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

18th
May 2016
Scott Leibrand
@scottleibrand
May 18 2016 00:00
that is possible, if the CGM is plugged in or available via openxshareble
but not sure anyone has finished setting it up yet
Paul Martin
@Spazholio
May 18 2016 00:00
tweek_tweak__by_renaixxxx.png
Scott Leibrand
@scottleibrand
May 18 2016 00:01
@jasoncalabrese does oref0-raw using data uploaded from xDrip
Paul Martin
@Spazholio
May 18 2016 00:01
Well, I currently have 2 hours of zero data. Doesn’t my data start after the 2 hour calibration is up? I’m very, VERY new to the Dexcom, and this is my first swap.
Scott Leibrand
@scottleibrand
May 18 2016 00:01
the FDA-approved data starts then, yes
but there is raw uncalibrated data on the receiver accessible over USB or BLE even before your first calibration
so if you're willing to use the previous sensor's calibration values, you can get a halfway decent estimate of BG well before 2 hours
Paul Martin
@Spazholio
May 18 2016 00:02
Well that’s interesting. I wasn’t aware of that.
Scott Leibrand
@scottleibrand
May 18 2016 00:02
and the raw data trend tends to be pretty accurate, so IMO it's worth a try looping with that data
Dana Lewis
@danamlewis
May 18 2016 00:04
I am not confident to loop on raw data, but it is still nice to use it to buoy manual decision making during the otherwise 2 hour gap
Paul Martin
@Spazholio
May 18 2016 00:05
Yeah, that’d be an interesting thing to see, but I don’t know that I’d be looping on it quite yet.
Oh, I meant to ask earlier - where does OpenAPS determine upper and lower limits from? I’m having some lows at night, and would like to bump it up a bit. Do I have to set that on the pump? Can I set that in NS, and have OpenAPS use those limits, or…?
Scott Leibrand
@scottleibrand
May 18 2016 00:06
@jasoncalabrese care to share your experience with oref0-raw?
targets come from the pump
Dana Lewis
@danamlewis
May 18 2016 00:07
@spazholio it pulls your pump target settings, so change it on your pump
Paul Martin
@Spazholio
May 18 2016 00:07
Ok, I was pretty sure the pump was considered the “system of record” for something that important, but I wasn’t sure.
What’s weird is that my lower limit is set to 80, but I’ve been hovering around 70 overnight. Not bad, but still lower than my threshold. =/
Paul Martin
@Spazholio
May 18 2016 00:12
90, not 80.
Scott Leibrand
@scottleibrand
May 18 2016 00:14
If you're riding below target, you should have negative IOB. If that isn't sufficient to bring you up, that might be something autosens can help fix, or you might need to tune your basals.
Dana Lewis
@danamlewis
May 18 2016 00:15
But it's a good plan regardless to up your target of you don't like where you are riding
Paul Martin
@Spazholio
May 18 2016 00:15
Yeah, I definitely need to change something. Wondering if my ISF is different at night, or...
@danamlewis Agreed. My endo and primary have long said that I am WAY too aggressive with my numbers. I take it as a complement, but I’m not 100% sure that’s how they’re intending it. =)
Dana Lewis
@danamlewis
May 18 2016 00:16
;)
Paul Martin
@Spazholio
May 18 2016 00:28
Oh, if I wanted to uncomment/insert come debugging code into oref0, do I need to run some sort of npm command afterward to make it available?
Scott Leibrand
@scottleibrand
May 18 2016 00:38
If you did npm link earlier, maybe not.
That links the src/oref0 libraries so changes go live faster.
(As you make them.)
Paul Martin
@Spazholio
May 18 2016 00:39
I had a page up that had something like 4 different npm commands to run whenever you switch a branch, but I can’t seem to find it now. Do you happen to know what I’m taking about? It was npm install, there was a link thrown into the mix somewhere….
Scott Leibrand
@scottleibrand
May 18 2016 00:40
That was probably from quick-src.sh
Paul Martin
@Spazholio
May 18 2016 00:40
Boom. Got it, thanks.
Paul Martin
@Spazholio
May 18 2016 00:47
Ok, I switched to the AMA branch, and everything appears to be working just fine. And @scottleibrand - you weren’t kidding about auto-sens.json taking a bit longer.
Scott Leibrand
@scottleibrand
May 18 2016 00:49
In exchange, it's able to detect sensitivity/resistance about 4x more often.
Paul Martin
@Spazholio
May 18 2016 00:50
Well, that’s only assuming I can find the 900MHz static generator that someone’s installed in my home.
Seriously, you’d think I’d have more loop failures at work than home, but MAN...
Scott Leibrand
@scottleibrand
May 18 2016 00:50
Cordless phone? ;-)
Paul Martin
@Spazholio
May 18 2016 00:50
No way - first thing I thought of. =)
And I’m curious - once I switched to AMA, I notice that it says "Optional feature Meal Assist not configured.” in my loop. Do you know offhand what triggers that? What it’s looking for that it’s not finding that causes it to throw that out?
Scott Leibrand
@scottleibrand
May 18 2016 00:55
There are new things you have to provide for AMA to work.
It's telling you your old meal assist config isn't good enough.
Since you needed an inferiority complex. :-p
Paul Martin
@Spazholio
May 18 2016 00:56
Oh, I figured. I just wasn’t sure what it was looking for.
I have so much more to learn, just to be able to aspire to having an inferiority complex. =)
Dana Lewis
@danamlewis
May 18 2016 00:57
Lol
Scott Leibrand
@scottleibrand
May 18 2016 00:57
There are lots of hints in the -h output of the various tools, and in setup.sh
Paul Martin
@Spazholio
May 18 2016 00:58
Yeah, I’m reading up on setup.sh now.
Wait…which various tools? C’mon, just guide me in the general direction. =)
Scott Leibrand
@scottleibrand
May 18 2016 00:59
Like oref0-meal etc.
the ones your reports run.
Probably oref0-determine-basal too.
Look for new optional inputs at the end.
Paul Martin
@Spazholio
May 18 2016 01:01

Huh. I do have oref0-meal in an alias (ns-meal-carbs) but nothing calls that…yet.

While I love your setup.sh script, I really don’t want to overwrite what I already have, so I’mma pore over it a bit. =)

Scott Leibrand
@scottleibrand
May 18 2016 01:03
Yeah, that's what I prefer people do. :-)
Paul Martin
@Spazholio
May 18 2016 01:12
This might be well outside the scope of something like OpenAPS, but have you guys ever considered factoring protein into calculations?
Paul Martin
@Spazholio
May 18 2016 01:46
@bewest After updating to the AMA branch, there seems to be some oddness with the nightscout command. Is this…well, “expected” isn’t the right word, but...
JaysonEwer
@JaysonEwer
May 18 2016 02:24

Hello @all. Can anyone shed light on the difference between these two devices:
'process://oref0-determine-basal/oref0-determine-basal/' and
'process://determine-basal/oref0/determine-basal'
They appear to be the same to me:

pi@Ewer-Open-APS:~/Documents/openaps $ openaps use determine-basal -h
usage: openaps-use determine-basal [-h] USAGE ...

optional arguments:
  -h, --help  show this help message and exit

## Device determine-basal:
  vendor openaps.vendors.process

  process - a fake vendor to run arbitrary commands



  USAGE       Usage Details
    shell     run a process in a subshell
pi@Ewer-Open-APS:~/Documents/openaps $ openaps use oref0-determine-basal -h
usage: openaps-use oref0-determine-basal [-h] USAGE ...

optional arguments:
  -h, --help  show this help message and exit

## Device oref0-determine-basal:
  vendor openaps.vendors.process

  process - a fake vendor to run arbitrary commands



  USAGE       Usage Details
    shell     run a process in a subshell
pi@Ewer-Open-APS:~/Documents/openaps $

Am I going crazy, or did I just get myself a little confused?

