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

19th
Sep 2016
Scott Leibrand
@scottleibrand
Sep 19 2016 00:20
Also, can't remember who was wanting G5 support in oref0-setup.sh and suggested openaps use cgm config --G5, but I just went ahead and added that. Please test and comment on openaps/oref0#216
To test, run cd oref0 && git checkout oref0-setup && git pull before running oref0-setup.sh
Russell Gardner
@RussGG
Sep 19 2016 00:24
I’m going through my reports and aliases to see what work and what doesn’t and I see an issue with the time being off somewhere.
pi@raspberrypi:~/ccgopenaps2 $ openaps do-oref0 get-profile://text/shell/oref0-monitor/profile.json reporting oref0-monitor/profile.json calculate-iob://text/shell/oref0-monitor/iob.json reporting oref0-monitor/iob.json determine-basal://text/shell/oref0-predict/oref0.json BG data is too old, or clock set incorrectly Sun Sep 18 2016 16:54:57 GMT-0700 (PDT) vs Sun Sep 18 2016 17:11:38 GMT-0700 (PDT) reporting oref0-predict/oref0.json
Any thoughts on where I should be checking?
Scott Leibrand
@scottleibrand
Sep 19 2016 00:25
the first timestamp is from your latest CGM glucose reading. the second is your Pi's system time.
so your CGM time might be off, or you might not be retrieving new CGM data
Russell Gardner
@RussGG
Sep 19 2016 00:26
I’m only pulling from NS
Scott Leibrand
@scottleibrand
Sep 19 2016 00:27
is the NS CGM data current? are you running the report to download BG data from NS right before running the oref0-predict/oref0.json report?
Russell Gardner
@RussGG
Sep 19 2016 00:29
I had just changed my monitor/glucose.json report to pull from NS and ran that a few minutes before
Scott Leibrand
@scottleibrand
Sep 19 2016 00:30
check the data in the file
Russell Gardner
@RussGG
Sep 19 2016 00:32
```pi@raspberrypi:~/ccgopenaps2 $ openaps report add monitor/glucose.json JSON ns shell get entries.json
added ns://JSON/shell/monitor/glucose.json
pi@raspberrypi:~/ccgopenaps2 $ openaps report invoke monitor/glucose.json
ns://JSON/shell/monitor/glucose.json
reporting monitor/glucose.json
pi@raspberrypi:~/ccgopenaps2 $ cat monitor/glucose.json
[
{
"direction": "SingleUp",
"trend": 2,
"dateString": "2016-09-18T23:54:57.000Z",
"sgv": 219,
"device": "share2",
"date": 1474242897000,
"_id": "57df29df822c8ba7b135689a",
"type": "sgv"
},
{
"direction": "FortyFiveUp",
"trend": 3,
"dateString": "2016-09-18T23:49:56.000Z",
"sgv": 206,
"device": "share2",
"date": 1474242596000,
"_id": "57df28b4822c8ba7b1356892",
"type": "sgv"
},
{
"direction": "FortyFiveUp",
"trend": 3,
"dateString": "2016-09-18T23:19:57.000Z",
"sgv": 156,
"device": "share2",
"date": 1474240797000,
"_id": "57df21ab822c8ba7b1356865",
"type": "sgv"
},
{
"direction": "Flat",
"trend": 4,
"dateString": "2016-09-18T23:14:57.000Z",
"sgv": 146,
"device": "share2",
"date": 1474240497000,
"_id": "57df207f822c8ba7b135685f",
"type": "sgv"
},
{
"direction": "Flat",
"trend": 4,
"dateString": "2016-09-18T23:09:55.000Z",
"sgv": 136,
"device": "share2",
"date": 1474240195000,
"_id": "57df1f53822c8ba7b1356858",
"type": "sgv"
},
{
"direction": "Flat",
"trend": 4,
"dateString": "2016-09-18T23:04:54.000Z",
"sgv": 130,
"device": "share2",
"date": 1474239894000,
"_id": "57df1e27822c8ba7b135684f",
"type": "sgv"
},
{
"direction": "Flat",
"trend": 4,
"dateString": "2016-09-18T22:54:55.000Z",
"sgv": 134,
"device": "share2",
"date": 1474239295000,
"_id": "57df1bcf822c8ba7b135683f",
"type": "sgv"
},
{
"direction": "Flat",
"trend": 4,
"dateString": "2016-09-18T22:49:55.000Z",
"sgv": 134,
"device": "share2",
"date": 1474238995000,
"_id": "57df1aa3822c8ba7b1356838",
"type": "sgv"
},
{
"direction": "Flat",
"trend": 4,
"dateString": "2016-09-18T22:45:15.000Z",
"sgv": 134,
"device": "share2",
"date": 1474238715000,
"_id": "57df1977822c8ba7b1356831",
"type": "sgv"
},
{
"direction": "Flat",
"trend": 4,
"dateString": "2016-09-18T22:19:57.000Z",
"sgv": 115,
"device": "share2",
"date": 1474237197000,
"_id": "57df139b822c8ba7b135680b",
"type": "sgv”
````
Scott Leibrand
@scottleibrand
Sep 19 2016 00:32
"dateString": "2016-09-18T23:54:57.000Z",
that's 35m ago
Russell Gardner
@RussGG
Sep 19 2016 00:35
I guess it seemed like I did right before.
Scott Leibrand
@scottleibrand
Sep 19 2016 00:36
you should have the report that refreshes that file running in the alias that does everything else
Russell Gardner
@RussGG
Sep 19 2016 00:40
This is for my son and he is at his moms right now with not very good internet. Right now on NS the last info is from 19 minutes ago. That must be the issue and not any of the devices
On the oref0-setup …...I received errors on step 1. I tried at my home directory and then tried it in my current openaps directory.
garykidd
@garykidd
Sep 19 2016 00:45
@scottleibrand does your new script accommodate both cgm and NS for BG data.....does it do both at same time...or do you switch?
Scott Leibrand
@scottleibrand
Sep 19 2016 00:52
it supports both: will use CGM data if connected, otherwise NS
@RussGG "errors" isn't very specific. :)
Russell Gardner
@RussGG
Sep 19 2016 00:53
pi@raspberrypi:~ $ cd ~/src && git clone -b dev git://github.com/openaps/oref0.git || (cd oref0 && git checkout dev && git pull) -bash: cd: /home/pi/src: No such file or directory -bash: cd: oref0: No such file or directory
I figured…. if I can’t get past step 1 I should go back to figuring out how to change my reports to pull from NS
Scott Leibrand
@scottleibrand
Sep 19 2016 01:10
You missed the first part (the mkdir)
Tim Mecklem
@tmecklem
Sep 19 2016 01:17
Got the new edison + srf + adafruit lipo charger open looping using BG from nightscout alongside the current closed loop pi3.
What’s a good capacity battery for the edison? We’ve got a 2300 mAh right now.
Russell Gardner
@RussGG
Sep 19 2016 01:19
I don’t see that, but that makes much more sense. Should I make a directory 'src' or does the command create it?
David
@davidkeddydb
Sep 19 2016 01:19
Hey everyone, I received a new error tonight on openaps do-everything. Getting an error on preflight: model.json raised ACK is 0 bytes: ? Any idea what is causing this?
Dana Lewis
@danamlewis
Sep 19 2016 01:29
Are you in range?
David
@davidkeddydb
Sep 19 2016 01:30
@danamlewis Yes, pump is almost sitting on top of the carelink usb
I have not hooked up yet, I am still running openloop not attached to the 722
@danamlewis unplugged carelink and replugged and voila
Dana Lewis
@danamlewis
Sep 19 2016 01:33
:+1:
Russell Gardner
@RussGG
Sep 19 2016 01:34
@scottleibrand I followed your link and it is differnt than I had. I ran the comand and it installed! Wahoo!!! I just entered my NS azure site and the next line said "no nightscout for you” but then is asking for my API secret. Are you guys big Seinfeld fans and I should just keep going?
Tim Mecklem
@tmecklem
Sep 19 2016 01:39
Maybe the dingo ate your bg
Russell Gardner
@RussGG
Sep 19 2016 01:40
@tmecklem I really laughed out load when I read that!!
Tim Mecklem
@tmecklem
Sep 19 2016 01:40
same for the “no nightscout for you"
:smile:
Russell Gardner
@RussGG
Sep 19 2016 01:52
@scottleibrand I think I’m going to cry! I think I just did (well…..you did) in 2 minutes what I have been working on for a long time. THANK YOU!!
Scott Leibrand
@scottleibrand
Sep 19 2016 01:54
