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

1st
Oct 2017
Deweyoxberg
@Deweyoxberg
Oct 01 2017 00:01
here goes test :P
sdneufer
@sdneufer
Oct 01 2017 00:01
fingers crossed!
Deweyoxberg
@Deweyoxberg
Oct 01 2017 00:02
HOLY BUUTS
WOOOOOOOOOOOOOO
sdneufer
@sdneufer
Oct 01 2017 00:02
sounds like it worked!
Deweyoxberg
@Deweyoxberg
Oct 01 2017 00:02
RAWRRRRRR
I could hug you so hard right now
sdneufer
@sdneufer
Oct 01 2017 00:04
Glad it works. How to document that may be interesting if it works as is for some and not for others. Guess it goes under if you get this error try this..
Deweyoxberg
@Deweyoxberg
Oct 01 2017 00:04
##/usr/local/libexec/bluetooth/bluetoothd &
sudo killall bluetoothd
sleep 5s
sudo /usr/local/bin/bluetoothd --experimental
sleep 15s
sudo hciconfig hci0 name $HOSTNAME
#sleep 20

##hciconfig hci0 up
##hciconfig hci0 noscan
whoa what the
that's... not a code block lol
there we go
Paul Dickens
@thebookins
Oct 01 2017 00:06
Hi All, in the process this morning of updating to the latest dev, with the exponential curves. Just wondering what to do (if anything) with our DIA. It's currently set at 4 hours - any reason to change with with the new curves?
Deweyoxberg
@Deweyoxberg
Oct 01 2017 00:13
@sdneufer : for the hci load connection parameters failed, bluetoothd can be run with --noplugin=sap
apparently has nothing to do with what was going on xD
katie disimone
@Kdisimone
Oct 01 2017 00:13
Thanks @YYCMichael and @sdneufer I’m coming in late to the convo. What was the issue originally? BT wasn’t picking up when known WiFi network was left?
Deweyoxberg
@Deweyoxberg
Oct 01 2017 00:15
@Kdisimone : short version is that on my current setup bluetooth on boot was not properly started. It doesn't throw any errors, but outofbox can't pickup wifi after setup.
using sudo killall bluetoothd, then starting with sudo/usr/local/bin/bluetoothd --experimental, then sudo hciconfig hci0 name $HOSTNAME works, but on power loss, the fix would be lost.
so all day was trying to figure out a way to have a small script launch after boot to restart bluetooth
But then @sdneufer pointed out I could throw those commands into /usr/local/sbin/bluetooth_patchram.sh which is loaded on boot, and comment out the existing bluetooth load which didn't work
I added a few sleeps in there as well, since the documentation does say to wait after starting bluetoothd otherwise errors get thrown
and now it works :D And I'm chairdancing and wiggling my butt all over the house like a super excited and proud doggo would :P (disclaimer: I am not a doggo)
katie disimone
@Kdisimone
Oct 01 2017 00:20

@YYCMichael interesting. There’s some weird WiFi issues that I’ve seen some people have lately. Wondering if it would help them.

You were able to get WiFi during setup script? But then WiFi wouldn’t go after that? Or did WiFi only fail when you added more networks.