Paul Martin
@Spazholio
May 18 2016 02:26
I think they’re the same. When I check my determine-basal.ini, I see:
[device "determine-basal"] fields = iob temp_basal glucose profile auto-sens meal cmd = oref0 args = determine-basal
[device "determine-basal"]                                                                                                                                                                                                                                                                                           
fields = iob temp_basal glucose profile auto-sens meal
cmd = oref0
args = determine-basal
Which - if I’m reading it right - means it’s just invoking oref0-determine-basal.
This is also assuming our determine-basal.ini files are the same. =)
JaysonEwer
@JaysonEwer
May 18 2016 02:28
This was a decision along the way, that I didn't keep up with. I think i'm beginning to wrap my head around this.
yeah, I think I've got things defined twice, Ive got some cleanup to do.
Paul Martin
@Spazholio
May 18 2016 02:28
Oh god. You’ve been doing this WAY longer than I have, and you’re just beginning to?! sobs
JaysonEwer
@JaysonEwer
May 18 2016 02:30
Well, I....foolishly went into maintenance mode, I can't keep up with all these young whippersnappers. :-D Upgrading code every 36 hours. ;-) So what I did know, i've slightly forgotten. :-D
Plus, I'm not sure what's WRONG WITH ME, but i've been feeling like CRAP, physically, since November of last year.:-/
I spend the better part of this afternoon in bed.
Paul Martin
@Spazholio
May 18 2016 02:31
I’m just waiting until they release one simple .sh that gets everything set up for me at once. Everything I need and want in one command. I’m sure someone will release that any day now.
Spazholio @Spazholio holds breath
Paul Martin
@Spazholio
May 18 2016 02:32
That kinda sorta sucks ass, man. November? Do the docs have any idea at all what it might be?
JaysonEwer
@JaysonEwer
May 18 2016 02:33
Yes, they're/ we-all-are ;-) moving in that direction...
:-/ No.
I'm beginning to wonder if i'm alergic to this platinum ROD stuck in my skin. :-D
Paul Martin
@Spazholio
May 18 2016 02:35
Assuming you’re talking about the CGM….it’s platinum?
JaysonEwer
@JaysonEwer
May 18 2016 02:35
part of it is
Paul Martin
@Spazholio
May 18 2016 02:37
…my wife’s a jeweler. I should keep that. Not for her to work with, but for her to throw in her gold dust bucket. Every few years, you ship that off to a refiner, and they give you money for the precious metals.
JaysonEwer
@JaysonEwer
May 18 2016 02:48
I"m trying to find concrete data on that...I am pretty sure it is part of the reaction that takes place....
Paul Martin
@Spazholio
May 18 2016 02:49
Yeah, I was looking for an MSDS or something on the sensors. Either way, my wife can take it to an old employer where they have this machine that will tell you exactly what percent of any kind of metal is in a given object.
It’s pretty sweet. It should be for how much it cost. =)
JaysonEwer
@JaysonEwer
May 18 2016 02:50
"Sensor Wire – The sensor wire is thin and flexible, and inserts just under your skin. It is attached to the sensor pod, and
is made of silver and platinum metal with polymer membranes. "
yep, it would be very cool to be able to salvage those tiny bits that are left at the top....I'm going to start saving mine now. :-D
you gave me a good idea
I gotta ge my loop backonline...its been down all day, :-/
Paul Martin
@Spazholio
May 18 2016 02:53
Oof. Have fun. And as hilarious as it may sound given my newness, let me know if I can help. =)
JaysonEwer
@JaysonEwer
May 18 2016 03:02
will do, thank you!
:-D
kamrausch
@kamrausch
May 18 2016 06:34
I am sure this question has been asked many times but what is special about the Medtronic pump currently being used in closed loop rig? Is anybody working on a similar rig for the omnipod?
Ben West
@bewest
May 18 2016 06:44
I spent six years to unlock the communications to medtronic
no one has yet similarly paid the toll for omnipod yet, thought they are working on it
my recommendation is to call vendor and ask for access
albertotrg
@albertotrg
May 18 2016 07:15
Hi there. This is Alberto from Spain, new with Nightscout and Computer Science Eng willing to help
@Ben West I got a Medtronic 640G. The communication now is different with 640G right?
I know you guys arr working with Contour Next Link (CNL). Wouldnt be easier to issue a bolus via CNL and sniff the communication pump<->CNL and try to decode this communication?
John Sjolund
@sjolundjohn
May 18 2016 07:52

Working through timezone issues now.

When pulling data from my Dexcom CGM, I note the following:
"system_time": "2016-05-18T07:46:50",
"display_time": "2016-05-18T09:45:11",