@RussGG hmm, you should only get "no NS for you" if you don't enter a NS host.
Did it end up setting NS up anyway?
Scott Leibrand
@scottleibrand
Sep 19 2016 02:02
Ok, that is just a display bug. I bet you can spot it at https://github.com/openaps/oref0/blob/dev/bin/oref0-setup.sh#L120
Scott Leibrand
@scottleibrand
Sep 19 2016 02:09
Fixed in the oref0-setup branch: openaps/oref0@4cbcb5a
Iain
@iainct
Sep 19 2016 02:24
Offline Xdrip to Openaps integration question: I had been carrying out some testing with G5-Xdrip, last night my G5 receiver developed the ERR121 fault, speaking to Dexcom it's a replace the receiver situation (i paper clip reset the bloody thing about 20 times...) I am now pulling G5-Xdrip data from Nightscout and have my Openaps loop up and running (hotspot on Nexus5x and pi-zero with Ti stick pulling cgm data from Nightscout), which is all well and good until I go out of cell tower range. Is there a method of generating/updating a glucose.json on my Raspberry Pi without the internet?? I am on the road at the moment terrible timing but I am pleased I have something in place to keep me looping at night. X-Drip is working awesome with my Nexus5x. So good in fact it was troublesome to temporarily disable it! (I changed the G5 Transmitter serial in the end).
Thanks for any assistance!
Russell Gardner
@RussGG
Sep 19 2016 02:42
@scottleibrand My site and secret are in there. Now should I follow the directiions starting at "running an open loop with oref0”?
Scott Leibrand
@scottleibrand
Sep 19 2016 02:54
Cool. Yeah, that sounds right. If you didn't specify differently, the oref0-setup script will set up a closed loop with max_iob 0, which basically means low glucose suspend (plus the ability to high-temp slightly to counteract any low temps when they're no longer needed). We haven't built any open loop stuff into it yet, so you can do that part manually.
@iainct the loop set up by oref0-setup will do that (pull glucose.json from a receiver plugged in via USB when available, and use NS otherwise). There is also a method in the openaps/docs wiki for actually uploading to NS from the rig.
@iainct if you want to use oref0-setup.sh with G5, that's now supported and just needs tested. see https://github.com/openaps/oref0/wiki/Setup-Script-Testing for details
Iain
@iainct
Sep 19 2016 03:22
@scottleibrand I am successfully pulling glucose data from night scout (generated by G5-Xdrip Beta), but I am trying to figure out if it is possible to avoid using the Internet when I am out of cell reception, as my G5 Receiver is broken, I would normally be connected via USB. I have not found anything on Xdrip to Openaps integration other than through night-scout. I will get by until my replaced receiver arrives. thanks
Scott Leibrand
@scottleibrand
Sep 19 2016 03:25
@iainct ah, gotcha. I know there are some people working on rigs that will integrate a wixel, but I'm not aware of any methods that have been created yet to get G5 data to an openaps rig without Internet. You could of course try Loop for offline use, though...
Hannah Crabtree
@lollydaggle
Sep 19 2016 03:37
For anyone who bought their pump from someone online (i.e. CL, Medwow, etc), how'd you go about paying for the pump and ensuring you'd actually receive it? Leap of faith? Paypal would be my normal route but there's no buyer protection for stuff you can't legally buy/sell.
I may have a lead for a backup pump or it may not even be old enough but this was a roadblock for me when acquiring my first
Scott Leibrand
@scottleibrand
Sep 19 2016 03:43
@lollydaggle do you have a prescription for the pump you're buying? If so, I think it is legal to buy it. And I suspect Paypal would still be willing to reverse a transaction if the seller simply refused to ship the item entirely.
as far as ensuring that the pump is in working condition, you might have the seller send you a video of turning it on and hitting the Esc button to show the firmware version.
Hannah Crabtree
@lollydaggle
Sep 19 2016 03:46
@scottleibrand I do have a prescription -- didn't realize that could matter from a legal standpoint but that makes sense. Video is definitely a good idea.
Russell Gardner
@RussGG
Sep 19 2016 03:46
@lollydaggle I bought mine through Medwow. I used Paypal, not even thinking about the issue of the prescription issue. The guy I bought it from appears to have quite a few and is an openAPS user. I had no issues with getting exactly what he said I was getting.
Tim Street
@tim2000s
Sep 19 2016 05:40
@lollydaggle Mine was Medwow sourced too. But bought from a different EU country so I'm not too sure how that would work legally.
Tim Street
@tim2000s
Sep 19 2016 05:46
@iainct There seems to be an experimental Mo go db on Pi that you could point xdrip at via mobile hotspot that would provide an offline store. Seems very much alpha at the moment.
jmcrawford
@jmcrawford
Sep 19 2016 05:47
@lollydaggle - we used both Medwow w/paypal and kijiji - in the case of Kijiji, my aunt lived close by and could go and meet with the seller at a coffee shop - so we had eyes on both of the pump and the seller. I then sent a payment as I didn't want my aunt to be at any risk of buying the pump 'illegally'. Like others - Medwow seller seemed to be selling multiple pumps and was well aware of the openAPS needs. The only 'blip' if you will with medwow is the pump came from the US and we are in Canada - we started to get concerned when it took about 4 weeks to finally come - it had been stuck in customs, but eventually cleared without any difficulty. So importing might take longer and I suppose could be trickier. In both cases, however, we got never used pumps, as described. I think a leap of faith is at least a small part of the equation!
Colin Lennon
@colinlennon
Sep 19 2016 10:09
@iainct My offline solution is - I run mongodb on my Edison rig, and I have the xdrip app on my Android phone upload glucose readings into mongodb. The phone and the Edison are connected via bluetooth tethering (but running a WiFi hotspot on the phone would also work). I've written a python script to export the values from mongodb and I use that script in my loop. Are you using an Edison or a Pi?
Tim Street
@tim2000s
Sep 19 2016 10:25
@colinlennon - I'm doing something not a thousand miles removed on my Pi rig with the 640 uploader.
The python and bash scripts generate a formatted file that openAPS uses instead of running monitor/glucose.
Colin Lennon
@colinlennon
Sep 19 2016 10:30
@tim2000s sounds good, looking forward to hearing your thoughts on how the algos compare too.
You didn't need to install mongodb on your Pi, did you? When I get time to document my setup, I'll test out installing mongodb on a Pi. Hopefully it is as easy as it was on the Edison.
Tim Street
@tim2000s
Sep 19 2016 10:53
No @colinlennon , I didn't due to my collection mechanism. If/when I flip to Dexcom (go full-time on OpenAPS) I'll need it. Had a few issues with getting Docker to run, but as it was an additional project I left it.
Colin Lennon
@colinlennon
Sep 19 2016 11:02
Thanks @tim2000s , I'll post a link on here when I've documented how my setup works. Should be possible to install mongodb directly rather than in a Docker container, I'll test that out as soon as time permits.
John Males
@johnmales
Sep 19 2016 12:11