Deweyoxberg
@Deweyoxberg
Oct 01 2017 00:21
uh... this was bluetooth, not wifi :)
That's my next problem....
I can get home wifi which I am on right now, but I cannot get mobile hotspot wifi. No DHCP, last I read.
Curious that you mention wifi failing after adding more networks... currently works with 4, but home only.
Still troubleshooting if it's because of signal frequency (5 vs 2.4) or something else.
Dana Lewis
@danamlewis
Oct 01 2017 00:51
@primags you will generally get better outcomes with at least a general carb estimate. Setting phone and or watch buttons are great for this. (Or use bolus wizard.) I usually enter carbs via phone/watch and bolus using easy bolus with the same mentality as you. But giving the system carb notice will help realtime with picking up where you need more insulin (or overdid it) and also helps with things like autosens and Autotune, so it can correctly attribute data to the right places
So not overt bite, but you won't get the most benefit out of it that you could.
Deweyoxberg
@Deweyoxberg
Oct 01 2017 01:07
@sdneufer : Field test confirmed :D bluetooth picked up when wifi was lost
Huzzah!
sdneufer
@sdneufer
Oct 01 2017 01:09
Happy its working!
Deweyoxberg
@Deweyoxberg
Oct 01 2017 01:17
May not be fixed. Able to get logs, but uploads to nightscout are a nogo. Curiouses...
Getting BG and pump data though
Bolus, nope.
Need to chill out wifi scanning too, every 8 seconds is crazy.
Deweyoxberg
@Deweyoxberg
Oct 01 2017 01:28
welps, there's the bolus. took almost 20 minutes to pop up
sdneufer
@sdneufer
Oct 01 2017 01:30
Data is on the edison quicker than in NS. I check my pump for what it last bolused or what is set as a temp as NS takes longer to show.
Deweyoxberg
@Deweyoxberg
Oct 01 2017 01:30
There a way to check the logs via papertrail for a bolus?
Scott Leibrand
@scottleibrand
Oct 01 2017 01:31
@thebookins if you leave DIA at 4h it'll ignore that and use 5h. You'll probably find it works best with larger/stacked boluses to use 6 or 7 hours though. That won't affect the peak much (peak time is set in preferences if necessary), just the tail.
@YYCMichael manual boluses only show up via IOB, plus the last bolus time if you're using SMB. The pumphistory contents aren't explicitly logged.
Deweyoxberg
@Deweyoxberg
Oct 01 2017 01:34
image.png
@scottleibrand : do you mean this:
Glenn Primack
@primags
Oct 01 2017 01:35
Just a heads up for those with IOS devices, I have an Iphone and was having trouble with Bluetooth tethering since the most recent update (IOS 11). IOS 11.01 just came out and fixes the issues so far. The symptom was that it would connect when wifi went down but would never drop the connection after wifi resumed.
@danamlewis That makes a lot of sense, thank you! I have an apple watch, I'll try to setup the carb recording.
Dana Lewis
@danamlewis
Oct 01 2017 01:53
@primags :+1: ifttt->NS is great for that.
Glenn Primack
@primags
Oct 01 2017 01:54
I'll check it out :smile:
Deweyoxberg
@Deweyoxberg
Oct 01 2017 02:03
any idea where the wifi scan interval is? Watching the logs fly and seems like edison is scanning like crazy while on BT. Want to chill it out a bit. Sometimes even scans three times in the same second.
like, this is crazy:
Sep 30 19:03:40 localhost dhclient: DHCPDISCOVER on wlan0 to 255.255.255.255 port 67 interval 13
Sep 30 19:03:42 localhost dhclient: DHCPDISCOVER on wlan0 to 255.255.255.255 port 67 interval 5
Sep 30 19:03:47 localhost dhclient: DHCPDISCOVER on wlan0 to 255.255.255.255 port 67 interval 7
Settle down wifi! :P
or is that edison finding local wifi's and saying "found something!"
Deweyoxberg
@Deweyoxberg
Oct 01 2017 03:18
Curious... rig is rebooting randomly with nothing in log as to why. hrm... /thinking
Paul Dickens
@thebookins
Oct 01 2017 03:19
thanks @scottleibrand , didn't realise the new curves used 5 hrs minimum. On that, my prefs have "bilinear". Is there documentation somewhere for this different curve options?
Deweyoxberg
@Deweyoxberg
Oct 01 2017 03:48
Reboot problem fixed... not enough power :P
Paul Dickens
@thebookins
Oct 01 2017 03:52
Thanks @danamlewis
Deweyoxberg
@Deweyoxberg
Oct 01 2017 03:57
Any fixes for pump status not refreshing in NS? Can see BG data, can see logs, last pump data 2 hours ago
Deweyoxberg
@Deweyoxberg
Oct 01 2017 04:21
Keep seeing this, related? pump-loop.log: Old pumphistory: process is running and remove the file manually to continue.'
Started up roughly 2 hours ago
Scott Leibrand
@scottleibrand
Oct 01 2017 05:20
That's the last line of a git lockfile error.
If you're using dev, re-do it with --nogit
Deweyoxberg
@Deweyoxberg
Oct 01 2017 07:39
seems to have self resolved overnight, but pump status information stopped refreshing again
Deweyoxberg
@Deweyoxberg
Oct 01 2017 07:45
At the moment I just see this in logs from papertrail:
Oct 01 00:43:11 DiabeticinJapan pump-loop.log: Listening for 30s silence before mmtuning: .No pump comms detected from other rigs
Oct 01 00:43:42 DiabeticinJapan pump-loop.log: mmtune: "868.390", 5, -52 
Oct 01 00:43:42 DiabeticinJapan pump-loop.log: Starting pump-loop at Sun Oct 1 16:43:41 JST 2017:
Oct 01 00:43:42 DiabeticinJapan pump-loop.log: Waiting up to 4 minutes for new BG: glucose.json newer than pump_loop_completed
