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

1st
Jun 2017
rsilvers129
@rsilvers129
Jun 01 2017 00:00
Is there something I have to do to get a Pushover alert when the rig thinks carbs are required?
Dana Lewis
@danamlewis
Jun 01 2017 00:00
yep. set up pushover and have SMB/oref1 enabled
right now in dev, pushover is a manual setup, but the PR i’m having folks test would add Pushover prompts to the advanced features menu to make setting it up easier.
garykidd
@garykidd
Jun 01 2017 00:03
and was just curious---the 1/3 a user definable variable or static in the oref1 code?
Dana Lewis
@danamlewis
Jun 01 2017 00:03
coded
garykidd
@garykidd
Jun 01 2017 00:03
thx
rsilvers129
@rsilvers129
Jun 01 2017 00:04
@danamlewis I have pushover working in general. I am using dev with SMB. But I have not got that message. What else do I need to enable?
garykidd
@garykidd
Jun 01 2017 00:04
intuitively was thinking that user definable setting might be beneficial....particularly if you may want SMBs more limited initially say for instance 1/4/ or 1/5 estimated dose.
Scott Leibrand
@scottleibrand
Jun 01 2017 00:05
@rsilvers129 have you seen anything in your pump-loop.log with carbsReq?
Dana Lewis
@danamlewis
Jun 01 2017 00:05
you can dig into the code and figure it out manually, or test the code from this PR: openaps/oref0#486
Scott Leibrand
@scottleibrand
Jun 01 2017 00:05
if not, you can set a really high temp target to trigger that
Scott Leibrand
@scottleibrand
Jun 01 2017 00:05
@garykidd there is a very narrow window there of sensible settings. if you get down to 1/4 or 1/5, you're probably better off without SMB and just using AMA
and anything higher than 1/3 is probably too aggressive for normal fast-acting insulins like humalog/novolog
rsilvers129
@rsilvers129
Jun 01 2017 00:08
@scottleibrand p; Eventual BG 51 < 90, setting 90m zero temp. 96m left and 0 ~ req 0U/hr: no temp required","COB":1,"eventualBG":51,"tick":-2,"deliverAt":"2017-05-31T23:09:18.210Z"}
bgUndershoot: 34 zeroTempEffect: 14 carbsReq: 4
Scott Leibrand
@scottleibrand
Jun 01 2017 00:08
and did carbsReq also show up in the reason field that time?
garykidd
@garykidd
Jun 01 2017 00:08
that makes sense to me @scottleibrand. I could also envision a construct where SMBs might only kick in if your off-target more than x%...otherwise it would stay with temp basals (oref0) methodology...what do you think?
Scott Leibrand
@scottleibrand
Jun 01 2017 00:08
if not, there may be an un-done TODO there
@garykidd I've explored a lot of those kinds of approaches, and settled on what we have now. But there is still room for improvement, particularly around when to zero temp vs. bolus snooze after a meal bolus...
rsilvers129
@rsilvers129
Jun 01 2017 00:10
@scottleibrand I don't think I ever got a pushover for CarsReq
Dana Lewis
@danamlewis
Jun 01 2017 00:10
@rsilvers129 you haven’t set up pushover
setting up pushover for NS is different
look at the code for the PR I posted if you want to do it manually, or try that PR as the easiest way of setting it up
rsilvers129
@rsilvers129
Jun 01 2017 00:13
Ok. Thanks.
RabidMoleMan
@RabidMoleMan
Jun 01 2017 00:25

@danamlewis my apologies for not having gotten back to you on this earlier. There was an untimely power outage.

In regards to the files raw-pump/settings.json, raw-pump/bg-targets-raw.json, and raw-pump/selected-basal-profile.json, I created a directory called "raw-pump" within the "myopenaps" directory and inside the "raw-pump" directory, added the files required as per the guide. (I am hoping I have interpreted this much correctly) The issue I now have is that there is an example for the first 2 files but not for selected-basal-profile.json one. I'm unsure as to what I need to put in there.

Cameron Chunn
@applehat
Jun 01 2017 01:00
This message was deleted
jtsackdex
@jtsackdex
Jun 01 2017 01:11
So after a couple of hickups of my own doing all seems to be running well. I don't think I'll have time to mess with offline stuff right now. My question is would it be better to carry my rig with me even tho I will not be anywhere near wifi tomorrow or should I just leave it at home or at the office until I get back? Which will be the easier reconnect?
rsilvers129
@rsilvers129
Jun 01 2017 01:26
Did you set up bluetooth tethering?
garykidd
@garykidd
Jun 01 2017 01:28
@scottleibrand does the bolus snooze/dia still drive bolus snooze control of temporary basal increase or does oref1 possible call to zero temp simply trump the oref0 temp basal (under a bolus snooze situation). thx-
jtsackdex
@jtsackdex
Jun 01 2017 01:33
@rsilvers129 I didn't have time nor will I for a couple of days. I will be out onsite so not really sure when I can get back to it. I'm do for a site change and was considering shelving the tSlim. At minimum I'll have overnight coverage.
rsilvers129
@rsilvers129
Jun 01 2017 01:33
The iPhone bluetooth setup was pretty easy. Like 10 minutes.
Scott Leibrand
@scottleibrand
Jun 01 2017 01:33
@garykidd Huh what? Bolus snooze only prevents low temping. It is disabled when SMB is active.
rsilvers129
@rsilvers129
Jun 01 2017 01:34
You can't use the rig without it.