Hi @scottleibrand I’m having an issue with the oref0-setup.sh script. While running openaps ns-loop I get this output:

openaps ns-loop
Starting ns-loop at Mon 19 Sep 21:25:11 AEST 2016:
ns://JSON/shell/cgm/ns-glucose.json
reporting cgm/ns-glucose.json
detect-sensitivity://text/shell/settings/autosens.json

/usr/local/lib/node_modules/oref0/lib/profile/isf.js:20
    isf_data = _.sortBy(isf_data.sensitivities, function(o) { return o.offset;
                                ^
TypeError: Cannot read property 'sensitivities' of undefined
    at Function.isfLookup (/usr/local/lib/node_modules/oref0/lib/profile/isf.js:20:33)
    at detectSensitivityandCarbAbsorption (/usr/local/lib/node_modules/oref0/lib/determine-basal/cob-autosens.js:54:24)
    at Object.<anonymous> (/usr/local/lib/node_modules/oref0/bin/oref0-detect-sensitivity.js:73:5)
    at Module._compile (module.js:456:26)
    at Object.Module._extensions..js (module.js:474:10)
    at Module.load (module.js:356:32)
    at Function.Module._load (module.js:312:12)
    at Function.Module.runMain (module.js:497:10)
    at startup (node.js:119:16)
    at node.js:906:3
reporting settings/autosens.json
morRefreshed temptargError: carb_ratio undefined out of bounds
 and meal-carbs
Uploadgrep ns-loop openaps.ini
ed

Drilling down through the reports in this alias

openaps invoke settings/autosens.json
detect-sensitivity://text/shell/settings/autosens.json