How / When does this get corrected to match the time on other devices?
John Sjolund
@sjolundjohn
May 18 2016 08:04
Does this sequence make sense? Note the two hour difference in time between timestamp and clock
pi@raspberrypi:~/openaps $ cat monitor/upload-status.json
{
  "device": "openaps://raspberrypi", 
  "pump": {
    "battery": {
      "status": "normal", 
      "voltage": 1.35
    }, 
    "status": {
      "status": "normal", 
      "timestamp": "2016-05-18T07:59:55.000Z", 
      "bolusing": false, 
      "suspended": false
    }, 
    "reservoir": 101.6, 
    "clock": "2016-05-18T09:58:59+02:00"
Oskar Pearson
@oskarpearson
May 18 2016 09:42
@dougtakeuti we are using nightscout these days. We upload 24 hours at a time.
What problem are you having?
John Sjolund
@sjolundjohn
May 18 2016 09:50
Hi - totally stuck with getting NS upload to work. I can get treatments uploaded (sometimes, I believe dependant upon sequence of pulling reports) and glucose never
Any tips on how to go about troubleshooting would be welcome
I plan to check in this evening (my time) when more people are online. Thanks for the patience. Good news is that open loop seems to work very well. Now just to get Nightscout upload and automation working.
Oskar Pearson
@oskarpearson
May 18 2016 10:17
Very strange. What errors / problems do you have?
Matt Walker
@walker0
May 18 2016 10:42
Do any of the US medtronic pumps have a setting to change to mmol/L or would I have to convert units?
John Sjolund
@sjolundjohn
May 18 2016 11:34
Hi @oskarpearson quite honestly I can't determine where problem is:
1) I can run the alias upload-recent treatments and status-upload and I do not get any errors on pi
2) Occasionally upload-recent-treatments does seem to update NightScout, however I have not been able to replicate sequence. I can confirm this because battery, clock and OpenAPS field is updated
3) I have never successfully gotten any glucose figures up to NightScout via OpenAPS
4) I have confirmed my Nightscout installation works, as I can update it via xDrip or via Medtronic 640g integration.
Colin Lennon
@colinlennon
May 18 2016 11:38
Hi @walker0 - my 722 is a US pump and has an option to use mmol/L (which I have selected as I'm in the UK so used to working in mmol/L)
OpenAPS uses mg/dL only, so you need to use units to convert
Paul Martin
@Spazholio
May 18 2016 11:42
Has anyone else ever had the Dexcom G4 be just…WAY off after inserting a new sensor? Last night - all night, my sensor kept reading things like 180-200 when I was actually closer to 130-150. No matter how many calculations I entered, it was still wonky. And this morning, it tells me I’m at 95, but in actuality I was at 65.
Oskar Pearson
@oskarpearson
May 18 2016 11:45
@sjolundjohn That's pretty strange. Which version of nightscout are you running?
John Sjolund
@sjolundjohn
May 18 2016 11:45
version 0.9.0-dev
Matt Walker
@walker0
May 18 2016 11:48
@colinlennon thanks, I'm happy to convert in openaps but I'd prefer to have pumps and bg monitors in mmol/l (Australian here)
John Sjolund
@sjolundjohn
May 18 2016 11:52
I just created a new pull of Nightscout so will update to the latest dev version as of today.
Colin Lennon
@colinlennon
May 18 2016 11:58
@walker0 Yeah, that's how I do it too. Let me know if you hit any problems. I had a few issues but got them resolved, so I should be able to help.
The only remaining issue is the BG values being uploaded to Nightscout. For example, if I enter a reading of 6.1 in the bolus wizard on the pump, it gets uploaded to Nightscout as 61. Haven't had a chance to try and fix that yet.
John Males
@johnmales
May 18 2016 12:05
@walker0 definitely can be done, that is what I am doing. Internally the medtronic pump is working with mg/dl for glucose readings. If you have the pump set to mmol/l then there are several conversions that need to be made from mmol/L to mg/dl for the purposes of looping
Specifically the bg-targets and insulin sensitivity reports need to have a conversion from mmol/L to mg/dl
Matt Walker
@walker0
May 18 2016 12:08
Yep that all makes sense, thanks
John Sjolund
@sjolundjohn
May 18 2016 12:10

@oskarpearson I even get the following, though it seems that latest event was much more recent, so I feel like that isn't getting updated correctly.

PREFLIGHT OK
Uploaded; most recent treatment event @ 2016-05-18T11:11:46+02:00

For the record time for me right now is 14:10

Hristina Dimova
@xpucuto
May 18 2016 12:11
@walker0 another option is simply to set the units in Azure/Heroku to be mmol . And you will easily see BG in mmol on your watch. But still stays the problem with viewing the cgm data in mg on the pump .
John Males
@johnmales
May 18 2016 12:17
@xpucuto If the pump is set to display mmol/L a few of the settings will be stored in the pump in mmol/L and will need to be converted as above
John Sjolund
@sjolundjohn
May 18 2016 12:29

Any ideas why:

openaps report add monitor/ns-glucose.json text ns-glucose shell

Is in the script for setting up Nightscout integration but not on the setup page docs?