When not on Wifi
Scott Leibrand
@scottleibrand
Jun 01 2017 01:34
@rsilvers129 the bluez compile takes about 45m
rsilvers129
@rsilvers129
Jun 01 2017 01:34
And when using G5.
Scott Leibrand
@scottleibrand
Jun 01 2017 01:34
Fastest thing to set up is a second wifi network for your phone's hotspot.
rsilvers129
@rsilvers129
Jun 01 2017 01:34
@scottleibrand Yes, but you don't have to watch it.
Scott Leibrand
@scottleibrand
Jun 01 2017 01:35
Yeah
garykidd
@garykidd
Jun 01 2017 01:42
@scottleibrand \thanks--I reversed the direction that bolus snooze 'pushes' for lack of better descpition...but you answered my question
jtsackdex
@jtsackdex
Jun 01 2017 01:43
Yeah, I don't have that kinda time. So what's the fastest handshake back to looping, keeping rig with me or just coming back to it? I know I'll just be on my basal settings while away and I'm ok with that.
Scott Leibrand
@scottleibrand
Jun 01 2017 01:43
Doesn't really matter if you aren't using a wifi hotspot.
mrshapir
@mrshapir
Jun 01 2017 01:44
Has anyone here noticed that they have to do anything special to their wpa_supplicant file to get connection to a Belkin router?
garykidd
@garykidd
Jun 01 2017 01:45
still...wondering then if some user defined variable could be established to prevent oref1 zero temp for a defined time frame post meal bolus? understanding this may be a less desirable than imbedded decision logic under a more universal algorithm...
jtsackdex
@jtsackdex
Jun 01 2017 01:45
@scottleibrand @rsilvers129 Thanks
rsilvers129
@rsilvers129
Jun 01 2017 01:47
Helping a friend set up his first rig. He has two boards. One of them just gave tons of python radio errors. He moved the Edison to the other board, and that one is not giving errors and the loop is running, but even with the pump on the Edison, it is only giving -99 signal. Nightscout is showing though that it is seeing the pump (says 1, or 2 minutes ago), but not seeing OpenAPS. "916.636", 0, -99 Starting pump-loop at Wed May 31 19:43:46 MDT 2017:
David
@davidkeddydb
Jun 01 2017 01:48
@garykidd @scottleibrand Well, you can go into the code and adjust from the 120mins right?
rsilvers129
@rsilvers129
Jun 01 2017 01:48
@jtsackdex Just leave it at home.
garykidd
@garykidd
Jun 01 2017 01:54
@davidkeddydb change what in the code from 120 mins? --thx
David
@davidkeddydb
Jun 01 2017 02:06
Screenshot 2017-05-31 22.04.39.png
@garykidd I think you can adjust down here but I am still making sure I understand the code in determine-basal.js
garykidd
@garykidd
Jun 01 2017 02:20
hmmm...does .js mean java script....maybe if a take a shot of .jd it'll make more sense.....
Scott Leibrand
@scottleibrand
Jun 01 2017 02:28
@davidkeddydb no, reducing the maximum length of the zero temp wouldn't help at all
@garykidd yes, .js is javascript
if we needed a user configurable preference to disable the zero temping behavior, it could be something like bolusSnoozeWithSMB, which would default to false, but if set to true would bolus snooze zero temps even when SMB is active...
David
@davidkeddydb
Jun 01 2017 02:30
@scottleibrand I won't do that then. I am really still reading determine-basal.js to make sure I understand all of it so I can make some suggestions.
@scottleibrand Thanks for the clues. I will either comment on the open issue or pm you once I get some ideas
mrshapir
@mrshapir
Jun 01 2017 02:41
@scottleibrand I am using some of your old advice about using belkin router that looks like it is connected to the edison when you do ifup wlan0 but wont ping google.com. I'm unable to edit this file though: /etc/dhcp3/dhclient.conf
"/etc/dhcp3/dhclient.conf" E212: Can't open file for writing
Scott Leibrand
@scottleibrand
Jun 01 2017 03:14
I don't remember ever suggesting editing that file...
mrshapir
@mrshapir
Jun 01 2017 03:16
@scottleibrand You sent someone else this link https://developers.google.com/speed/public-dns/docs/using#linux
Aaron Michelson
@Aaybob
Jun 01 2017 03:31
I got ssh on my iphone! Thanks to the documentation effort. My v1.2 explorer rig was buggy but this time it loaded all the software dev and oref0 0.5.0. It's all so wonderful. My last endo took pictures of my explorer rig with autotune printout paper for a future conference. His dietician stayed in fantasy land about what artificial pancreases would do said they were all the talk but no one had seen one.
David
@davidkeddydb
Jun 01 2017 03:31
@scottleibrand is their a maximum that the bolussnoozer_dia_divisor? You can set it to 2 to reduce the zero temping to half. If my DIA is 5 hours, Can I get the temping to last an hour by setting the bolussnooze_dia_divisor to 5
Scott Leibrand
@scottleibrand
Jun 01 2017 03:47
@mrshapir you can just do the resolv.conf stuff without touching dhclient.conf I think
@davidkeddydb you can set it as high as you want. keep in mind that because the snoozeBG is being reduced gradually over whatever timeframe you choose, it will often start low-temping before that time is up. so if you did 5, then it might start low-temping sooner than 1h after the bolus, but bolus snooze would be completely gone by 1h.