/usr/local/lib/node_modules/oref0/lib/profile/isf.js:20
    isf_data = _.sortBy(isf_data.sensitivities, function(o) { return o.offset;
                                ^
TypeError: Cannot read property 'sensitivities' of undefined
    at Function.isfLookup (/usr/local/lib/node_modules/oref0/lib/profile/isf.js:20:33)
    at detectSensitivityandCarbAbsorption (/usr/local/lib/node_modules/oref0/lib/determine-basal/cob-autosens.js:54:24)
    at Object.<anonymous> (/usr/local/lib/node_modules/oref0/bin/oref0-detect-sensitivity.js:73:5)
    at Module._compile (module.js:456:26)
    at Object.Module._extensions..js (module.js:474:10)
    at Module.load (module.js:356:32)
    at Function.Module._load (module.js:312:12)
    at Function.Module.runMain (module.js:497:10)
    at startup (node.js:119:16)
    at node.js:906:3
reporting settings/autosens.json
C-Ville
@C-Ville
Sep 19 2016 12:14
roaming is slow but generally works on my edison rig, however, if I lose connection to my cell phone hotspot, it seem like the edison WLan goes into energy save mode and never wakes up again. Does anyone know of a way to disable energy saving for WiFi on the Edison
Tim Street
@tim2000s
Sep 19 2016 12:16
@colinlennon ,👍
Colin Lennon
@colinlennon
Sep 19 2016 12:28
@C-Ville are you on Yocto or Ubilinux now? I have a cron job that runs every minute and executes a connmanctl connectso that a dropped connection is re-established quickly. That's for BT tethering but should work for WiFi too
C-Ville
@C-Ville
Sep 19 2016 12:30
@colinlennon sorry forgot to mention I am on ubilinux. I believe conmanctl is not enabled/installed on ubilinux
Colin Lennon
@colinlennon
Sep 19 2016 12:33
@C-Ville Oh okay, I'm not familiar with Ubilinux, but hopefully someone else can give you some advice
Tim Street
@tim2000s
Sep 19 2016 13:23
@colinlennon I've just found this for natively installing mongo on Pi. Looks dead easy. Should be fine with xDrip: www.yannickloriot.com/2016/04/install-mongodb-and-node-js-on-a-raspberry-pi/
Colin Lennon
@colinlennon
Sep 19 2016 13:29
@tim2000s Excellent, thanks! Sounds like it's even easier than it was on the Edison (and it was pretty easy on there). Good news :smile:
Colin Lennon
@colinlennon
Sep 19 2016 13:39
I think @oschumac 's solution is the best one for someone starting from scratch. My approach suited me because I had already built an xdrip wireless bridge and it's working perfectly. Waiting for some more info on GPIO access on the new 900MHz Explorer and RiledUp boards - an ideal solution would maybe be one of those boards plus a Wixel, all in a single case.
Tim Street
@tim2000s
Sep 19 2016 13:58
I think, if you're buying. G5 and not going for Loop, then the xdrip beta does this quite well. I'm fairly sure that porting it to linux from android can't be that hard.
C-Ville
@C-Ville
Sep 19 2016 16:01
@colinlennon I hope that the Explorer board will have UART connections, that would then easily allow for a Wixel.
Chris Oattes
@cjo20
Sep 19 2016 16:03
There are two USB connectors on an explorer board, a console and an OTG
I think the only other connectors available are battery and the CC1110 header
Colin Lennon
@colinlennon
Sep 19 2016 16:28
USB OTG should work with a Wixel, but it might mean more power consumption (5v output). It might even be possible to provide power to the Explorer board and to an xdrip wireless bridge from a single battery, put it all in one box. Would still need mongodb on the Edison for offline looping.
Looking at the images @danamlewis posted of the Explorer board setup, I'm sure I could cope with carrying that and an xdrip. It looks tiny! :smile:
Chris Oattes
@cjo20
Sep 19 2016 16:29
Explorer can't currently provide 5V. There is a request to provide 5V if one of the two inputs has a 5V connection, I can't remember which way round it was
diabeticgonewild
@diabeticgonewild
Sep 19 2016 16:29

I know this is crazy, but can somebody give me access to a "Nightscout site"?

I need a "stream of numbers" in order to BS blood sugar logs for Medicare, or I won't be able to get test strips.

Yes, I can "make up my own numbers" but having a "stream of numbers" will be so much easier....