Oct 01 00:44:14 DiabeticinJapan pump-loop.log: Radio ok. Listening: .No pump comms detected from other rigs
Oct 01 00:44:38 DiabeticinJapan pump-loop.log: Old pumphistory: AttributeError: 'ReadPumpModel' object has no attribute 'read_status'
Oct 01 00:44:38 DiabeticinJapan pump-loop.log: Error, retrying
Oct 01 00:44:38 DiabeticinJapan pump-loop.log: pump_loop_completed more than 15m old; waiting for 30s silence before mmtuning
and repeats over hours
Deweyoxberg
@Deweyoxberg
Oct 01 2017 07:56
saw this flow through a few minutes ago
Oct 01 00:50:20 DiabeticinJapan pump-loop.log: Old pumphistory: Refreshed pumphistoryWarning: setting mealCOB to 0 because currentDeviation is null/undefined
Oct 01 00:50:20 DiabeticinJapan pump-loop.log:  and meal.json
Oct 01 00:50:25 DiabeticinJapan pump-loop.log: {"carbs":120,"boluses":4,"mealCOB":0,"currentDeviation":null,"maxDeviation":0,"minDeviationSlope":0}
Oct 01 00:50:25 DiabeticinJapan pump-loop.log: BG data is too old (it's probably this), or clock set incorrectly.  The last BG data was read at Sun Oct 01 2017 15:59:30 GMT+0900 (JST) but your system time currently is Sun Oct 01 2017 16:50:25 GMT+0900 (JST)
Is the refresh a failsafe?
PieterGit
@PieterGit
Oct 01 2017 08:42
@scottleibrand : can you rebase openaps/oref0#686 to dev. I accidentally did not select the dev branch and I can't change it after creating
PieterGit
@PieterGit
Oct 01 2017 09:53
@YYCMichael : what kind of pump do you have? There is also something wrong with your rig not getting correct cgm values (BG data is too old) or your pump and cgm time are not in sync
tepidjuice
@tepidjuice
Oct 01 2017 11:05
Referencing openaps/oref0#685. How do people stay in range without manually pre-bolusing? Personally when I eat CHO without a prebolus 10-30 minutes, or at least a high temp basal a la 'eating soon', before I struggle to keep my bgl below 200.
Sorry 180... multiplication by 18 is hard :)
fidoman
@efidoman
Oct 01 2017 11:44
Updating to the latest dev. When I do an npm run global-install, I get a missing file error. I've done the git checkout dev and git pull and those are fine. Any ideas?
npm ERR! Linux 3.10.98-poky-edison
npm ERR! argv "/home/.rootfs/usr/bin/node" "/home/.rootfs/usr/bin/npm" "install"
npm ERR! node v6.11.3
npm ERR! npm v3.10.10
npm ERR! path /root/src/oref0/node_modules/share2nightscout-bridge/node_modules/request/node_modules/hawk/node_modules/hoek
npm ERR! code ENOENT
npm ERR! errno -2
BTW - that file /root/src/oref0/node_modules/share2nightscout-bridge/node_modules/request/node_modules/hawk/node_modules/hoek is missing from my disk.
fidoman
@efidoman
Oct 01 2017 12:25
I got it to work ... I just had to remove oref0 and repull it from scratch. Had to run oref0 setup twice but could've been I keyboard interrupted it. Anyway, all is good now.
Glenn Primack
@primags
Oct 01 2017 12:26
Wasn’t able to find this - what are the symptoms of your azure site hitting its quota limit? Does the entire site stop and it is obvious? I attempted to look at quotas but the information always says 0 percent.
fidoman
@efidoman
Oct 01 2017 12:55
I love this new eSMB feature. Already testing it with breakfast and it does what I was doing manually (meal manual bolusing) but does it much more scientifically and automatically. Really incredible stuff here! I'm going to watch it closely over the next few days and report back on my testing on the PR https://github.com/openaps/oref0/pull/650#issue-256564183
Deweyoxberg
@Deweyoxberg
Oct 01 2017 13:13
@PieterGit : Medtronic 722; how do I fix the time sync?
Tim Street
@tim2000s
Oct 01 2017 13:54
@YYCMichael The code should do that.
On a different note , it sounds like you got Bluetooth to work. Mind pr img into the docs in troubleshooting? Seems like a bluez 5.44 issue.
Deweyoxberg
@Deweyoxberg
Oct 01 2017 13:57
uh, english? lol
it's only 0800; I'm not awake yet :P
jaylagorio
@jaylagorio
Oct 01 2017 14:56
@scottleibrand @danamlewis The plural of anecdote is not data, but: A couple nights ago I announced a 45g meal, no bolus, never went over 175. Aaaamazing. Last night at a wedding I honestly didn't know what to guess so I set a 3 hour eating soon temp at 100 (with eSMB enabled for temps). Ended the night at 275 at 11 AM, didn't get to 150 until 2 AM. Less than amazing.
jquitos
@jquitos
Oct 01 2017 14:57
@YYCMichael ... Would your problem solving of the BT issue work for a rig/phone going out of range of eaxh
Sorry... Each other? I must have a phone where even the BTAutotether doesn't work... Possibly from the rig not restarting BT correctly...so not entirely the fault of the phone. But seems like I have to reconnect via ssh when home on wifi or through serial coms and issue the sudo BT command to get it working again
Deweyoxberg
@Deweyoxberg
Oct 01 2017 15:00
@jquitos : possibly. If you find that manually executing the sudo commands to establish a BT network is the only fix, could give my fix a try. Make sure to backup the affected file(s) first.
sdneufer
@sdneufer
Oct 01 2017 15:06
@tim2000s
Argh on phone. I can do the PR when i get to my computer.
Deweyoxberg
@Deweyoxberg
Oct 01 2017 15:13
I would if I knew what was being asked / how to do it :P
sdneufer
@sdneufer
Oct 01 2017 15:16
jaylagorio
@jaylagorio
Oct 01 2017 15:20