Dana Lewis
@danamlewis
Jun 01 2017 03:55
Woot, thanks to @Kdisimone for some epic testing (plus @tomasboudr), got some improvements to oref0-setup.sh merged into dev. Should look a little (a lot!) different the next time anyone pulls and runs dev setup :)
katie disimone
@Kdisimone
Jun 01 2017 04:00
@Aaybob there’s a great app (if you are willing to pay $1.99) that is like terminus but even better. Called SimpleSSH. I think it could be setup like the HotButton app for androids perhaps...
Dana Lewis
@danamlewis
Jun 01 2017 04:16
At this point, given how many people are using oref1 despite it being in dev, I am planning to go ahead and merge the oref1 docs into the main docs so it’ll be more easily accessible. It can be edited after the fact (as can all our docs), but I’m going to let openaps/docs#777 sit until tomorrow in case anyone goes “ahhh! XYZ must be on that page before it goes more live. (And for an easier view of that oref1 page, see https://github.com/openaps/docs/blob/14ee3ea85d1737db1216319e91a2ace62ca0d560/docs/docs/walkthrough/phase-4/oref1-features.md)
rsilvers129
@rsilvers129
Jun 01 2017 04:27
@Kdisimone SSH on iphone is useful. I am using Terminus and it works well.
Scott Leibrand
@scottleibrand
Jun 01 2017 04:31
*termius (no n)
katie disimone
@Kdisimone
Jun 01 2017 04:41
There’s no N? Oh dear…i’ve been reading it wrong forever then
now, how the heck am I supposed to pronounce termius. Sounds like a disease.
Martin Haeberli
@mhaeberli
Jun 01 2017 04:53
if i run oref0-rerun and it gets updated .js from github (public dev, not my own branch) do I need to re-run npm install?
also I’m having good luck with blink (and it also does mosh) (and you can install mosh on the Edison as well, the usual apt-get install mosh
Scott Leibrand
@scottleibrand
Jun 01 2017 05:18
mosh is super awesome. I'll have to check out blink
Martin Haeberli
@mhaeberli
Jun 01 2017 05:20
:+1:
Scott Leibrand
@scottleibrand
Jun 01 2017 05:21
hmm, $20: is it worth it?
Martin Haeberli
@mhaeberli
Jun 01 2017 05:21
i kinda think so but I can justify it to the IRS … :-)
so if I say OPENAPS_FIELDS status-symbol status-label iob meal-assist rssi to the nightscout instance, what will my pill look like?
pill.tiff
mine just looks like this - what am I missing?
Martin Haeberli
@mhaeberli
Jun 01 2017 05:29
@scottleibrand while we’re on the topic, I’m having a little problem with my reverse ssh tunnel … I have a cron script that repeatedly connects to a bridge public IP address if the connection has dropped, but I haven’t been able to connect back into the Edison via that node since the first day or so - I just get ‘connection refused’ … I remember I think you posted a recipe here in this thread
katie disimone
@Kdisimone
Jun 01 2017 05:44
@mhaeberli OpenAPS pill has all the info hidden inside of you hover or click on it
francesc0-cgm
@francesc0-cgm
Jun 01 2017 06:05
@scottleibrand changed in src/lib/meal/total.js on my rig the cob decay to the old 1.5*etcetc
to make change effective could i rerun the script or i have to run it like the first time?
francesc0-cgm
@francesc0-cgm
Jun 01 2017 06:11
@tim2000s too
Tim Street
@tim2000s
Jun 01 2017 06:16
That will work from there @francesc0-cgm
francesc0-cgm
@francesc0-cgm
Jun 01 2017 06:18
so only with rerunning script?
PieterGit
@PieterGit
Jun 01 2017 07:09
@scottleibrand could you have a look at https://github.com/openaps/oref0/pull/470#pullrequestreview-38515433 when you have some spare time?
Anybody interested in testing the new token based authentication features for Nightscout? I think I need one or more volunteers to testdrive the new and more secure mechanism. I hope to do finish the polishing and the merge resolving with current dev this weekend.
Miiko
@tw_miiko_twitter
Jun 01 2017 07:54
@danamlewis @scottleibrand @all Hi, after setting of oref0, so end phase 2, I can not see kernel.config in the root derectory. Is something missing?
Tim Street
@tim2000s
Jun 01 2017 10:46
@francesc0-cgm once you've changed total.js there is no need to do anything else. It works every time the loop runs.
francesc0-cgm
@francesc0-cgm
Jun 01 2017 10:59
Ah ok
So you have not to rerun any script if you change the code. It will work with your change made in src dir
@scottleibrand now i will try to edit smb to make it effective with x23...found how to use it when basal is >0.15 but i can't find where to tell him less than 0.10 per time
Martin Haeberli
@mhaeberli
Jun 01 2017 11:45
@Kdisimone thx!
dan lightner
@danimaniac
Jun 01 2017 14:10
@Jane83 I find that my OpenAPS rig 1) has not significantly lowered my a1c...I was at 6.0-6.2 already...but 2) it narrows my standard deviation (fewer high and low swings) and stabilizes my graph, 3) has significantly increased my time in range, 4) has nearly eliminated the all day / all night effort necessary to achieve tight control (I would estimate my effort has been reduced by 70-80%), and 5) makes me the cool kid on the block! I don't know anyone else in the area with a functioning APS...neither does my Endo!
fidoman
@efidoman
Jun 01 2017 14:52
My endo apt went very well. He had a college medical student shadowing him so I was a little worried about how that would go. However, he was very supportive, had lots of good questions that I happily answered and demonstrated my NS on my iphone / watch. I showed them the rig and talked about the safety features and all the cool functional features. He changed my prescriptions without issue and documented on my chart that I was successfully using a DIY closed loop system. He urged me to keep it quiet because the FDA could force me to surrender the device (I told him that the founding developers of the system were already in communications with the FDA). He then said that Medtronic may try to stop it by changing the protocols. His reasoning was because someone could have a negative experience on DIY closed loop and give closed loop a bad reputation across the board. He ended by saying congratulations, you are using a system that is well ahead of the latest FDA approved systems coming out and he added "Be careful."