PM me privately, please. I really appreciate it!
Chris Oattes
@cjo20
Sep 19 2016 16:30
what? Why do you need a 'stream of numbers' to get test strips?
diabeticgonewild
@diabeticgonewild
Sep 19 2016 16:31
In order to get > 3 test strips/day on insulin, Medicare requires BG logs. It's such bull****. My endo told me to NEVER have < 6 tests/day, written on my logs...
I have autonomic and peripheral neuropathy, NOT due to my diabetes, and that's why I get 6 test strips per day.
My neuropathy is due to rare autoimmune diseases...
Chris Oattes
@cjo20
Sep 19 2016 16:32
@colinlennon: And yes, it is tiny. It's too big for my small case from Maplin
diabeticgonewild: Can't you just show them the logs of the >= 6 tests a day you absolutely have to do?
diabeticgonewild
@diabeticgonewild
Sep 19 2016 16:32
Like Medicare audits people's records, multiple times, if you get > 3 tests/day on MEdicare
Cause, sometimes I just "check out" for the day, with my medical problems, and I DON'T always test >= 6x/day
So yes, I would appreciate a stream of numbers...It will greatly save me time...
I also BS my numbers, cause it's quicker to make up numbers than to retrieve them from the meter...
BS.jpg
That's what my records will look like...typed in
I have a year to do...so it will take me awhile...
diabeticgonewild
@diabeticgonewild
Sep 19 2016 16:38
Like, Medicare audits my endo medical records, WITHOUT my permission. The US government can pretty much do whatever they want.
And, if I don't turn over my "numbers" to my endo. The US government goes after my endo, NOT me...
And I don't want to screw over my endo...
Hannah Crabtree
@lollydaggle
Sep 19 2016 16:42
@RussGG @tim2000s @jmcrawford Thanks for the input! I feel a lot better about buying a pump over the internet now.
722/522s shouldn't have the security/firmware issue? That was my understanding. Will ask for a picture of the pump and a video though if the price is right...
Eric
@ecc1
Sep 19 2016 16:47
Why do you conclude that you need to run the mongo server on the Edison for offline looping? There's enough history on the pump and CGM to last until you regain connectivitiy; why not just backfill then? That approach seems more robust to me (and useful, because cloud)
diabeticgonewild
@diabeticgonewild
Sep 19 2016 16:50
halfway.jpg
Halfway there on my first monthly record....so 1/24 of the way there...in order to get test strips...
Eric
@ecc1
Sep 19 2016 16:51
I'm debugging a classic PID controller for my APS. If it works, it will be ignoring the IOB (other than reporting it to NS) since the effect of the IOB should all be included in the integral term (cumulative error). LIving dangerously ...
Actually after last night, I want to print up a T shirt that says "My robot pancreas is trying to kill me"
Chris Oattes
@cjo20
Sep 19 2016 16:52
ecc1: how's the tuning of that going? I had a go and getting a PID setup that was anywhere near sensible seemed very hard
Either a lot of overshoot or ringing
or taking about 12 hours to correct a 180
diabeticgonewild
@diabeticgonewild
Sep 19 2016 16:53
@cjo20 you can do a PID pretty easily via MATLAB...
Chris Oattes
@cjo20
Sep 19 2016 16:53
I know how to create a PID, it's the tuning that's difficult to get right
C-Ville
@C-Ville
Sep 19 2016 16:53
if sparkfun is going to build the explorer board I hope that it will fit with the other block boards, that way we could have UART for Wixel although at the cost of a slightly thicker device
Eric
@ecc1
Sep 19 2016 16:54
Too early to say yet. I wish I could speed everything up x10 or more to run tests :-)
Chris Oattes
@cjo20
Sep 19 2016 16:54
C-Ville: I believe it is sparkfun-block compatiable. It's only got one connector, like the battery board
Eric
@ecc1
Sep 19 2016 16:54
And you're right, the "debugging" is actually all tuning.
Chris Oattes
@cjo20
Sep 19 2016 16:55
You don't even need matlab to do a PID, it's about 10 lines of C code
C-Ville
@C-Ville
Sep 19 2016 16:55
@cjo20 that would not matter to much, but OTG for Wixel would in my view defeat the purpose of a small device with bulky USB connectors
Scott Leibrand
@scottleibrand
Sep 19 2016 16:57
@C-Ville yes @cjo20 is correct: you can stack a UART board in between the Edison and the Explorer board if you need that. Also, @fjungman's new RiledUp board is bigger but should have built in wixel connect points.
@johnmales what does the sensitivity data in the input file look like?
Tim Street
@tim2000s
Sep 19 2016 16:59
@lollydaggle mine's a 522 running the correct firmware.
Colin Lennon
@colinlennon
Sep 19 2016 17:00
@ecc1 re. offline looping - I need a way of getting Dexcom readings into my loop. On what you said about pulling data from the pump, you're maybe thinking of a situation where a Medtronic CGM is being used?
Eric
@ecc1
Sep 19 2016 17:01
I use Dexcom too. I fetch its history over BLE.
C-Ville
@C-Ville
Sep 19 2016 17:02
@ecc1 unfortunately my dex does not provide BLE, old G4 :-(
Eric
@ecc1
Sep 19 2016 17:03
Maybe we have our wires crossed (no pun intended). Are you connecting a G4 receiver via USB to something, or are you receiving the G4 sensor transmissions using a Wixel, or ...?
C-Ville
@C-Ville
Sep 19 2016 17:05
@ecc1 I have my G4 data send to my xdrip (wixel2BLE)
meaning that my loop goes dexcom ->xdrip -> cell phone ->NS cloud ->cell phone (hotspot) ->edison-> pump and reverse
which left me without my loop while on vacation (no cellphone reception), this is why I am fully with @colinlennon that a nice offline solution is so useful
Chris Oattes
@cjo20
Sep 19 2016 17:08
@ecc1 I'd be interested to see your PID results when you have some available to share
Kevin Wells
@wkcwells
Sep 19 2016 17:08
This message was deleted
Eric
@ecc1
Sep 19 2016 17:08
Ah, OK. I'm using the G4 receiver; not optimal for size and weight, but it's been a lifesaver in the form of redundancy
diabeticgonewild
@diabeticgonewild
Sep 19 2016 17:09
Yeah, I just finished my first record out of 12.
Chris Oattes
@cjo20
Sep 19 2016 17:09
diabeticgonewild: You could probably write something in matlab to generate it for you
diabeticgonewild
@diabeticgonewild
Sep 19 2016 17:09
I decided to use RANDOM.org and generate random numbers between a certain range, with 7 columns, with 30 records in each column
I was think MATLAB, but it's not as simple as it looks, and I don't want to debug code...
C-Ville
@C-Ville
Sep 19 2016 17:10
@ecc1 my G4 receiver can only do OTG, and OTG does not seem to work for me when powering the Edison via battery
diabeticgonewild
@diabeticgonewild
Sep 19 2016 17:10
RANDOM.org is PERFECT! Hahahaha!
@cjo20 , thanks!!!!
Chris Oattes
@cjo20
Sep 19 2016 17:11
it'll work until they use the numbers to work out your average glucose and work out you couldn't possibly have the a1c your bloodwork suggests :P
diabeticgonewild
@diabeticgonewild
Sep 19 2016 17:15
NAH, my endo said that they are a bunch of idiots anyway! :D
Eric
@ecc1
Sep 19 2016 17:16
@C-Ville I think I understand your design constraints now, so I agree that local storage of the xdrip data seems reasonable. Thanks for walking me through it!
C-Ville
@C-Ville
Sep 19 2016 17:18
20160830_225208.jpg
I got it all in this box, and also mongo and NS running without need of a SD card, and then could not use it on vacation as I was to stupid to get autostart to work with mongo and NS
diabeticgonewild
@diabeticgonewild
Sep 19 2016 17:23
Actually, I could just do a normal distribution...good enough...with a decent SD/Variance....
rocket science, apparently...
Then I can sort it into a vector in MATLAB
Then it will be more "legit"...
Chris Oattes
@cjo20
Sep 19 2016 17:24
It might be easier in future to just get logs off your meter every month or two
diabeticgonewild
@diabeticgonewild
Sep 19 2016 17:25
Nah, it takes too long.
Thanks for the suggestion, though!
Chris Oattes
@cjo20
Sep 19 2016 17:28
Then your endo can actually see your BG numbers too
diabeticgonewild
@diabeticgonewild
Sep 19 2016 17:38