I can't get this report setup
Hristina Dimova
@xpucuto
May 18 2016 12:39
@johnmales :+1:
John Sjolund
@sjolundjohn
May 18 2016 12:45
Oh man, I figured out most of the problem. When I was setting up the alias for the uploads, I had replaced $NIGHTSCOUT_HOST and $API_SECRET with my actual values. Seems that when I copy/paste as is, it works much better
Still not getting glucose up there though
Andrew Baugh
@baughaw
May 18 2016 14:02
Anyone have any thoughts on this new device? www.getchip.com 9 dollars seems like it might be worth exploring.
Oskar Pearson
@oskarpearson
May 18 2016 14:04
@sjolundjohn Are you cleaning up your glucose values with Nate's scripts?
John Sjolund
@sjolundjohn
May 18 2016 14:04
No - could you share?
Oskar Pearson
@oskarpearson
May 18 2016 14:04
Our command is ns-upload $NIGHTSCOUT_HOST $API_SECRET entries.json recent_glucose_mm.json
recent_glucose_mm.json is created with mm-format-ns-glucose --oref0 recent_glucose.json > recent_glucose_mm.json
John Sjolund
@sjolundjohn
May 18 2016 14:05
I am still having problem with ns-glucose, should what I pasted above work?
Oskar Pearson
@oskarpearson
May 18 2016 14:05
I actually misremembered. You shouldn't have to filter through Nate's scripts
John Sjolund
@sjolundjohn
May 18 2016 14:05
pi@raspberrypi:~/openaps $ openaps report add monitor/ns-glucose.json text ns-glucose shell
usage: openaps-report add [-h] report {base,text,stdout,JSON} device ...
openaps-report add: error: argument device: invalid choice: 'ns-glucose' (choose from u'calculate-iob', u'determine-basal', u'dexcom', u'get-profile', u'nightscout', u'ns-status', u'oref0', u'tz', u'units')
pi@raspberrypi:~/openaps $
Oskar Pearson
@oskarpearson
May 18 2016 14:05
Our recent_glucose.json is generated with:
[report "recent_glucose.json"]
device = pump
hours = 24.0
use = iter_glucose_hours
reporter = JSON
John Sjolund
@sjolundjohn
May 18 2016 14:21
@oskarpearson with the creation of recent_glucose_mm.json, where does it find/place the file?
I assume I should add upload/recent_glucose to both ends?
if I want it in upload folder?
ie:
mm-format-ns-glucose --oref0 upload/recent_glucose.json > upload/recent_glucose_mm.json
Strange results:
pi@raspberrypi:~/openaps $ cat upload/recent_glucose.json
[
  {
    "packet_size": 0, 
    "name": "DataEnd", 
    "date": "2016-05-18T14:01:00", 
    "date_type": "none", 
    "_tell": 1, 
    "op": "0x01"
  }, 
  {
    "packet_size": 4, 
    "name": "10-Something", 
    "raw": "4d-78-12-10", 
    "date": "2016-05-18T13:56:00", 
    "date_type": "minSpecific", 
    "_tell": 2, 
    "op": "0x10"
  }, 
  {
    "packet_size": 0, 
    "name": "DataEnd", 
    "date": "2016-05-18T09:56:00", 
    "date_type": "none", 
    "_tell": 7, 
    "op": "0x01"
  }, 
  {
    "name": "GlucoseSensorData", 
    "date_type": "prevTimestamp", 
    "_tell": 8, 
    "sgv": 0, 
    "date": "2016-05-18T09:51:00", 
    "packet_size": 0, 
    "op": 0
  }, 
  {
    "name": "GlucoseSensorData", 
    "date_type": "prevTimestamp", 
    "_tell": 9, 
    "sgv": 0, 
    "date": "2016-05-18T09:46:00", 
    "packet_size": 0, 
    "op": 0
  }, 
  {
    "packet_size": 4, 
    "name": "10-Something", 
    "raw": "49-69-12-10", 
    "date": "2016-05-18T09:41:00", 
    "date_type": "minSpecific", 
    "_tell": 10, 
    "op": "0x10"
  }
`
Andrew Baugh
@baughaw
May 18 2016 14:28
John are you using Enlite sensors?
C-Ville
@C-Ville
May 18 2016 14:29
blob
@/all had my first strange experience with the loop today. was well in range and dropping, however the loop proposed to increase basal, which in turn brought me close to hypo. Does anybody have an idea what I should look at to narrow down the problem.
John Sjolund
@sjolundjohn
May 18 2016 14:30
@baughaw Dexcom
ah, this is medtronic I guess
hmm, can't get this sorted
Andrew Baugh
@baughaw
May 18 2016 14:31
ah ok, the data you posted looked very similar to my Enlite data which can't be uploaded yet.
Dana Lewis
@danamlewis
May 18 2016 14:35
@C-Ville will want to start looking through logs to see what the decision making process was
C-Ville
@C-Ville
May 18 2016 14:39
@danamlewis will do that later today thanks
John Sjolund
@sjolundjohn
May 18 2016 14:49
Is there a suggested person I should reach out to to troubleshoot Nightscout upload?
I have most working, but lacking consitency AND getting glucose figure up there
Oskar Pearson
@oskarpearson
May 18 2016 15:24
@sjolundjohn Putting them in the upload directory is fine, yeah
@sjolundjohn the sgv keys there should be something in the 100s range if everything's going right
It's in mg/dl (US units)
Not sure why yours are zero
Are you in the USA or in the EU?
Are you perhaps running the unit conversion code when you shouldn't be?
JaysonEwer
@JaysonEwer
May 18 2016 15:28
@Spazholio regarding the Dexcom G4 being way off. Yes, it has happened, recently, but not often. I
Paul Martin
@Spazholio
May 18 2016 15:29
On a FB group, I was admonished for over-calibrating. I fingerstick a lot, so I was calibrating a lot. I was told that this is…not good. Limit it to twice a day.
Oskar Pearson
@oskarpearson
May 18 2016 15:30
@C-Ville That's not good. It's probably best to run the logs and extra output past @scottleibrand so he can figure out what's going on
@Spazholio That is what we were told by the Medtronic engineer too
JaysonEwer
@JaysonEwer
May 18 2016 15:30
Yes, that is something I tend to overdo as well. I've been checking, but not calibrating the Dexcom.
Oskar Pearson
@oskarpearson
May 18 2016 15:31
BTW: If you've got accuracy problems, Siobhan found things to be much more accurate when the medtronic sensor is on her arm, and it lasts a lot longer too. It loses signal a lot more though.
Paul Martin
@Spazholio
May 18 2016 15:31
I was told something different by my doc before I actually had the device, but given the preponderance of evidence and advice, I think I’ll stick to twice a day for a bit.
I’ve been considering the arm for my next sensor. While not “sanctioned” nearly everyone seems to do it and have no issues.
On the side of the arm, towards the back - correct?