@scottleibrand Possible bug in logging?

Checking reservoir: reservoir level before: 25.9, suggested: 25.9 and after: 25.9

And then it does a 0.3U bolus (which is the right move). Am I misreading the intention of this log line?

sdneufer
@sdneufer
Oct 01 2017 15:29
The wifi section link on the troubleshooting page seems to be broken. Also, @YYCMichael it might make more sense to put the bluetooth changes you did on the bluetooth page?? @tim2000s thoughts? http://openaps.readthedocs.io/en/latest/docs/Customize-Iterate/bluetooth-tethering-edison.html
Deweyoxberg
@Deweyoxberg
Oct 01 2017 15:30
I was thinking starting up the wifi.md page to keep information specific to networking in there
the tether page is more "how to setup"
PieterGit
@PieterGit
Oct 01 2017 15:31
@YYCMichael : i would first manually sync the clocks of your pump and cgm to a ntp based clock (e.g. your PC/Mac). There is also a script for syncing oref0-set-device-clocks, see https://github.com/openaps/oref0/blob/master/bin/oref0-set-device-clocks.sh
checking first is better, because you then know that there is a problem
John Benjamin
@Bender1061
Oct 01 2017 15:36
with people that have multiple rigs running in their house. How you do test a new Branch? I would carry my new PRed rig on me, but you might get another rig to weasel it's way in and then you have two (or more ) rigs doing different calculations. Yes I could unplug the other rigs, but does anyone know of an easy way to disable a rig so that it does not interfere with your test rig. I know this is a weird request, and I'm more of wondering what others do? By the way @scottleibrand openaps/oref0#677 sounds amazing, I'm going to test that out right now!
sdneufer
@sdneufer
Oct 01 2017 15:37
@YYCMichael Do you have a full link to the wifi.md page. I found a broken link to it in the troubleshooting page.
sdneufer
@sdneufer
Oct 01 2017 15:43
@Bender1061 My question as well.
Deweyoxberg
@Deweyoxberg
Oct 01 2017 15:43
I am making it.
wifi.md doesn't exist
sdneufer
@sdneufer
Oct 01 2017 15:44
I noticed, however there is a wifi debug page here: https://openaps.readthedocs.io/en/2017-05-21/docs/Resources/wifi.html
Deweyoxberg
@Deweyoxberg
Oct 01 2017 15:45
well then....
sdneufer
@sdneufer
Oct 01 2017 15:45
my thought exactly ;)
Deweyoxberg
@Deweyoxberg
Oct 01 2017 15:45
I'll update the link in my pull request and add the bluetooth stuff
sdneufer
@sdneufer
Oct 01 2017 15:47
cool... I can go walk the dog who is staring at me intently...
Deweyoxberg
@Deweyoxberg
Oct 01 2017 15:49
link fix submitted for review :)
oh, wifi.md does exist...
but it's in /docs/docs/docs/wifi.md
(why three docs)
what the...
But if I make a fork, it vanishes?
GitHub why you do this!
Deweyoxberg
@Deweyoxberg
Oct 01 2017 15:57
Okay maybe someone can explain how github documentation is set up
sdneufer
@sdneufer
Oct 01 2017 15:58
Redundancy??? Need those that understand the docs better @Kdisimone to help..
Deweyoxberg
@Deweyoxberg
Oct 01 2017 15:58
OpenAps / Docs / Docs / Docs (why)
OpenAps / Docs / Docs / Docs / Resources, with a few .md files like my-first-pr.md an troubleshooting.md
OpenAps / Docs / Docs / Docs / Troubleshooting, with several .md files like General_linux_troubleshooting.md
Very confused by the structure.
And I know I found wifi.md but good luck if I can find it again :D
Deweyoxberg
@Deweyoxberg
Oct 01 2017 16:05
Going to wait before trying to stuff this info somewhere
katie disimone
@Kdisimone
Oct 01 2017 16:20

@mike @sdneufer so the OpenAPS docs are held in the docs repo...but there’s a “service” called read the docs (RTD) that periodically rebuilds the displayed docs from those markdown files in the docs repo. Before making major doc changes, dana has RTD preserve a version of the docs in that state. So when you click on the green “v: latest” link in the docs, you’ll see links to not just the “latest” RTD build of the docs, you’ll also see those old preserved versions listed by dates.

So if you try to find or edit off the old md files for the preserved builds....things could get sticky because the files and folders and names have likely changed since that preserved version.

If you start a PR directly from the page you’d like to edit (using the “latest” version of RTD build), then click “edit in GitHub”, that button will take you straight to the md page you want to edit and automatically generate a new branch for you to do the Pr from. It’s very slick and easy that way