@cjo20 here is my MATLAB code for my "fake numbers":

R = lognrnd(log((142^2)/sqrt(10000+142^2)),sqrt(log(10000/(142^2)+1)),[31,7])

Average BG: 142
Variance: 10,000

31 rows (days)
7 columns (tests)

Remember that this is a random number generator for a lognormal distribution...

MATLAB is such a beautiful thing....
diabeticgonewild
@diabeticgonewild
Sep 19 2016 17:46
R = int64(lognrnd(log((142^2)/sqrt(12000+142^2)),sqrt(log(12000/(142^2)+1)),[31,1]))
R =

                   37
                  331
                  137
                   50
                  235
                   46
                  103
                  187
                   63
                   89
                  115
                  427
                  154
                   86
                   96
                   93
                  275
                   93
                   65
                  133
                  105
                  193
                  217
                  236
                  175
                  173
                  413
                   51
                   89
                   35
                  211
Just copying bull****ed numbers, column by column now...
C-Ville
@C-Ville
Sep 19 2016 17:56
I recently had this strange key error on my 715. After I did put it into a rice bag and dried out it started working again, however the ESC button is now flaky, meaning I have to press it really hard. I saw a youtube video that you can disassemble the key part and clean it. Has anyone tried that, I am concerned how I will glue the key cover back on
Scott Leibrand
@scottleibrand
Sep 19 2016 18:08
@C-Ville we currently use our bedside Pi rig for offline use with the G4 receiver plugged in via USB. It's a bit bulky, but extremely reliable. I think you can do the same with an Edison and Sparkfun base block if it has a 5V USB battery attached, but we haven't really tried that yet. Also, if you get it working, there's an openaps use cgm command to tell the Dexcom to stop drawing power over USB.
C-Ville
@C-Ville
Sep 19 2016 18:19
@scott I just took 2 weeks without looping and I hope that I will have something smaller lined up next time I leave country. These where my first 2 weeks without a loop after I started pump therapy and I can tell it is annoying, so much more to look @ .... Now I truly know what I am gaining with OpenAPS
to be honest I even had to switch back to pen and pencil, because my medtronic did not like to go hiking with me, looks like the pump is really sensitive to humidity.
Eric
@ecc1
Sep 19 2016 18:26
@C-Ville Did you get your autostart sorted out? For programs running as root, I usually just put it in /etc/rc.local. For user programs (also works for root), you can use the @reboot entry in a crontab.
C-Ville
@C-Ville
Sep 19 2016 18:34
I tried some sort of skript, /etc/init.d there for MongoDB which did not work at all
so I gave up 2h before we headed of to our vacation in France
looks like I took the wrong path. It is tempting, but I will try to not make another attempt tonight
PieterGit
@PieterGit
Sep 19 2016 20:03
@scottleibrand : small error in oref0-setup.sh: API_SECRET is added to crontab even it's there already
Scott Leibrand
@scottleibrand
Sep 19 2016 20:04
ah, thx. just need to add a grep first
PieterGit
@PieterGit
Sep 19 2016 20:29
new setup seems to work fine, but i needed some bootstrap stuff (see #218)
garykidd
@garykidd
Sep 19 2016 21:27
SO took my rig with me. Never got internet access while away..couldn't connect to my wifi or BT hotspot on the go. Intereting that when I rolled into driveeway --it connected to my home network straight away. Anyone have any idea on how I can try to troubleshoot?
C-Ville
@C-Ville
Sep 19 2016 21:36
@garykidd what setup do you have?
ParkerBenda
@ParkerBenda
Sep 19 2016 21:37
From Dallas jdrf:
When I originally asked about you sharing this with our support group, the response I received from the International Board included this: We cannot promote OpenAPS. FDA is very unhappy about this and there is a good chance that they will shut them down. It's been a fine line for me. I love the innovation that they are driving, but it is a risk for non-experts. I've showed slides of OpenAPS to show the potential of innovation and then say that we want a thriving ecosystem of FDA approved and reimbursed systems and that is the focus at JDRF.
PieterGit
@PieterGit
Sep 19 2016 21:39
@scottleibrand (or others) : help with openaps/oref0#218 would be very welcome. i'm officialy stuck on what's going wrong
garykidd
@garykidd
Sep 19 2016 21:47
@C-Ville I have an Rpi3 trying to connect thru mobile hotspot and failed both thru BT and wifi. I can, however, automatically connect to home network. WHen I try to manually connect thru BT I get a 'succesfully connected but the connected device has no services' or something like to that affect. and when I attempt to connect to manually to mobile wifi hotspot it doesnt like the PSK I enter--though I know its right...it reports back something to the effecct of 'failed to reconfigure'....but I am positive I have the correct PSK in there....and I can connect from other device from mobile hotspot---very confusing.
Scott Leibrand
@scottleibrand
Sep 19 2016 21:49
@PieterGit commented on the issue. not entirely clear what you're trying to do
Dana Lewis
@danamlewis
Sep 19 2016 21:53