C-Ville
@C-Ville
May 18 2016 15:34
@oskarpearson thanks for pointing me there @ceben80 suggested doing that too. I will try to sum that up as much as I can and then provide the information. I was a little surprised after having the loop running for about a week and seeing this for the first time. Only on day one where my programmed basal was way to high I realized something that looked like oscillation in the BG
Paul Martin
@Spazholio
May 18 2016 15:36
Does the loop take into account fingersticks when calculating tempp basals? Meaning, if my CGM is saying “160” but my fingerstick is “120” (and I don’t calibrate my CGM, no matter how wrong that feels) will it factor the fingerstick into the temp basal? Or is that just something to show on NS?
JaysonEwer
@JaysonEwer
May 18 2016 15:38
As far as I know, it only shows on NS. The decisions are made based on the CGM readings.
Dana Lewis
@danamlewis
May 18 2016 15:49
No, it doesn't take into account finger sticks. How would it know?
It only knows what you tell it ;)
JaysonEwer
@JaysonEwer
May 18 2016 15:53
Thanks @danamlewis. BTW, I think my auto-sens is working, but my BGI data is not populating the center column of the analysis of the data. I'm sure i'm missing something, but I"m wondering if you might know where I should look? Do I need a BGI report to be added to my auto-sens analysis?
Here is a recent enact:
May 18 11:43:38 Ewer-Open-APS openaps: reporting enact/enacted.json
May 18 11:43:39 Ewer-Open-APS openaps: {
May 18 11:43:39 Ewer-Open-APS openaps:   "bg": 135,
May 18 11:43:39 Ewer-Open-APS openaps:   "temp": "absolute",
May 18 11:43:39 Ewer-Open-APS openaps:   "snoozeBG": 105,
May 18 11:43:39 Ewer-Open-APS openaps:   "recieved": true,
May 18 11:43:39 Ewer-Open-APS openaps:   "rate": 1.5,
May 18 11:43:39 Ewer-Open-APS openaps:   "reason": "Eventual BG 105>100 but Avg. Delta -4.00 < Exp. Delta -2.3; setting current basal of 1.5 as temp",
May 18 11:43:39 Ewer-Open-APS openaps:   "eventualBG": 105,
May 18 11:43:39 Ewer-Open-APS openaps:   "timestamp": "2016-05-18T11:43:38.488937",
May 18 11:43:39 Ewer-Open-APS openaps:   "duration": 30,
May 18 11:43:39 Ewer-Open-APS openaps:   "mealAssist": "Off: Carbs: undefined Boluses: undefined Target: 100 Deviation: -11 BGI: -2.14",
May 18 11:43:39 Ewer-Open-APS openaps:   "tick": -1
May 18 11:43:47 Ewer-Open-APS openaps: }
Paul Martin
@Spazholio
May 18 2016 15:55
@danamlewis Well, the fingersticks are in the pump, and OpenAPS pulls info directly from the pump, so I figured the data was there so OpenAPS might work with it.
JaysonEwer
@JaysonEwer
May 18 2016 15:56
I see at the end of the mealAssist row the calcuated BGI, but that data is not making it into the following analysis:
May 18 11:43:22 Ewer-Open-APS openaps: p=0.60: 0.33, 0.00, 1.00
May 18 11:43:22 Ewer-Open-APS openaps: p=0.58: 0.33, 0.00, 0.77
May 18 11:43:22 Ewer-Open-APS openaps: p=0.56: 0.33, 0.00, 0.67
May 18 11:43:22 Ewer-Open-APS openaps: p=0.54: 0.00, 0.00, 0.33
May 18 11:43:22 Ewer-Open-APS openaps: p=0.52: 0.00, 0.00, 0.33
May 18 11:43:22 Ewer-Open-APS openaps: p=0.50: 0.00, 0.00, 0.33
May 18 11:43:22 Ewer-Open-APS openaps: p=0.48: -0.33, 0.00, 0.00
May 18 11:43:22 Ewer-Open-APS openaps: p=0.46: -0.33, 0.00, 0.00
May 18 11:43:22 Ewer-Open-APS openaps: p=0.44: -0.33, 0.00, 0.00
May 18 11:43:22 Ewer-Open-APS openaps: p=0.42: -0.67, 0.00, -0.33
May 18 11:43:22 Ewer-Open-APS openaps: p=0.40: -0.67, 0.00, -0.33
May 18 11:43:22 Ewer-Open-APS openaps: p=0.38: -0.67, 0.00, -0.33
May 18 11:43:22 Ewer-Open-APS openaps: p=0.36: -1.00, 0.00, -0.67
May 18 11:43:22 Ewer-Open-APS openaps: p=0.34: -1.33, 0.00, -0.67
May 18 11:43:22 Ewer-Open-APS openaps: p=0.32: -1.33, 0.00, -0.92
May 18 11:43:22 Ewer-Open-APS openaps: p=0.30: -1.50, 0.00, -1.00
May 18 11:43:22 Ewer-Open-APS openaps: p=0.28: -1.67, 0.00, -1.33
May 18 11:43:22 Ewer-Open-APS openaps: p=0.26: -2.00, 0.00, -1.33
May 18 11:43:22 Ewer-Open-APS openaps: Mean deviation: 0.76
May 18 11:43:22 Ewer-Open-APS openaps: Sensitivity within normal ranges
May 18 11:43:22 Ewer-Open-APS openaps: Basal adjustment 0.00U/hr
May 18 11:43:22 Ewer-Open-APS openaps: Ratio: 100%: new ISF: 45.0mg/dL/U
May 18 11:43:23 Ewer-Open-APS openaps: auto-sens://text/shell/monitor/auto-sens.json
May 18 11:43:23 Ewer-Open-APS openaps: reporting monitor/auto-sens.json
May 18 11:43:30 Ewer-Open-APS openaps: Optional feature Meal Assist not configured.
May 18 11:43:30 Ewer-Open-APS openaps: {"ratio":1}
May 18 11:43:30 Ewer-Open-APS openaps: {"delta":-1,"glucose":135,"avgdelta":-4}
May 18 11:43:30 Ewer-Open-APS openaps: {"duration":25,"rate":1.05,"temp":"absolute"}
May 18 11:43:30 Ewer-Open-APS openaps: {"netbasalinsulin":1.1,"activity":0.0095,"basaliob":0.415,"hightempinsulin":1.75,"bolussnooze":0,"iob":0.415}
May 18 11:43:30 Ewer-Open-APS openaps: {"max_bg":100,"dia":3,"min_bg":100,"max_iob":3,"sens":45,"max_daily_basal":2.75,"type":"current","current_basal":1.5,"max_basal":4.5}
May 18 11:43:31 Ewer-Open-APS openaps: determine-basal://text/shell/enact/suggested.json
May 18 11:43:31 Ewer-Open-APS openaps: reporting enact/suggested.json
May 18 11:43:31 Ewer-Open-APS openaps: {"temp":"absolute","bg":135,"tick":-1,"eventualBG":105,"snoozeBG":105,"mealAssist":"Off: Carbs: undefined Boluses: undefined Target: 100 Deviation: -11 BGI: -2.14","reason":"Eventual BG 105>100 but Avg. Delta -4.00 < Exp. Delta -2.3; setting current basal of 1.5 as temp","duration":30,"rate":1.5}
May 18 11:43:38 Ewer-Open-APS openaps: pump://JSON/set_temp_basal/enact/enacted.json
May 18 11:43:38 Ewer-Open-APS openaps: reporting enact/enacted.json
Oh, do I just need to get the rest of the meal assist setup?
Dana Lewis
@danamlewis
May 18 2016 16:06
@Spazholio you could change the algorithm to take them into account. But it doesn't today. (But yes, you could pull the data from your pump)
Paul Martin
@Spazholio
May 18 2016 16:07
@danamlewis Understood. And if I’m looking at NS properly, it doens’t display a dual-wave bolus as a dual-wave, does it? Just shows the “normal” part of the dual-wave?