Cameron Chunn
@applehat
Jun 01 2017 15:29
lol @ "having to surrender the device to the FDA".
Did you tell him Medtronic already stopped it? =(
Ryan Dalisky
@rdalisky
Jun 01 2017 15:36
@ericfeibelman Sounds like a medical student talking out their bunger.
dan lightner
@danimaniac
Jun 01 2017 15:36
@ericfeibelman I want an FDA badge...Come with me, sir. You are in possession of a non-FDA approved medical device. I am authorized to confiscate such by force, if necessary.
Cameron Chunn
@applehat
Jun 01 2017 15:58
Sir. SIR. Are you transmitting data through the air in a none FDA approved manner?
FDA / FCC Joint Task Force to deal with these DIABETICS and their UNAPPROVED DEVICES.
Think of the children!
@danamlewis is actually the leader of a diabetic crime syndicate with the evil goal of making peoples lives easier.
Ryan Dalisky
@rdalisky
Jun 01 2017 16:02
🤘Breakin' the law!!
Cameron Chunn
@applehat
Jun 01 2017 16:04
This message was deleted
Diadon81
@Diadon81
Jun 01 2017 16:59
Tommorow I am going to Endocrine Science Center to show my Openaps rig to head of Child with Diabet Cure Department.
in Russia
My doctor just share info about me and my compensation results to him and he wants to see me in a person :)
Diadon81
@Diadon81
Jun 01 2017 17:05
sorry for my terrible english:) romanian language group so painful for russian speaking man :)
Cameron Chunn
@applehat
Jun 01 2017 17:07
@Diadon81 thats awesome! And dont worry, you're english is fine =)
Diadon81
@Diadon81
Jun 01 2017 17:08
@ericfeibelman I think community should develop pump which can be used with openaps instead of medtronic.
So tomorrow I will present my rig to one of the biggest Endocrine Science center in Russia, so, I am little worry :)
John Seaber
@jseaber
Jun 01 2017 17:12
@Diadon81 - Totally agree. I develop electronics for a living and have given this some thought, especially after it took weeks to find a used Minimed 722. Developing an opensource insulin pump would be a huge undertaking, as well as a massive safety risk. To sell in the US, we'd need FDA approval, meaning a company or organization would need to develop the project independently of OpenAPS.
Diadon81
@Diadon81
Jun 01 2017 17:13
Unfortunatelly in Russia, we only good in developing weapons and rockets:) all other research in a very poor condition
John Seaber
@jseaber
Jun 01 2017 17:14
:)
Diadon81
@Diadon81
Jun 01 2017 17:15
@jseaber Why you need FDA approval to sell a opensource pump? Is it prohibited to sell something if I informed about risks and agree with that?
John Seaber
@jseaber
Jun 01 2017 17:16
It's fine to build an opensource project. The US government restricts the sale of certain devices, like insulin pumps.
Diadon81
@Diadon81
Jun 01 2017 17:16
What if you split this pump on two parts (which can be easily assembled) and sell it as a simple electronic components?
John Seaber
@jseaber
Jun 01 2017 17:16
So the design is okay. Selling it would be illegal.
I think it would be best to seek FDA approval. Safety is important!
Diadon81
@Diadon81
Jun 01 2017 17:17
Like a gas and a bottle it is not prohibited, but all together is a Molotov coctail:)))
Cameron Chunn
@applehat
Jun 01 2017 17:17
A pump would have to be handled a lot like the current APS rigs are.
A guide to taking off-the-shelf stuff and making something with them
Diadon81
@Diadon81
Jun 01 2017 17:18
Yes, splitting this device on easily assembled components
Cameron Chunn
@applehat
Jun 01 2017 17:18
I've considered playing with the idea - even found a supplier in china who would sell the same stepper motor gear box that medtronic uses.
Diadon81
@Diadon81
Jun 01 2017 17:20
yep. I also dig in this topic, motor is most important component, so I think using 3d printing for prototyping isn't so hard to make a pump.
Matthias Granberry
@mgranberry
Jun 01 2017 17:20
If you were designing a device like an insulin pump for wider distribution w/o FDA approval you would want to market it as a research infusion pump (e.g. for delivery of drugs to mice, pets, or maybe high-control hydroponics)
Cameron Chunn
@applehat
Jun 01 2017 17:21
^ this. I also wouldn't really be interested in making and selling pumps - I'd be more interested in designing one and giving a guide to how to make your own.
John Seaber
@jseaber
Jun 01 2017 17:22
We could go the route of RaspberryPi: Didn't they build the first RPI, and only later seek FCC approval?
Diadon81
@Diadon81
Jun 01 2017 17:23
@applehat but you should consider a risk, that not evererybody can assemble a pump or worse, assemble it right.
Cameron Chunn
@applehat
Jun 01 2017 17:23
In theory, you could stack s GPIO board on our edison rigs to drive a pump motor.
@Diadon81 that same risk exists with the OpenAPS rigs - in both a hardware and software sense.
John Seaber
@jseaber
Jun 01 2017 17:24
It's a greater risk IMO. Firmware would need to be flawless.