@ParkerBenda JDRF has also historically not allowed NS to be a part of local events, etc. for similar reasons (So not surprised to see this). But worth noting we do have a good relationship with them, and I am actually talking with Aaron Kowalski about he & I co-authoring a piece about setting expectations for closed looping, because we both strongly feel the need to address some misconceptions in the broader community that may hurt adoption when there’s a commercial option available. And all the while, JDRF and we as an open source community each continue to advocate however we can to achieve the same end goals: better technology in the hands of patients as soon as possible.
ParkerBenda
@ParkerBenda
Sep 19 2016 21:54
I personally couldn't care less about the jdrf. Just curious under what pretense the FDA would say openaps was violating laws
Dana Lewis
@danamlewis
Sep 19 2016 21:54
they’re not saying that
ParkerBenda
@ParkerBenda
Sep 19 2016 21:55
Oh
Dana Lewis
@danamlewis
Sep 19 2016 21:55
they’re publicly saying they’re “monitoring” the open source community, which is old news, and their take is that they have ‘enforcement discretion'
ParkerBenda
@ParkerBenda
Sep 19 2016 21:55
Oh
Dana Lewis
@danamlewis
Sep 19 2016 21:55
however legally, their job is to regulate products that are commerically distributed
there’s no commercial distribution happening, so nothing to be done right now. (unless something changes, or someone tries to do something commercial without going through regulatory processes)
(and then other people/orgs take that general gist and paraphrase it and it comes across as it did to you.)
garykidd
@garykidd
Sep 19 2016 21:56
sonia sotomayor is a T1D...maybe if there are issues that ever get that far, she'd vote to hear them...kind of kidding
diabeticgonewild
@diabeticgonewild
Sep 19 2016 21:59
The FDA can't take down OpenAPS. It is published openly on the Internet.
Jason Calabrese
@jasoncalabrese
Sep 19 2016 22:04
@danamlewis when Heidi talked to Arron at AADE he said the same "FDA is very unhappy about this" I think this is JDRF FUD, and its not what I've ever heard from you, Ben, or anyone else thats talked to the FDA. I don't know why they are intentionally spreading that rumor,
but we're going to be at the JDRF ride in Tucson again this year, and the walk, etc, so maybe it would be fun place for an OpenAPS meetup
Dana Lewis
@danamlewis
Sep 19 2016 22:06
I generally suggest don’t believe anything you don’t hear firsthand from the mouth of someone employed by the FDA or in writing in an official FDA email/document/etc…otherwise “telephone” happens and things get taken out of context as they get discussed down the line. So maybe you shouldn’t believe me, either ;) but definitely please don’t feel the FUD (fear, uncertainty, doubt) from anyone else, either!
ParkerBenda
@ParkerBenda
Sep 19 2016 22:10
Thanks @danamlewis
Dana Lewis
@danamlewis
Sep 19 2016 22:10
:+1: thx for opening the discussion!
Eric
@ecc1
Sep 19 2016 22:20
Worst case, we can start the Pirate Pancreas Party.
diabeticgonewild
@diabeticgonewild
Sep 19 2016 22:21
My cousin who works for a medical device company said that the FDA was "unhappy" about it too. He had more to say about it too, but overall he thinks it is still neat...
Jason Calabrese
@jasoncalabrese
Sep 19 2016 22:23
I'd just like to track down where that "unhappy" is coming from, it's being repeated, but who's the source
diabeticgonewild
@diabeticgonewild
Sep 19 2016 22:23
Probably multiple...
C-Ville
@C-Ville
Sep 19 2016 22:23
@ecc1 as long as nobody seizes my AP, I would not mind pirate pancreas parties
David Young
@wgs4
Sep 19 2016 22:24
I haven’t poked around in a little over 3 months.. I had a pretty big deal surgery and work stuff happen that derailed my efforts to get my unit up and going for my sister. But she is on my case big-time now so I’m fully recovered and back in the game.
diabeticgonewild
@diabeticgonewild
Sep 19 2016 22:24
He said that if it happened [to his small company that he works for as an electrical engineer], that the FDA "would shut them down"
David Young
@wgs4
Sep 19 2016 22:24
I assume I’ve got a ton of reading to do to catch up
Jason Calabrese
@jasoncalabrese
Sep 19 2016 22:24
device companies might be "unhappy", but doesn't make sense for the FDA to say that, so I assume it's coming from somewhere else
C-Ville
@C-Ville
Sep 19 2016 22:24
@garykidd by mobile hotspot, do you mean your cell phone used as a hotspot
Oskar Pearson
@oskarpearson
Sep 19 2016 22:24
Great, David. Yep, you have, I'm sure! :)
David Young
@wgs4
Sep 19 2016 22:25
Has the documentation had any signifcant updates?
diabeticgonewild
@diabeticgonewild
Sep 19 2016 22:25
@jasoncalabrese I wouldn't make any assumptions.
Dana Lewis
@danamlewis
Sep 19 2016 22:26
@wgs4 yes!
stay tuned for an email about a new setup script you’ll want to try :)
(and welcome back)
David Young
@wgs4
Sep 19 2016 22:26
you have my email? a setup script would be great.
Oskar Pearson
@oskarpearson
Sep 19 2016 22:26
I'm not sure if the new setup scripts have made it into the docs yet - but soon no doubt
diabeticgonewild
@diabeticgonewild
Sep 19 2016 22:27
If you use the templates posted in Github, you can set it up pretty quickly without setup scripts.
And from that you could make an install script pretty quickly, probably within a day or 2
And do a very nice job...
PieterGit
@PieterGit
Sep 19 2016 22:28
I can recommend the setup scripts. they really make life easier
David Young
@wgs4
Sep 19 2016 22:29
awesome
Dana Lewis