If you already have the fork of the docs repo, you can skip the first step in that page
Deweyoxberg
@Deweyoxberg
Oct 01 2017 16:22
I saw that, but noticed the troubleshooting page refered to wireless troubleshooting as wifi.md which doesn't exist.
Started to make one, but then found a wifi.md buried ... but I can't find it again... :D
It had good stuff on it
katie disimone
@Kdisimone
Oct 01 2017 16:23
@YYCMichael ahh, that’s probably an old link path from before the last major docs overhaul in July.
Deweyoxberg
@Deweyoxberg
Oct 01 2017 16:25
So how do we get the info from the ghost wifi.md file to a new wifi.md file
Also typing .md is annoying because I keep hitting / :P =laugh=
Scott Leibrand
@scottleibrand
Oct 01 2017 16:26
@jaylagorio a 100 temp target is not an eating soon as far as oref0 is concerned. You have to set something below 100 to get any of the behaviors triggered by a low temp target.
katie disimone
@Kdisimone
Oct 01 2017 16:27
@YYCMichael what’s the current page that you’re finding the dead link?
The current troubleshooting page you’re talking about?
jaylagorio
@jaylagorio
Oct 01 2017 16:28
@scottleibrand Ah, I thought any temp target labeled as Eating Soon would trigger SMB when that configuration value was set to true. Looks like I'll be trying that again with a 99 value at the next wedding.
Scott Leibrand
@scottleibrand
Oct 01 2017 16:28
The reservoir check is to make sure no other rig (or human) has bolused since the pumphistory pull. If the reservoir reading goes down, it retries the SMB checks to pull in the new bolus (and then doesn't SMB within 2m of another one).
jaylagorio
@jaylagorio
Oct 01 2017 16:28
Ah, got it. So misunderstanding of the log on my part.
Heading "wifi and hotspot issues"
I sent in a PR to fix the link
Scott Leibrand
@scottleibrand
Oct 01 2017 16:30
@Bender1061 when I want to test on just one or a few rigs, I do sudo service cron stop; sudo shutdown -r +60; killall oref0-pump-loop, adjusting the 60 for however many minutes I want them off. The scheduled reboot ensures that cron gets re-enabled at that time. If you want it left off indefinitely you can skip that part.
That’s the old page you’re talking about seeing and not finding, right?
Deweyoxberg
@Deweyoxberg
Oct 01 2017 16:32
yes :D
Scott Leibrand
@scottleibrand
Oct 01 2017 16:33
@jaylagorio unless you're going to be highly active, I would also recommend entering conservative carb estimates as meal announcements so it can get started on SMBs without having to wait for UAM to detect a sustained BG rise.
jaylagorio
@jaylagorio
Oct 01 2017 16:34
OK, will do!
John Benjamin
@Bender1061
Oct 01 2017 16:35
@scottleibrand Thank you, that is just the type of thing I was looking for!
katie disimone
@Kdisimone
Oct 01 2017 16:45
@YYCMichael ok...yup...that’s from the old docs. It has a separate branch in github. It’s not on master branch. https://github.com/openaps/docs/blob/old-docs/docs/docs/Resources/wifi.md
But rather than point back to that page, better to cherry pick if there’s parts of the old one you want to bring forward...just copy and paste from that old page maybe?
Dana Lewis
@danamlewis
Oct 01 2017 16:53
@jaylagorio btw - what is your normal target?
@YYCMichael rtd expects certain files in subdirectories vs top level directories for building. At some point we accidentally added an additional sub directory, but w so many people using docs and various links, haven't pushed on cleaning it up for concern of breaking more than helping. Happy to help you find things if you'd like more detail on rationale on what is where. +1 for bringing in any content from old versions that you think still needs to be in current docs.
PopcornPam
@PopcornPam_twitter
Oct 01 2017 17:11
Any ideas why I would see "smb disabled' in the pill? Not on a temp target, and 206 with 2 up arrows
jaylagorio
@jaylagorio
Oct 01 2017 17:12
@danamlewis At that time it would have been 120 - 130
I think I would have liked to have used a temp to 100 to say "during this time frame be more aggressive because I'm not confident I know what I'm eating" and I didn't express that correctlyu
Brian Rabinovitz
@bdr1177
Oct 01 2017 17:17
@scottleibrand When you say temp target of 100 is not ES for oref0, do you just mean for SMB and other advanced functions? Sometimes I use 100 as an ES target because my daughter's normal target is 130. I assumed that some people (especially adults) using 80 as an ES target were doing so because their normal targets we closer to 100.
jaylagorio
@jaylagorio
Oct 01 2017 17:18
^^
Deweyoxberg
@Deweyoxberg
Oct 01 2017 17:18
@danamlewis : bringing in the info was the idea :) Just was trying to figure out how things were setup
Dana Lewis
@danamlewis
Oct 01 2017 17:20
@jaylagorio that's why I was asking. A below your target low target should allow smbs. (@scottleibrand may be forgetting some people have higher baseline targets). But may be something else going on re settings for why that didn't trigger SMB's. (But still want to do some carb estimate if you want it to work faster ;)).
@bdr1177 correct thinking process
jaylagorio
@jaylagorio
Oct 01 2017 17:20
That sounds good, I'll keep that in mind for next time!
Dana Lewis
@danamlewis
Oct 01 2017 17:21
Ahh. Debating with @scottleibrand . It would have to be 99 to trigger SMB's
jaylagorio
@jaylagorio
Oct 01 2017 17:22
OK! Whatever the number I have to use is what I'll use
Scott Leibrand
@scottleibrand
Oct 01 2017 17:22
A 100 temp target vs. normal 120-130 would activate high temps but not SMBs.
Dana Lewis
@danamlewis
Oct 01 2017 17:22
Different parts of the code base. So it doesn't know your baseline target for SMB . (But per @bdr1177 , a 100 would work for temp basal increase)
jaylagorio
@jaylagorio
Oct 01 2017 17:22
For what it's worth, there was some crazy high temping going on
Dana Lewis
@danamlewis
Oct 01 2017 17:22
Heh
jaylagorio
@jaylagorio
Oct 01 2017 17:22
So at least that's consistent
Scott Leibrand
@scottleibrand
Oct 01 2017 17:23
Also, no reason you can't set an ES of 80 even with a 130 normal target, if you are actually expecting enough food to raise BG by 50 points.
jaylagorio
@jaylagorio
Oct 01 2017 17:23
Sure, true
I think I just used it incorrectly due to some misconceptions, this chat will be helpful for next time around
I have to say though that I was crazy impressed with the 45g popcorn I had the night before though
Brian Rabinovitz
@bdr1177
Oct 01 2017 17:25