Diadon81
@Diadon81
Jun 01 2017 17:24
@applehat but there are some safety caps, that reduce a risk, how do you plan to put these caps in a pump, especially for wrong dosage?
John Seaber
@jseaber
Jun 01 2017 17:24
Exactly.
Cameron Chunn
@applehat
Jun 01 2017 17:24
Tho improperly dosing insulin is a higher risk - its still the same concept - and part of the reason I'd much prefer to design something and give away the plans then have liability fall on me when someone does something stupid
There is a very real reason that no one has made an Open Source pump yet.
Both issues in accuracy and liability.
It's still something i'd dabble with, tho.
Diadon81
@Diadon81
Jun 01 2017 17:26
@applehat agree with you
So if reuse edison for pump control, we need a motor a 3d printer :)
and
Cameron Chunn
@applehat
Jun 01 2017 17:27
Stepper motor + gear box
John Seaber
@jseaber
Jun 01 2017 17:27
I have a 3D printer and multi-axis CNCs.
Dana Lewis
@danamlewis
Jun 01 2017 17:28
Ya'll might want to create a separate channel for this kind of discussion
Cameron Chunn
@applehat
Jun 01 2017 17:28
Oops. You're right @danamlewis. Sorry!
You guys can join https://gitter.im/medtronic-flash/Lobby if you want - The discussion there has evolved into a general "alternative game plan to old medtronic pumps" discussion.
Dana Lewis
@danamlewis
Jun 01 2017 17:30
:+1:
Kichul Kim
@nic2mit
Jun 01 2017 17:46
@ericfeibelman thanks for sharing your endo visit. I guess we all have different experiences. For me, it was my endo who told me about openaps back in this past Feb and encourage me to build the Rig. He has my NS website and still monitor my process time to time lol. I think openaps is getting really good reputation within diabetic endo community. Thanks to @danamlewis and @scottleibrand and others!
fidoman
@efidoman
Jun 01 2017 17:55
@nic2mit - Nice and agreed. How do you do a thumbs up? +1
I'm getting the following error repeatedly in my logs - on the dev branch. Any ideas of what I should check? Starting supermicrobolus pump-loop at Thu Jun 1 13:00:37 EDT 2017 with 2 second wait_for_silence:
MDT CGM configured; not wct/smb-enacted.json’: No such file or directory
ValueError: byte must be in range(0, 256)
grep: enact/smb-enacted.json: No such file or directory
ValueError: byte must be in range(0, 256)
cp: cannot stat ‘enact/smb-enacted.json’: No such file or directory
enact/smb-enacted.json: cat: enact/smb-enacted.json: No such file or directory
Error, retrying
Sleeping 2;
Dana Lewis
@danamlewis
Jun 01 2017 17:57
Should resolve after the first SMB
And to do a thumbs up or any emoji, start by typing a colon : and the emoji menu should pop up. A thumbs up is : +1 : without spaces between the colons and the content
fidoman
@efidoman
Jun 01 2017 18:00
I've SMB'd all morning long before this happened over lunch
@danamlewis - :+1:
Dana Lewis
@danamlewis
Jun 01 2017 18:00
Hm, really? And smb'd after this update occurred?
Wonder if it is related to the mdt CGM related fix that was added. Are you on mdt or something else?
Diadon81
@Diadon81
Jun 01 2017 18:01
Just one question, If I put max_iob = 0 is it means, that pump will never try to increase a temp basal rate and only stop delivery on low?
Dana Lewis
@danamlewis
Jun 01 2017 18:02
@Diadon81 it primarily will low temp, but it will high temp you if you have negative iob, only to the point that iob returns to zero. Then it will stop high temping.
fidoman
@efidoman
Jun 01 2017 18:02
@danamlewis - nope. SMB should have happened after this error occurred, but I had to reboot rig to get SMB working again.
Sgoldberg89
@Sgoldberg89
Jun 01 2017 18:03
Oh hello
Diadon81
@Diadon81
Jun 01 2017 18:03
@danamlewis in that case, how to use rig only for autotune, don't allow it to set a temp basal?
fidoman
@efidoman
Jun 01 2017 18:03
@danalewis- MDT and enlite sensor connected to the pump
Sgoldberg89
@Sgoldberg89
Jun 01 2017 18:04
(I never know how to announce that I'm about to ask a question, feel like I'm awkwardly standing in a room)
Dana Lewis
@danamlewis
Jun 01 2017 18:04
@Diadon81 you can not enable cron
Cameron Chunn
@applehat
Jun 01 2017 18:04
My endo was confused and didn't understand what my rig really did. Then I talked to her husband who was super excited about it and talked about Bigfoot Biomedical.
Sgoldberg89
@Sgoldberg89
Jun 01 2017 18:05
ANYWAY, are there instructions for using the CGM with the edison offline? I mean I know there are, but I can't find them, so if someone is like, in the middle of doing it and wants to post the link I'd be super duper grateful
Diadon81
@Diadon81
Jun 01 2017 18:05
@danamlewis this is a good point, thank you
Sgoldberg89
@Sgoldberg89
Jun 01 2017 18:06
just kidding
the big tab called offline looping sort of escaped me
Dana Lewis
@danamlewis
Jun 01 2017 18:08
;)
fidoman
@efidoman
Jun 01 2017 18:15
Just my opinion on offline looping. If you have a choice and are able to consider enlite CGM with MDT pump it is really easy to do with enlite CGM connected to the MDT pump. I let that happen all the time for commute rides, etc. It is automatic.
Tom Boudreau
@tomasboudr
Jun 01 2017 18:16
Random Q. Can the explorer board red power light be turned off?