@danamlewis
Sep 19 2016 22:29
it’ll go out to openaps-dev in a bit, @wgs4
diabeticgonewild
@diabeticgonewild
Sep 19 2016 22:29
I'm just saying it's not hard to figure out if you look in certain places on OpenAPS Github outside the docs
Oskar Pearson
@oskarpearson
Sep 19 2016 22:35
@danamlewis We had the talk + Q&A with the NHS London doctors today.. there are really some super interested doctors here, and a lot of people eager to learn more :)
We think it went really well
Dana Lewis
@danamlewis
Sep 19 2016 22:36
awesome, great job @oskarpearson
Oskar Pearson
@oskarpearson
Sep 19 2016 22:36
Mostly Siobhan! :) But thanks
Thanks for the slides - I really like them btw. We did stuff slightly differently because of what the primary contact wanted to cover, but it was great to formalise thinking. Great job too ;)
Dana Lewis
@danamlewis
Sep 19 2016 22:38
happy to share anytime! :) :+1:
eyim
@eyim
Sep 19 2016 22:50
@danamlewis The commercial vs non-commercial distinction isn't apparent from the Mark Wilson presentation (wonderful presentation by the way). https://www.youtube.com/watch?v=L52OP7d4xpY I don't know who the FDA person speaking at 26:45 min is but she states "100% of these devices are under FDA jurisdiction. So technically it is not legal to build an artifical pancreas device and use it without FDA approval. However, we have something called enforcement discretion and we regulate by risk . . ." Not trying to say whether the FDA is right or not. Just trying to point times they have stated their stance on openaps.
My personal opinion is that regardless of their stance, I will continue to use and support the effort as I know it is life changing and can help the lives of many diabetics.
garykidd
@garykidd
Sep 19 2016 22:57
@C-Ville I do yes
Dana Lewis
@danamlewis
Sep 19 2016 23:04
@eyim yea. she’s wrong, unfortunately..an individual can do whatever they want to/for themselves..the distinction is distributing a device, which is what they regulate.
Lynne Butts
@flutelynne_twitter
Sep 19 2016 23:04
@danamlewis I can see the motto on a t-shirt..."don't fear the FUD"
Dana Lewis
@danamlewis
Sep 19 2016 23:04
but :+1: for supporting and helping people in the ways that we can, even though we can't and don’t distribute devices :)
@flutelynne_twitter :)
Ryan Dalisky
@rdalisky
Sep 19 2016 23:05
I'm a doctor, have been looping for 3 days. I plan on writing a paper about my experience with openaps. And trust me it is already night and say.
*day
Dana Lewis
@danamlewis
Sep 19 2016 23:06
:+1: cool @rdalisky, would love to see it when it’s ready!
ps :confetti_ball: for you :)
Ryan Dalisky
@rdalisky
Sep 19 2016 23:08
I'll present it here in Michigan, and hopefully I grab some attention. This is seriously years ahead of anything commercial, and the built-in safeguards are wonderful. I'll make sure to link everyone here when the paper is done.
Dana Lewis
@danamlewis
Sep 19 2016 23:08
:+1:
bonus points if you reference the study!
Ryan Dalisky
@rdalisky
Sep 19 2016 23:10
Reference the study with openaps.org? You're darn right I am going to!
woot! :) Awesome
Ryan Dalisky
@rdalisky
Sep 19 2016 23:15
Thank you for that, it will definitely help. My poster is going to show the before and after effects of using openaps as a Cheif resident in General Surgery. That's going to turn some heads 😎
Dana Lewis
@danamlewis
Sep 19 2016 23:16
:sparkles: :clap:
garykidd
@garykidd
Sep 19 2016 23:43
great presentation by mark wilson above. I love the car/drive analogy. It occurred to me that the car before hot wiring is not only lacking instrumentation let alone any intelligent controls...it goes too fast...steers to slow, its brakes don't slow you nearly quickly enough...there's often a dirty windshield which you can't see out of...and you're driving at night on a twisty roads with endless cliffs at roads edge...and then there's the police that tell you you can't fix your car --that's against the law---just be careful and wear your seatbelt (as in just take your insulin) and you'll be fine.
@danamlewis I'm sure you probably already have this...but how bout a form to fill out when one initiates loop...not just your name..but a decent level of information on every participant....I think it should be coded too. The coded data could/should be aggregated and shared broadly....to anyone that wants access.....I like the idea of groups of people aggregated together and the data (real time) being aggregated....the data could be really useful in helping to generate better algorithms.
garykidd
@garykidd
Sep 19 2016 23:49
maybe everyone already knows that's where this is headed. I'd share my data.
that’s more so we can add people to email lists to notify in case of bugs, but that’s also how we roll the community “count” of DIY loopers
and I recently added a Q about interest in sharing data for future research projects
Scott Leibrand
@scottleibrand
Sep 19 2016 23:57
@PieterGit fixed the API_SECRET duplication thing
garykidd
@garykidd
Sep 19 2016 23:59
Yup...I saw this form...but thinking of significantly deeper level of info (btw I filled out the form)...like all my NS data for a year or something. Even if we could do it for a week initially for lats say 30 users all running the same openaps setup...could be pretty powerful even with that limited information.....I could easily envision financial support to get someone to pay for the costs of obtaining the information and hosting db, sites, etc.