Does it make sense to try SMB with normal basals between 0.25 and 0.35? I understand it would be disabled whenever basals are below 0.30. But would it be enabled whenever a temp above 0.30 is set?
PopcornPam
@PopcornPam_twitter
Oct 01 2017 17:26
As far as I understand, high temps (activity temp) disables smb. in this case, how does it work to enter carbs via nightscout at the same time as activity temp?
Brian Rabinovitz
@bdr1177
Oct 01 2017 17:26
Sorry, I meant temp basals not temp targets.
Scott Leibrand
@scottleibrand
Oct 01 2017 17:28
@popcornpam with a >100 temp target and carbs, oref0 will try to high temp for the carbs instead of SMB.
Tore Bjørndalen
@BjoTore_twitter
Oct 01 2017 17:28
Got this situation after dev update:
Oct 01 19:01:34 ToresEdison pump-loop.log: ed
Oct 01 19:01:38 ToresEdison pump-loop.log: Could not parse input data: [SyntaxError: /root/myopenaps/monitor/iob.json: Unexpected end of input]
Oct 01 19:01:38 ToresEdison pump-loop.log: Traceback (most recent call last):
Oct 01 19:01:38 ToresEdison pump-loop.log: File "/usr/local/bin/openaps-report", line 82, in <module>
Oct 01 19:01:38 ToresEdison pump-loop.log: app( )
Oct 01 19:01:38 ToresEdison pump-loop.log: File "/usr/local/lib/python2.7/dist-packages/openaps/cli/init.py", line 51, in call
Oct 01 19:01:38 ToresEdison pump-loop.log: self.run(self.args)
Oct 01 19:01:38 ToresEdison pump-loop.log: File "/usr/local/bin/openaps-report", line 75, in run
Oct 01 19:01:38 ToresEdison pump-loop.log: output = app(args, self)
Oct 01 19:01:38 ToresEdison pump-loop.log: File "/usr/local/lib/python2.7/dist-packages/openaps/cli/subcommand.py", line 52, in call
Oct 01 19:01:38 ToresEdison pump-loop.log: return self.method.main(args, app)
Oct 01 19:01:38 ToresEdison pump-loop.log: File "/usr/local/lib/python2.7/dist-packages/openaps/reports/invoke.py", line 48, in main
Oct 01 19:01:38 ToresEdison pump-loop.log: reporters.Reporter(report, device, task)(output)
Oct 01 19:01:38 ToresEdison pump-loop.log: File "/usr/local/lib/python2.7/dist-packages/openaps/reports/reporters/init.py", line 33, in call
Oct 01 19:01:38 ToresEdison pump-loop.log: self.blob = self.serialize(data)
Oct 01 19:01:38 ToresEdison pump-loop.log: File "/usr/local/lib/python2.7/dist-packages/openaps/reports/reporters/init.py", line 31, in serialize
Oct 01 19:01:38 ToresEdison pump-loop.log: return self.method.serialize(render(data), self)
Oct 01 19:01:38 ToresEdison pump-loop.log: File "/usr/local/lib/python2.7/dist-packages/openaps/vendors/process.py", line 56, in prerender_json
Oct 01 19:01:38 ToresEdison pump-loop.log: return json.loads(data)
Oct 01 19:01:38 ToresEdison pump-loop.log: File "/usr/lib/python2.7/json/init.py", line 338, in loads
Oct 01 19:01:38 ToresEdison pump-loop.log: return _default_decoder.decode(s)
Oct 01 19:01:38 ToresEdison pump-loop.log: File "/usr/lib/python2.7/json/decoder.py", line 366, in decode
Oct 01 19:01:38 ToresEdison pump-loop.log: obj, end = self.raw_decode(s, idx=_w(s, 0).end())
Oct 01 19:01:38 ToresEdison pump-loop.log: File "/usr/lib/python2.7/json/decoder.py", line 384, in raw_decode
Oct 01 19:01:38 ToresEdison pump-loop.log: raise ValueError("No JSON object could be decoded")
Oct 01 19:01:38 ToresEdison pump-loop.log: ValueError: No JSON object could be decoded
Oct 01 19:01:38 ToresEdison pump-loop.log: SMB pump-loop failed. pump_loop_completed more than 15m old; waiting for 40s silence before mmtuning
Oct 01 19:02:02 ToresEdison network.log: Starting oref0-online.
What to do? Try a new setup? It doesn't fix it self(4 hours now)
The file is empty, null bytes
Scott Leibrand
@scottleibrand
Oct 01 2017 17:30
@bdr1177 with basals that small, it will only SMB when they're high enough. you'll need to keep a close eye on them to make sure that it doesn't give too much insulin when rounding things to the nearest 0.1U
@BjoTore_twitter check your preferences.json for syntax errors first
Brian Rabinovitz
@bdr1177
Oct 01 2017 17:31
Okay. We'll hold off then, since my daughter is at school much of the time, which isn't really good for testing out new things.
Scott Leibrand
@scottleibrand
Oct 01 2017 17:31
if that parses as valid json (jq . ~/myopenaps/preferences.json) then you'll need to drill down and run the iob.json and profile.json reports manually to see what it's puking on that results in an empty iob.json
@bdr1177 when you're ready, you could start by enabling SMB only for low temp targets, and activate it manually for testing with a <= 99 target whenever you can watch it
Brian Rabinovitz
@bdr1177
Oct 01 2017 17:34
@scottleibrand So would that be using SMB in conjunction with ES (with target < 100)? In that case, would it SMB throughout the ES period, or only when it sees carbs?
Sorry if I'm misunderstanding. I read the SMB docs some time ago, but haven't focused on them since our basals were so low.
Tore Bjørndalen
@BjoTore_twitter
Oct 01 2017 17:35
Good tip as always from you two:. Found it: "curve": "utra-rapid",
Scott Leibrand
@scottleibrand
Oct 01 2017 17:38
@bdr1177 if you only set "enableSMB_with_temptarget": true and leave all the others to false, it will SMB as needed for the duration of the <100 temptarget (as long as basal is high enough), regardless of whether there are carbs. (but without carbs, any SMBs will be minimal, and with basals that low you may never get any)
Mike
@libxmike
Oct 01 2017 18:20
@scottleibrand I finally did setup one of my rigs to oref1/jubilinux 0.2.0 and looking for curve setup for Fiasp/ultra-rapid but cant find it, can you tell me where to set it? it is last missing link
danamlewis @danamlewis is broken record :) https://github.com/openaps/oref0/pull/650#issue-256564183
cameronrenwick
@cameronrenwick
Oct 01 2017 18:25
is it possible to determine the strength of the pump's comms (not mmtune from the EB/edison)... just wondering if there's a way to measure the com strength of a pump
Scott Leibrand
@scottleibrand
Oct 01 2017 19:00
Not sure what you mean. That's what mmtune is measuring.
cameronrenwick
@cameronrenwick
Oct 01 2017 19:02
ok thx Scott... I guess I was thinking mmtune was determining strength from the EB point of view.. so mmtune is giving the the relationship between the strength of the pump and EB then is what you're saying?
Tore Bjørndalen
@BjoTore_twitter
Oct 01 2017 19:04