Summer pants are thinner. And I'm glowing.
Also just tested oref1 by eating many goldfish without bolusing. Felt like freedom!
Dana Lewis
@danamlewis
Jun 01 2017 18:18
:fish: :confetti_ball:
Sgoldberg89
@Sgoldberg89
Jun 01 2017 18:20
@ericfeibelman Not sure I understand, but enlite sensors are my kryptonite... or maybe I'm their kryptonite... well either way they don't seem to work on me so :(
Also, @danamlewis I just got out my USB cables to set this up only to realize I don't have those cables! So to be continued
Scott Leibrand
@scottleibrand
Jun 01 2017 18:26
@ericfeibelman that ValueError: byte must be in range(0, 256) might be related to a max basal >= 6.4 U/hr - how high is yours?
Tom Boudreau
@tomasboudr
Jun 01 2017 18:30
@Sgoldberg89 I used to use enlites. Have you tried them on your upper arm. Made a 100% difference for me.
I use my G5 I have now. there as well
fidoman
@efidoman
Jun 01 2017 18:49
@scottleibrand - Yes, it was. I just reset it to 6.35 U/hr which should be fine for me as a max anyway.
Scott Leibrand
@scottleibrand
Jun 01 2017 18:50
especially with SMB also available. :)
dan lightner
@danimaniac
Jun 01 2017 18:51
@tim2000s @peterfulvi @tatigger @others Can we / how do we use Autotune with xDripAPS data. I am BT tethered to an android phone using xDrip for offline monitoring and xDripAPS for offline looping. 100% offline. My rig is never online anymore and I do not upload anything to NS although I have a NS site already setup. Autotune is set to run nightly via cron but never succeeds as it attempts to pull data from NS. I would like to run autotune offline.
Cameron Chunn
@applehat
Jun 01 2017 18:53
@danimaniac you could, in theory, remove xDripAPS and install a full copy of Nightscout on your edison...tho I'm not sure if anyone has done this and Im also not sure if the edison has enough power to accommodate it.
Scott Leibrand
@scottleibrand
Jun 01 2017 18:53
@danimaniac you'll be trailblazing then. originally the autotune tools were designed to also work on pumphistory.json and glucose.json, so that part might still work, but you'll need to do the plumbing to replace the autotune.sh script that pulls all the stuff from NS etc.
or, yeah, a full local NS might also work. it might require adding an SD card, though
Cameron Chunn
@applehat
Jun 01 2017 18:55
I'd be interested in blazing that trail with you @danimaniac. I've always liked the idea of my loop being completely self contained offline, and only uploading data for remote monitoring (or potentially for sharing to other rigs if I get to that point).
dan lightner
@danimaniac
Jun 01 2017 18:58
@applehat @scottleibrand Well, I am not much of a trailblazer in this regard. I can follow directions, though, so I was hoping another xDripAPS'er had undertaken this. I also only have 1 rig and it works great as is...don't want to bork it! Maybe a second test rig is in my future?
Cameron Chunn
@applehat
Jun 01 2017 19:00
@danimaniac I am currently kicking around an idea in my head that I plan to pursue that would replace xDripAPS with a lighter weight process and remove the need to run a Bluetooth PAN tether. I could probably look into doing offline autotune at the same time.
dan lightner
@danimaniac
Jun 01 2017 19:05
@applehat Count me as interested. I will help however I can...I just don't know how much assistance I can provide. I love the idea of eliminating the online NS site and operating completely offline...including autotune.
Cameron Chunn
@applehat
Jun 01 2017 19:10

My idea involves running a bluetooth "server" on the RIG side that will accept new glucose values, and then a small app on the android side that listens for xDrip's event that it has a new blood glucose value, and pushing it to the rig. The would illuminate the need for an internet connection on the rig when its away from wifi.

The idea could, in theory, go both ways - and the app could poll data from the rig to push into nightscout online if needed.

I just hate having to keep a full Bluetooth PAN connection open (because it hasn't been the most reliable thing for me), and I think a lightweight communication protocol that wasn't always active could be beneficial.

dan lightner
@danimaniac
Jun 01 2017 19:15
@applehat I have just the opposite experience. Once I got my BT tether properly setup it has never once dropped connection. I completely disabled my wpa_supplicant file so my rig never even looks for wifi connections. In the meantime...I have read the docs concerning xDrip uploading to xDripAPS on the rig AND simultaneously uploading to NS. I can't seem to get it configured properly, though. I would settle for that as a temporary autotune solution.
Cameron Chunn
@applehat
Jun 01 2017 19:17
Oh, thats easy @danimaniac. In xDrip (im using xDrip+ recent nightly, so make sure your up to date) under "BaseURL", I litterally just have it set as http://secret@myaddress/api/v1/ http://secret@192.168.44.107:5000/api/v1
just put both upload addresses with a space in between
and xDrip+ will use both.
fidoman
@efidoman
Jun 01 2017 19:19
@Sgoldberg89 - Yep, we are all different with different experiences. My point was that using an enlite sensor connected directly to an x22 MDT pump will automatically loop offline without any special configuration. Sorry you can't use them. I've found them to be mostly spot on. Never tried on the upper arm.
Diadon81
@Diadon81
Jun 01 2017 19:20
@ericfeibelman I am also using a rig with Enlite and for me it is most comfortable configuration.
francesc0-cgm
@francesc0-cgm
Jun 01 2017 19:25
@scottleibrand i edited the .js on src to get cob till 1.5 dia. Then i found out the line to enable smb on custom basal value. I don't find
Sorry i can't find where enabling smb lower than 0.1
However I found the line where alghorim talks about smbing when 0.1 is needed...if i edit it to 0.05 will it work?