I’m running an experiement with Keto Chow, so I’m trying to very, very accurately document everything. =)
Dana Lewis
@danamlewis
May 18 2016 16:14
It does not display (or calculate iob correctly) for dual waves
Scott Leibrand
@scottleibrand
May 18 2016 16:14
@C-Ville can you post a screenshot of the NS retrospective view from the time right after t set the high-temp? I suspect you had a lot of negative IOB, so when BG flattened out and upticked slightly it high-temped to get IOB back up.
Paul Martin
@Spazholio
May 18 2016 16:15
@danamlewis Excellent, thank you.
Scott Leibrand
@scottleibrand
May 18 2016 16:16
@C-Ville if you can include the OpenAPS pill tooltip that would also be helpful.
C-Ville
@C-Ville
May 18 2016 16:17
@scottleibrand is there a way to reactively make that visible?
Scott Leibrand
@scottleibrand
May 18 2016 16:19
@C-Ville If you ENABLE it, it will show up for retrospective data if the ns-status upload has been running.
If not, the same data should be in your logs and the git history.
JaysonEwer
@JaysonEwer
May 18 2016 16:34
Hello @all, Does anyone have a pointer to how to setup mealAssist? Or is it recommended that I jump to AMA at this point?
Dana Lewis
@danamlewis
May 18 2016 16:36
Start with meal assist, so you'll have many of the reports done for AMA
JaysonEwer
@JaysonEwer
May 18 2016 16:36
Ok, thanks.
is it documented, i'm not finding it anywhere?
Dana Lewis
@danamlewis
May 18 2016 16:42
It is not well documented despite me begging, I mean encouraging, the last 3 people who went through setup ;)
This script will help guide setup
JaysonEwer
@JaysonEwer
May 18 2016 16:43
AWESOME, i'll work on getting it documented!
Dana Lewis
@danamlewis
May 18 2016 16:43
please pr some English :) to help us add documentation
JaysonEwer
@JaysonEwer
May 18 2016 16:43
Thank you, @danamlewis
Dana Lewis
@danamlewis
May 18 2016 16:43
Yay :) thank you
JaysonEwer
@JaysonEwer
May 18 2016 16:43
pr some English...:-D
Dana Lewis
@danamlewis
May 18 2016 16:43
(Also if/when you get stuck, there's been several convos in gitter, so searching through here will be helpful)
Paul Martin
@Spazholio
May 18 2016 16:43
Once I get it set up properly and I’m confident in what the hell I did, I’m happy to assist with documenting. At this point though, it’s be like a monkey documenting how he put that rocket ship together. :laughing:
Dana Lewis
@danamlewis
May 18 2016 16:44
What do you think we did when building openaps @Spazholio ? :p
Paul Martin
@Spazholio
May 18 2016 16:44
That’s different! Somehow...
Dana Lewis
@danamlewis
May 18 2016 16:44
Uh huh
:p
JaysonEwer
@JaysonEwer
May 18 2016 16:45
@Spazholio Yes, I've felt the same way before, but i've now come full circle and any bit will help, or hurt, which is just more reason for more eyeballs and PR's of ENGLISH! :P
Paul Martin
@Spazholio
May 18 2016 16:45
Honestly though, I’m still finding cruft from when I went through the “official” docs vs @scottleibrand’s ssetup.sh vs @bewest’s usage of oref0. =)
JaysonEwer
@JaysonEwer
May 18 2016 16:46
POE's....PR's of ENGLISH....usage, make sure you POE that!!
Dana Lewis
@danamlewis
May 18 2016 16:46
All the more reason to PR. Official is just slightly more polished, but never perfect and always in need of additions and corrections to improve :)
(And if you* don't improve it, who will? :))
*you = each person reading this
Scott Leibrand
@scottleibrand
May 18 2016 16:48
openaps/docs#166
Any volunteers to fix that up if we merge it as is?
JaysonEwer
@JaysonEwer
May 18 2016 16:49
cool, thanks, Scott, i'll work from there.
Scott Leibrand
@scottleibrand
May 18 2016 16:50
or you can fork his repo, make your edits, and PR them to openaps/docs dev
JaysonEwer
@JaysonEwer
May 18 2016 16:50
@scottleibrand If you look at what I posted above, does it look correct, aside from not having the BGI info included?
with respect to my auto-sens, that is?
Paul Martin
@Spazholio
May 18 2016 16:52
Unrelated to the discussion here, but does anyone know if the ns device is still working if I’m on AMA? It’s not working, and the -h flag isn’t showing what it used to. @bewest?
Scott Leibrand
@scottleibrand
May 18 2016 16:53
@JaysonEwer you probably don't have enough pumphistory data being passed to autosens.
Dana Lewis
@danamlewis
May 18 2016 16:53
(I kidnapped @bewest today so he might be slow to respond for the next few hours. Sorry!)
JaysonEwer
@JaysonEwer
May 18 2016 16:54
28 hours
Paul Martin
@Spazholio
May 18 2016 16:54
No worries, I knew you were all out doing stuffs. =)
JaysonEwer
@JaysonEwer
May 18 2016 16:55
@scottleibrand I can see the BGI calculation on the one line, its just not being included in the analysis for some reason.
This output seems to indicate auto-sens is working, no?
May 18 12:58:37 Ewer-Open-APS openaps: Adjusting basal from 1.5 to 1.46
May 18 12:58:37 Ewer-Open-APS openaps: Adjusting sens from 45 to 46.4
Paul Martin
@Spazholio
May 18 2016 17:01
I hope so, because that’s what I’ve got right now. =)
(I envy that sensitivity, btw)
JaysonEwer
@JaysonEwer
May 18 2016 17:02
But your BGI's are being populated in the analysis, right? The center column?
Paul Martin
@Spazholio
May 18 2016 17:05
Well, don’t have a center column. I think the output’s different in AMA? Here’s what I have:
++++-+-++++++++-++++----------++-+++++++++--+----------++->>>>>++--=+----+--------+++-----+------++++++++-+++++++++++--+-+++++++++++++>++->+>>++-->>++-->>++->>>!-+!>>+-++++--++++->>>-+---+---++-+--+++++!!!!!!!!!!!!!!!!!!!!!!!!!!!!---+-+--------->>>>>+++++++++++++++-++++++++++++-++-+-+--+->++->>++++++-+++ 40% of non-meal deviations negative (target 45%-50%)
Excess insulin resistance detected: ISF adjusted from 20 to 17.857142857142854
JaysonEwer
@JaysonEwer
May 18 2016 17:09
Whew, I got it! .....I think. :_D
:-D
Paul Martin
@Spazholio
May 18 2016 17:11
Nice!
Hey, how much data does NS keep? Does it only keep ~24h worth and purge the older stuff from the DB, or does it only display the last ~24h, but I can query the DB if I want older stuff?
JaysonEwer
@JaysonEwer
May 18 2016 17:13