@scottleibrand scott So if I want to generate profile.json , it was also empty, how to do that? Just curious since I tried to figure it out and couldn't. (did not want to wait more and ran setup again so all is good now)
Katie Aldridge
@kcrcgm
Oct 01 2017 19:08
Forgive me if this info is written up somewhere, but I have autotune questions. 1) When running nightly, how many days worth of data does it use to make its calculations? 2) If you ever want to turn it off, can you do that without rerunning the setup script? 3) I looked at settings/autotune.json to see what my basals are today (first time running as part of the script). It was a little hairy to read... Is there a way to see a simple table (like what you get when you run it manually)? 4) What improvements were made to autotune in the dev version? Thank you! I'm excited to be trying this out and want to understand it. 😊
Scott Leibrand
@scottleibrand
Oct 01 2017 19:17
@cameronrenwick it's reporting RSSI: received signal strength indication. So it's only tuning for maximum strength at the selected receive frequency, and then using that same frequency for transmission.
@ToreBj openaps report invoke settings/profile.json
@kcrcgm 1) it incrementally adjusts the previous day's results based on the last 24h, just like it does (multiple times) when running manually.
2) you can set your autosens min/max to 1 to disable both, but I generally don't recommend that.
3) there is an autotune_recommendations_report file in the autotune directory that is what you're used to seeing.
4) in dev, autotune now tunes CR directly, eliminates the upward bias on ISF and CR, and smooths out mealtime basals based on surrounding hours when there's no data for tuning a particular hour due to meals.
Katie Aldridge
@kcrcgm
Oct 01 2017 19:25
@scottleibrand That's awesome, thank you. I'm using my pump bolus wizard to enter carbs and meal boluses. Is that still ok?
Dana Lewis
@danamlewis
Oct 01 2017 19:26
Yes
cameronrenwick
@cameronrenwick
Oct 01 2017 19:27
thanks for the info @scottleibrand .. really helps my learning!
Katie Aldridge
@kcrcgm
Oct 01 2017 19:27
How often do you manually change the settings in your pump to match the autotune profile?
Dana Lewis
@danamlewis
Oct 01 2017 19:28
Me personally? Not often. I'm lazy :smile: but also don't usually run into the caps on Autotune. Really depends on how often your body is changing and needs different things, plus how reasonable your baseline was before
Niels Hartvig
@hartvig
Oct 01 2017 19:29
Has anyone tried to switch CGM on a rig? We just switched from MDT to G5 and after re-running the setup, openaps no longer reads or reports pump data (nothing in NS - but it read NS CGM data fine). By judging from mmtune logs it seems to find the pump alright. Here's how the logs look:
Oct 01 21:24:43 august pump-loop.log: Listening for 30s silence before mmtuning: .No pump comms detected from other rigs
Oct 01 21:25:24 august pump-loop.log: mmtune: "868.462", 5, -86 waiting for 52 second silence before continuing
Oct 01 21:26:20 august pump-loop.log: Radio ok. Listening: .No pump comms detected from other rigs
Oct 01 21:26:20 august pump-loop.log: Unsuccessful supermicrobolus pump-loop at Sun Oct 1 21:26:20 CEST 2017
Oct 01 21:27:02 august pump-loop.log: Starting supermicrobolus pump-loop at Sun Oct 1 21:27:02 CEST 2017 with 30 second wait_for_silence:
Oct 01 21:27:03 august pump-loop.log: Waiting up to 4 minutes for new BG: glucose.json newer than pump_loop_completed
Oct 01 21:27:37 august pump-loop.log: Radio ok. Listening: .No pump comms detected from other rigs
Oct 01 21:28:42 august pump-loop.log: Preflight OK. Old pumphistory-24h, waiting for 30 seconds of silence: Radio ok. Listening: .No pump comms detected from other rigs
Oct 01 21:28:54 august pump-loop.log: Old pumphistory-24h refreshValueError: month must be in 1..12
Oct 01 21:28:54 august pump-loop.log: SMB pump-loop failed. pump_loop_completed more than 15m old; waiting for 30s silence before mmtuning
Scott Leibrand
@scottleibrand
Oct 01 2017 19:30
Do you always get that same month value error on every loop?
Dana Lewis
@danamlewis
Oct 01 2017 19:30
Did you do share Bridge in NS?
Niels Hartvig
@hartvig
Oct 01 2017 19:31
@scottleibrand no, not every time - I've also gotten a "need more than 0 values to unpack" error. Yes, I've setup share bridge in NS
Scott Leibrand
@scottleibrand
Oct 01 2017 19:31
That's usually just an indicator of flaky pump comms
Katie Aldridge
@kcrcgm
Oct 01 2017 19:31
@danamlewis Ha! I'm switching back to humalog from Fiasp, so I may (or may not) hit the caps. I'll just keep an eye on it this week. 😊 Thanks!
Scott Leibrand
@scottleibrand
Oct 01 2017 19:32
See if you can get RSSI up to at least -70 and try a few more times.
Marco
@CaptainBalou
Oct 01 2017 19:34
@hartvig I also have pure signal strength and face these kinds of errors sometimes. Have we got a bad explorer board? When have you bought yours?
Niels Hartvig
@hartvig
Oct 01 2017 19:36
Mine is from spring this year, but has worked just fine until this weekend (where we swapped enlite CGM with Dexcom G5)
Dana Lewis
@danamlewis
Oct 01 2017 19:36
@kcrcgm yea if you are switching insulin types, you will likely want to adapt basals if needed sooner/keep a closer eye on it
Marco
@CaptainBalou
Oct 01 2017 19:37
@hartvig Mine was working fine as well but signal strength decaded more and more. But if that happens with this change and from one day to the other our issues are probabaly not the same.
Dana Lewis
@danamlewis
Oct 01 2017 19:37
@CaptainBalou occasional reading of history errors doesn't mean bad board. It means it didn't get a proper read for one of 1,800 or so possible reasons ;)
Marco
@CaptainBalou
Oct 01 2017 19:38
@danamlewis Yes understand. :-)
Can someone assist me what the correct output of this command should be?
openaps report invoke monitor/iob.json
I am investigating were thos „_date“-errors in the log could come from...
@danamlewis But nevertheless my EB seems to be nearly broken. :-( The distance is 5 centimeter and the singal strength is -71 so this is not what it was 3 months ago. :-(
Niels Hartvig
@hartvig
Oct 01 2017 20:05
Re-ran setup script without activating 0ref1 features and it's working now. Now to figure out how to get 0ref1 features again...
Dana Lewis
@danamlewis
Oct 01 2017 20:09
Added the note about autotune changes to the 0.6.0 PR. What else is missing from these notes? https://github.com/openaps/oref0/pull/650#issue-256564183
Marco
@CaptainBalou
Oct 01 2017 21:06