Diadon81
@Diadon81
Jun 01 2017 19:29

start to getting following error on my dev rig

BG data is too old (it's probably this), or clock set incorrectly.  The last BG data was read at Thu Jun 01 2017 22:17:00 GMT+0300 (MSK) but your system time currently is Thu Jun 01 2017 22:27:07 GMT+0300 (MSK)
  "reason": "BG data is too old (it's probably this), or clock set incorrectly.  The last BG data was read at Thu Jun 01 2017 22:17:00 GMT+0300 (MSK) but your system time currently is Thu Jun 01 2017 22:27:07 GMT+0300 (MSK)"
Checking system clock against pump clock:
Checking deliverAt: null is within 1m of current time: Thu Jun  1 22:27:10 MSK 2017
date: invalid date ‘null’
Error, retrying
Sleeping 3;

any suggestion, clock on pump and rig less the one minute difference.

Scott Leibrand
@scottleibrand
Jun 01 2017 19:37
@francesc0-cgm this is code: we can't talk in generalities. :) can you paste me a link to exactly which line you're talking about?
@Diadon81 looks like the problem is with your BG data being stale, not with rig/pump clocks
dan lightner
@danimaniac
Jun 01 2017 19:40
@applehat I will give the multiple REST API's a try. As I read the docs, the multiple REST API's was for connecting to two "personal" networks...not one personal network and one mobile network connection.
Cameron Chunn
@applehat
Jun 01 2017 19:41
its all HTTP protocol. Your phone will route the requests correctly. xDrip won't care.
dan lightner
@danimaniac
Jun 01 2017 19:43
...sounds like a simple PR is my future...THEN the offline autotune later?
Cameron Chunn
@applehat
Jun 01 2017 19:44
Let xDrip upload to both for now and rely on the internet a bit for autotune... then we can see what can be done in the future when I have time to tinker.
Thats the quick fix that works right now
Dana Lewis
@danamlewis
Jun 01 2017 20:04
let me know if anyone’s pulled dev this morning/today - I want to see if something correctly showed up in preferences or if it needs a fix
Tim Street
@tim2000s
Jun 01 2017 20:19
@danimaniac @cameronchunn I haven't tried but your xdripaps treatments data has an ipaddress and port. I'd go and point Autotune at that and see how it gets on.
francesc0-cgm
@francesc0-cgm
Jun 01 2017 20:20
@scottleibrand
'''
''' // only allow microboluses with COB or low temp targets, or within DIA hours of a bolus
// only microbolus if 0.1U SMB represents 20m or less of basal (0.3U/hr or higher)
if (microBolusAllowed && enableSMB && profile.current_basal >= 0.3) {
// never bolus more than 30m worth of basal
maxBolus = round(profile.current_basal/2,1);
// bolus 1/3 the insulinReq, up to maxBolus
microBolus = round(Math.min(insulinReq/3,maxBolus),1);
'''
Sorry i'm from the phone...
'0,3' should become 0.15
Scott Leibrand
@scottleibrand
Jun 01 2017 20:24
yeah, that part is right I think
that line rounds the microBolus amount to 1 significant digit (the nearest 0.1U)
sounds like you want to multiply by 2 inside the round() and then divide by two outside it, in order to get it rounded to the nearest 0.05U
francesc0-cgm
@francesc0-cgm
Jun 01 2017 20:26
Then also this
// if no microBolus required, snoozeBG > target_bg, and lastCOBpredBG > target_bg, don't set a zero temp
if (microBolus < 0.1 && snoozeBG > target_bg && lastCOBpredBG > target_bg) {
durationReq = 0;
It is to be edited to microbolus<0.05
It has^
cameronrenwick
@cameronrenwick
Jun 01 2017 20:26
anyone have any advice for me? I'm away and trying to log my rig into a wifi network. There is a wifi name, a username and a password. In our wpa_supplicant file I've never needed to add a user name in addition to the wifi and pw. Any suggestions on how to do this?
where should the user name go in the wpa_supplicant file?
francesc0-cgm
@francesc0-cgm
Jun 01 2017 20:28
@scottleibrand didn't understand how to edit this to make it bolusing 0.05...
francesc0-cgm
@francesc0-cgm
Jun 01 2017 20:28
I found them but editing code lines is over my actual knowledge
dan lightner
@danimaniac
Jun 01 2017 20:29
@tim2000s Yes, that's how I verify that xDripAPS receives the data it needs. I point my phone's browser to my rig's <localipaddress>:5000/api/v1/entries. You are suggesting it MIGHT be as simple as editing the autotune.sh script to pull from the local ip address instead of the online NS instance? I think that's what @scottleibrand meant above when he mentioned doing the plumbing to replace the autotune.sh script?