Except that I need to add the report to my loop. When I manually ran the report I got the following as I was watching syslog:

May 18 13:07:32 Ewer-Open-APS openaps: auto-sens://text/shell/monitor/auto-sens.json
May 18 13:07:32 Ewer-Open-APS openaps: reporting monitor/auto-sens.json
May 18 13:07:40 Ewer-Open-APS openaps: {"carbs":0,"boluses":0,"mealCOB":10}
May 18 13:07:40 Ewer-Open-APS openaps: {"ratio":0.97}

But on the next loop, It went back to the following:

May 18 13:10:28 Ewer-Open-APS openaps: reporting monitor/auto-sens.json
May 18 13:10:36 Ewer-Open-APS openaps: Optional feature Meal Assist not configured.
May 18 13:10:36 Ewer-Open-APS openaps: {"ratio":0.97}

Notice the lack of the cabs, boluses, mealCOB line.

Correct, the later.
I can run reports back to the first day I started using it in December.
Paul Martin
@Spazholio
May 18 2016 17:15
Ok, cool. I don’t know that I’d need to (or that I could - not a fan of MongoDB’s query structure) but it’s nice to know that I have it.
JaysonEwer
@JaysonEwer
May 18 2016 17:17
Are you using mongo, even though you are hosting the app yourself?
Paul Martin
@Spazholio
May 18 2016 17:26
Yes. All the install docs indicated MongoDB, so MongoDB is what I used. Why, is MySQL an option?
Paul Martin
@Spazholio
May 18 2016 17:32
Also - does anyone have a recommendation for a battery that’s small, can last at least a day, and supports passthrough charging? My battery is way too bulky...
JaysonEwer
@JaysonEwer
May 18 2016 17:32
Well, I knew you were just not wanting to use Azure...Wasn't sure you reasoning there, just wondering? :-)
Paul Martin
@Spazholio
May 18 2016 17:34
Well, I’m a sysadmin/programmer by trade, I already own my own server, and I’ve never been a fan of hosting solutions that can just up and go away or start charging (more) at any given moment.
JaysonEwer
@JaysonEwer
May 18 2016 17:35
yeah
I would imagine a different DB could be setup, but I haven't played with it myself.
Paul Martin
@Spazholio
May 18 2016 17:36
I know nothing about node, much less Mongo. I figure that’s one aspect I’ll likely just leave alone for now. =)
JaysonEwer
@JaysonEwer
May 18 2016 17:36
thats cool
benhamou
@benhamou
May 18 2016 18:15
Hello I am continuously struggeling with uploading my pump data: even I get the info that "openaps" was updated one minute ago, I cannot see the treatments. might that be due to my dump safari browser
dougtakeuti
@dougtakeuti
May 18 2016 18:59
@oskarpearson It's not a problem, but it is something I'd like to improve. I'd like to upload continuously while Julia is at school. The method I use is to read the last 30 min of glucose data from the pump and upload that set of data every time the loop runs. It works, but I believe this leads to azure reaching my quota/limit and stopping the site. What I'd like to do is to retrieve the last glucose data from NS, compare to pump glucose data and only upload what is new.
Colin Lennon
@colinlennon
May 18 2016 19:31
Getting an error when installing openaps on Edison (Yocto) -
easy_install -ZU openaps
This message was deleted
Installed /usr/lib/python2.7/site-packages/smmap-0.9.0-py2.7.egg
Finished processing dependencies for openaps
Error in atexit._run_exitfuncs:
Traceback (most recent call last):
  File "/usr/lib/python2.7/atexit.py", line 24, in _run_exitfuncs
    func(*targs, **kargs)
  File "/usr/lib/python2.7/multiprocessing/util.py", line 284, in _exit_function
    info('process shutting down')
TypeError: 'NoneType' object is not callable
Error in atexit._run_exitfuncs:
Traceback (most recent call last):
  File "/usr/lib/python2.7/atexit.py", line 24, in _run_exitfuncs
    func(*targs, **kargs)
  File "/usr/lib/python2.7/multiprocessing/util.py", line 284, in _exit_function
    info('process shutting down')
TypeError: 'NoneType' object is not callable
Error in sys.exitfunc:
Traceback (most recent call last):
  File "/usr/lib/python2.7/atexit.py", line 24, in _run_exitfuncs
    func(*targs, **kargs)
  File "/usr/lib/python2.7/multiprocessing/util.py", line 284, in _exit_function
    info('process shutting down')
TypeError: 'NoneType' object is not callable
C-Ville
@C-Ville
May 18 2016 19:47
@scottleibrand you are right it showed negative IOB, which was strange because at this time IOB from my meal was still positive. I just dumped my /var/log/messages with grep applied to it to an output file. I will try to extract the important part
@scottleibrand battery on the pump was low too, but I would assume that should not create weired behaviour apart from losing connection to the carelink
John Sjolund
@sjolundjohn
May 18 2016 20:15
openaps report add monitor/ns-glucose.json text ns-glucose shell $sgv_url
What is meant to go after shell for sgv_url?
Saskia
@Missdazzle
May 18 2016 20:53
@/all Stopped by a few days ago.... Now I want to show you something because it might be a good idea for other newbies. As everything is so new my brain just couldn´t cope with all the information thus I had to get it into something I can have in front of me and GRAB. This is my approach to playing around with the aliases:
20160518_114313.jpg
20160518_144137.jpg
Dana Lewis
@danamlewis
May 18 2016 20:54
Cool!
Saskia
@Missdazzle
May 18 2016 20:56
:)
Paul Martin
@Spazholio
May 18 2016 20:56
Nice! I tried doing this with a notepad and failed misrably. I think I need to make something like this. Maybe some extras so I can show dependencies for different reports.
John Sjolund
@sjolundjohn
May 18 2016 20:58
Happy to report that I was able to solve all open bugs on my end. Ended up making a new instance and using setup scripts + scripts as guides to get everything working
Saskia
@Missdazzle
May 18 2016 20:59
@Spazholio Well it´s so much easier to shuffle the pieces around...
John Sjolund
@sjolundjohn
May 18 2016 20:59

Now just need to be able to confirm cron is functioning correctly and in theory I should have a working system.

Have been running open loop during the days and working great :-)

Onwards and upwards (and off to bed)

Saskia
@Missdazzle
May 18 2016 20:59
@sjolundjohn Congrats to you!!!
Paul Martin
@Spazholio
May 18 2016 21:01
@sjolundjohn Impressive, man. Congrats!
@Missdazzle Definitely agreed. =)
John Sjolund
@sjolundjohn
May 18 2016 21:03
Thanks!
So.. any tips for how to confirm and figure out if cron is functioning well are appreciated. Currently I am looking at notes in /var/mail/pi - but it isn't telling the full story I want of confirming loop ran properly
Matt Pressnall
@logichammer
May 18 2016 21:11
did I see that @bewest created an oref0 command to verify that a pump was compatible? Looking in all the branches and can't find anything...but I'm a github n00b.
Paul Martin
@Spazholio
May 18 2016 21:11
@sjolundjohn I have everything tee’d to /var/log/openaps/loop.log and I just tail that so I can see.
Dana Lewis
@danamlewis
May 18 2016 21:23
@logichammer yes but i haven't seen verification that it works. Might be quickest to start -h'ing your way around to see if you can find it. And points for documenting if you verify the existence ;)
Matt Pressnall
@logichammer
May 18 2016 21:25
with 722 and firmware 2.4a, I should be good...having trouble talking to a pump right now...need to see what's what
frustrating
Dana Lewis
@danamlewis
May 18 2016 21:25
Yea no questions there
Matt Pressnall
@logichammer
May 18 2016 21:28
working now...will see if I can track down that util!
Dana Lewis
@danamlewis
May 18 2016 21:30
:+1:
Ben Kamens
@kamens_twitter
May 18 2016 22:09

Just dropping in to send deep thanks to all. Currently regularly/manually triggering closed loops at work, have mobile rig but haven’t turned on cron yet. Coming soon.

Am extremely humbled by and grateful for all the work done by others here. I’ve seen enough to know that I’m only aware of the tip of the iceberg w/r/t this effort.

