitter you need to do a bolusing check and only pull pumphistory if it's false
@jasoncalabrese and @Pazoles Added IOB and COB as top-level entries to suggested.json openaps/oref0@27603c9
Brooke Pazoles
@PazolesBrooke_twitter
Jun 05 2016 23:51
It just came through as a meal bolus just now. So it's double entered, but both from the pump
if it is still in process of bolusing, it won't have added the bolus to the record yet, but when it does, it backdates it
so when you go to check for more recent treatments, the bolus often doesn't show up
Brooke Pazoles
@PazolesBrooke_twitter
Jun 05 2016 23:53
Thanks @scottleibrand !