Scott Leibrand
@scottleibrand
Jun 01 2017 20:29
@francesc0-cgm that line I posted reads microBolus = round(Math.min(insulinReq/3,maxBolus),1); - walk through that step by step and see if you can explain exactly what it does, and my *2, /2 suggestion should make more sense
@danimaniac what @tim2000s is suggesting is simpler: have the rig upload to the xDripAPS on localhost and also to your real NS, and then have autotune pull from the real NS
francesc0-cgm
@francesc0-cgm
Jun 01 2017 20:32
Ok i will try...thanks
Tim Street
@tim2000s
Jun 01 2017 20:35
I'm suggesting if you're offline completely you might want to point at the local database. I upload to NS and feed the local db so have left Autotune to run from NS.
cameronrenwick
@cameronrenwick
Jun 01 2017 20:42
@danamlewis thanks for the pointer to the wifi... not having any joy tho. No typos in my edits to the file.. any other suggestions? I've got wifi on my phone and macbook with the supplied wifi name, username and pw so I know they're functioning
rsilvers129
@rsilvers129
Jun 01 2017 20:44
I entered 5 carbs twice using IFTTT and I noticed the logs said "Removed duplicate bolus/carb entries:1"
francesc0-cgm
@francesc0-cgm
Jun 01 2017 20:57
Sorry @scottleibrand i can't understand how to manage this...i red about the round math.min but i can't understand how it works with your x2 and then /2
Atm is current basal (so min is 0.3) / 3
Then ,1
If i put 0.15 as min basal for smb working...and i multiply x 2 it will become 0.15x2 = 0.3 then i have to put ,2 to divide it by 2? It will be again 0.15
fidoman
@efidoman
Jun 01 2017 21:05
Is this what your looking to do with the 2 /2? microBolus = (round(Math.min(insulinReq 2/3,maxBolus),1)/2) ... the reason you want to do that (I think) is because the round function only goes to the .1 decimal because the 2nd parameter to round is 1. You can't make that parameter 2 because then you would get numbers like .17 which are invalid. Therefore the multiple by 2 inside the round and then the divide by 2 outside the round. This gives you these possibilities 0.05, 0.1, 0.15, etc. Make sense?
francesc0-cgm
@francesc0-cgm
Jun 01 2017 21:08
Yes it is...i believed it limits only to 1 value...it will be for smb 0.5 with a x23
To try it on a kid
Because 0.1 is too much
@ericfeibelman why insulinreq 2/3 ?
It is muliplied per 2 then divided by 3?
fidoman
@efidoman
Jun 01 2017 21:12
that is the multiply by 2 inside the round to allow the smaller .05 increments. then you compensate for that multiply by two by dividing by two outside the round. Mathematically if you take a number and both multiply by two and divide by two it will cancel each other out. But doing the multiply by 2 there inside the round allows the smaller 0.05 increments.
francesc0-cgm
@francesc0-cgm
Jun 01 2017 21:12
Don't understand why to divide it by 3
fidoman
@efidoman
Jun 01 2017 21:13
That divide by 3 was already in the code. It is how the SMB works ... only give 1/3 of insulinReq dosage at a time
francesc0-cgm
@francesc0-cgm
Jun 01 2017 21:14
Ahhh ok got it
Now i try to change it in src and i will try without a pump
Sorry without pump attached
fidoman
@efidoman
Jun 01 2017 21:14
yes, good idea. The comment above was missing a "" for the 2/3. This markup language is removing my times sign
francesc0-cgm
@francesc0-cgm
Jun 01 2017 21:15
? Don't get it
I have to write "2/3" ?
fidoman
@efidoman
Jun 01 2017 21:16
asterick was removed from my code suggestion
francesc0-cgm
@francesc0-cgm
Jun 01 2017 21:16
Ah ok so i it is insulinreq"asterisk"2
/3
fidoman
@efidoman
Jun 01 2017 21:17
yes "asterisk 2/3" but this text field doesn't allow me to write an asterisk
francesc0-cgm
@francesc0-cgm
Jun 01 2017 21:17
I know
fidoman
@efidoman
Jun 01 2017 21:17
yep - sorry about that
francesc0-cgm
@francesc0-cgm
Jun 01 2017 21:17
Thank you very very much
You was so kind
By myself i won't be able to understand that
fidoman
@efidoman
Jun 01 2017 21:18
yw - hope it works in your test
francesc0-cgm
@francesc0-cgm
Jun 01 2017 21:19
Hope me too
It could be great for kids with high sensitivity to use smb
Scott Leibrand
@scottleibrand
Jun 01 2017 21:45
@ericfeibelman you can put code inside backticks (`) to avoid it rendering the * as an italics indicator
so for example microBolus = (round(Math.min(insulinReq*2/3,maxBolus),1)/2)
and thanks for the help. :)
sdneufer
@sdneufer
Jun 01 2017 22:56
@applehat I have been wondering if for offline xdrip would it be possible to plug wixel into the explorer board uart port? I have a wixel plugged to the pi I originally used for openaps which sends readings to my Android however I've been wondering if it would be possible to write something to read from the edison uart port. Then perhaps it could be offline or upload to ns. Could this work?
rsilvers129
@rsilvers129
Jun 01 2017 23:23
What does it mean when the yellow triangle is lit on my OpenAps pill? Everything seems to be working with the loop.
rsilvers129
@rsilvers129
Jun 01 2017 23:27
I had the OPENAPS_WARN variable set to true. I just set it to false and the yellow triangle went off. I see no failures. It was updating and working.
Mark Flasher
@MarkFlasher_twitter
Jun 01 2017 23:31
Does anyone here have the link to th e omnipod hacking group?
Mark Flasher
@MarkFlasher_twitter
Jun 01 2017 23:39
@danamlewis thanks.
fidoman
@efidoman
Jun 01 2017 23:41
@scottleibrand :+1:
Dana Lewis
@danamlewis
Jun 01 2017 23:42
http://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-4/oref1-features.html <— code is still in dev, but put the oref1 docs in the regular docs so people have no excuse not to read :D if they want to test out dev-related advanced features.