tazitoo
@tazitoo
May 18 2016 22:11
@logichammer - just popping in from work so haven't followed your thread - but checking the 'temp_basal_type' on your pump is 'absolute' and not 'percentage' (that tripped me up for a bit)
Dana Lewis
@danamlewis
May 18 2016 22:21
@kamens_twitter congrats on the progress you've been making. And thank you :)
Ben Kamens
@kamens_twitter
May 18 2016 22:22
:thumbsup: :)
Dana Lewis
@danamlewis
May 18 2016 22:22
@sjolundjohn kudos on the progress! :)
Scott Leibrand
@scottleibrand
May 18 2016 22:23
@C-Ville I think I see what happened prior to your low. Looks like the system had been low-temping for quite awhile after your 10U bolus 3 hours earlier, then BG flattened out when it should have been falling (based on insulin activity). as a result, the predicted eventual BG (based on that deviation of 39 and a net IOB of 0.4U) was 124 at 6:31am, so it decided to high-temp slightly. at the time it did so, basal IOB (the effect of OpenAPS) was -0.9U, and over the next half hour it delivered enough extra insulin to bring that up to -0.5U (but net IOB actually dropped to -0.3 as the boluses finished decaying). by that time, BG had started dropping again, so it zero-temped again to bring basal IOB and net IOB down to -0.9 before you started rising again (presumably from a carb correction).
so it looks like the low was actually caused by the tail end of your 10U meal bolus, after the 100g carbs had finally stopped absorbing. OpenAPS had been low-temping to help prevent that, but its -0.9U worth of low temps was insufficient to prevent a low, just to help keep it from being worse.
the high-temps were unfortunate, but did not end up reversing the net effect of openAPS, which remained at a net-negative IOB throughout.
Scott Leibrand
@scottleibrand
May 18 2016 22:32
@C-Ville in terms of what we can do to help next time, I agree with your intuition that improving connectivity would help a lot. that would allow OpenAPS to cancel high-temps and set low-temps faster as BG starts to drop. In addition, you might find that setting up meal assist (or AMA) and doing a smaller meal bolus might help a lot. OpenAPS can do the rest of the meal insulin over time, but if BG isn't rising as much as expected, it can forego that, or even switch to low-temping instead, to help prevent post-meal lows.
C-Ville
@C-Ville
May 18 2016 22:41
@scottleibrand thanks for the quick analysis. Unfortunately my meal values may be off quite a bit as I just switched to a pump last week and only got back to meals at work on tuesday this week. I swapped batteries and will check on AMA, for now I was just happy that I got the loop up and running. I was not so much surprised that the loop could not save me from a hypo when over bolusing, but that it started even providing insulin. And yes you are correct I had some glucose to bring me back up. Thanks again.
Scott Leibrand
@scottleibrand
May 18 2016 22:46
Sounds good. Keep bringing these kinds of things up as you see them. It's great to analyze these and understand where we might need to keep a close eye on the algorithm.
Dana Lewis
@danamlewis
May 18 2016 22:49
:+1:
JaysonEwer
@JaysonEwer
May 18 2016 22:57
I can't believe it took me as long as it did to get mealAssist and Autosens setup. They are HUGELY helpful!! They weren't difficult and have made a very positive difference today. 👏👏
Dana Lewis
@danamlewis
May 18 2016 22:58
Woot @JaysonEwer ! So glad to hear!
Paul Martin
@Spazholio
May 18 2016 23:01
:thumbsup:
Now do AMA! :grin:
Matt Pressnall
@logichammer
May 18 2016 23:14
go @JaysonEwer ! And like @Spazholio said: go AMA! 1st draft of documentation coming by Friday...you can run through it and tell me what I missed...
Dana Lewis
@danamlewis
May 18 2016 23:14
Woohoo!
Paul Martin
@Spazholio
May 18 2016 23:27
If I’m consistently seeing my ISF adjusted from 20 to ~18, after a few days would it make sense to possibly consider changing the setting on my pump? I’ve been wondering if my ISF has been set incorrectly for quite some time, and this might be indicating that it could be...
Dana Lewis
@danamlewis
May 18 2016 23:28
I would give it some more time. But yes it might affirm your intuition and other observational tools that suggest making a change.
eyim
@eyim
May 18 2016 23:28
@scottleibrand @C-Ville That is similar to what I see sometimes as well and tried to describe in #101. I'm running auto sens but luckily haven't had to deal with a low like that recently
Scott Leibrand
@scottleibrand
May 18 2016 23:34
@Spazholio be aware that autosens can't tell whether it's ISF or basals that need adjusting.
Paul Martin
@Spazholio
May 18 2016 23:34
@danamlewis Yeah, I’m going to wait a bit. I’m noticing that I’m hovering lower than I’d like even after changing the targets on my pump, which is also contributing to my thinking that something is slightly awry.
@scottleibrand Understood.
I’m still (re)learning how protein affects me nowadays, so there’s lots of variables present.
Matt Pressnall
@logichammer
May 18 2016 23:36
silly question: in the ini files, can you comment things out? I want to drop different pump serials in easily
Paul Martin
@Spazholio
May 18 2016 23:36
I second @logichammer’s question. =)
Scott Leibrand
@scottleibrand
May 18 2016 23:37
Unfortunately you can't rely on comments to stick around when the ini gets rewritten.
Matt Pressnall
@logichammer
May 18 2016 23:37
ahhh...good point...separate file then
Paul Martin
@Spazholio
May 18 2016 23:38
Oh, that brings up a good point - when the loop runs, what is getting overwritten in openaps.ini?
Matt Pressnall
@logichammer
May 18 2016 23:41
I don't think anything does while looping, Paul...but pump.ini get overwritten I think every time you call model
openaps.ini should only get overwritten when adding / modifiying elements but in a readonly state like execute reports, I do not believe it writes to the file
William Fallon
@inform880
May 18 2016 23:46
This message was deleted
dougtakeuti
@dougtakeuti
May 18 2016 23:48
Hi, somehow I can't find the link to the description of meal-assist algorithm. Could someone point me to it please?
eyim
@eyim
May 18 2016 23:49
@logichammer In my openaps.ini file I have two pump in there - pump and pump2 and I put them right next to each other. When I switch pumps, I simply edit the openaps.ini file and switch the names of the pumps so pump2 becomes pump and pump becomes pump2
Matt Pressnall
@logichammer
May 18 2016 23:49
ahhhh...good idea!
JaysonEwer
@JaysonEwer
May 18 2016 23:50
@logichammer thanks for the first draft, I'll jump into the AMA stuff a little later. I still have a few notes to add to th meal.json creation and I am still intrigued by my BMI not being present when the 28 hours of pump history are processed. Somethings not 100% correct there.
William Fallon
@inform880
May 18 2016 23:50
I'm trying to run openaps on a machine loaded with debian jessie, and I'm stuck at setting up my TI stick. I installed everything necessary, but it says it can't find dev/mmeowlink. I found that the stick was dev/ttyACM0, but when I tried to use that directly, it said I didn't have permission. Anybody have any ideas? I know it works somewhat cause I manually set permissions for dev/ttyACM0, but then it tells me this:
serial.serialutil.SerialException: [Errno 2] could not open port /dev/mmeowlink: [Errno 2] No such file or directory: '/dev/mmeowlink'
Also, whenever I use chmod, when I restart my system or unplug the TI stick, it just goes back to the previous permissions
Paul Martin
@Spazholio
May 18 2016 23:51
@logichammer When my loop is running, if I’m editing openaps.iniand save it, vim yells at me and tells me that the file has changed since I opened it.
@inform880 What are the permissions on /dev/ttyACM0?
Matt Pressnall
@logichammer
May 18 2016 23:52
Ahhh...didn't know that...I'd do a git diff to see what is changing, then
I'll take a look a bit later to see what is changing
Paul Martin
@Spazholio
May 18 2016 23:52
Mine (on a Pi3) are crw-rw-r-- 1 root plugdev 166, 0 May 18 19:50 /dev/ttyACM0
@logichammer While I’m obviously in a git repo, I’m not doing ANYTHING with it. No committing. I should probably start doing that one of these days. =)
JaysonEwer
@JaysonEwer
May 18 2016 23:54
Yeah, same here, I'm sure it would be very beneficial in the long run.
William Fallon
@inform880
May 18 2016 23:55
@Spazholio
ls -l ttyACM0
crwxrwxrwx. 1 root root 166, 0 May 18 19:39 ttyACM0
Paul Martin
@Spazholio
May 18 2016 23:57
Wait, I just re-read your error message. That indicates that /dev/mmeowlink doesn’t exist. What happens when you type ls -l /dev/mmeowlink?
William Fallon
@inform880
May 18 2016 23:58
It doesn't exist. My rules for my usb devices are not working for some reason.
Paul Martin
@Spazholio
May 18 2016 23:59
By “rules” you mean the udev rule you set up?