During investigation on strange pump log output I found out that I see a lot of errors like this on different report commands:

Profile less than 60m old. Temp duration >20m. Refresh  _date: '938f110111' }
ed pumphistoryWarning: clock input Invalid Date is unzoned; please pass clock-zoned.json instead
 and meal.json

So I tried to figure out were that comes from. I found this output on a lot of reports issued by pump-loop.

root@captainsloop1:~/myopenaps# openaps report invoke monitor/meal.json
meal://text/shell/monitor/meal.json
skipping bolus wizard entry with 0 g carbs and no insulin
Removed duplicate bolus/carb entries:6
No duration found for 0 U/hr basalundefined { _type: 'ChangeSensorSetup2',
  _description: 'ChangeSensorSetup2 2017-10-01T17:15:26 head[2], body[34] op[0x50]',
  timestamp: '2017-10-01T17:15:26+02:00',
  _body: '20011e00781e003c3c25f1e3b44600803721011e00781e003c3c25f1e3b446008037',
  _head: '5000',
  _date: '9a8f110111' } { _type: 'TempBasal',
  temp: 'absolute',
  _description: 'TempBasal 2017-10-01T17:15:19 head[2], body[1] op[0x33]',
  timestamp: '2017-10-01T17:15:19+02:00',
  _body: '00',
  _head: '3300',
  rate: 0,
  _date: '938f110111' } { _type: 'BasalProfileStart',
  _description: 'BasalProfileStart 2017-10-01T17:15:19 head[2], body[3] op[0x7b]',
  timestamp: '2017-10-01T17:15:19+02:00',
  _body: '223000',
  profile_index: 17,
  _head: '7b11',
  rate: 1.2,
  offset: 61200000,
  _date: '938f110111' }
No duration found for 0 U/hr basalundefined { _type: 'ChangeSensorSetup2',
  _description: 'ChangeSensorSetup2 2017-10-01T17:15:26 head[2], body[34] op[0x50]',
  timestamp: '2017-10-01T17:15:26+02:00',
  _body: '20011e00781e003c3c25f1e3b44600803721011e00781e003c3c25f1e3b446008037',
  _head: '5000',
  _date: '9a8f110111' } { _type: 'TempBasal',
  temp: 'absolute',
  _description: 'TempBasal 2017-10-01T17:15:19 head[2], body[1] op[0x33]',
  timestamp: '2017-10-01T17:15:19+02:00',
  _body: '00',
  _head: '3300',
  rate: 0,
  _date: '938f110111' } { _type: 'BasalProfileStart',
  _description: 'BasalProfileStart 2017-10-01T17:15:19 head[2], body[3] op[0x7b]',
  timestamp: '2017-10-01T17:15:19+02:00',
  _body: '223000',
  profile_index: 17,
  _head: '7b11',
  rate: 1.2,
  offset: 61200000,
  _date: '938f110111' }
reporting monitor/meal.json

I see that this is not the expected outcome of this command but I also cannot deal with it to see what goes wrong. Any help would be appreciated.

Scott Leibrand
@scottleibrand
Oct 01 2017 21:24
The pumphistory parsing code expects the TempBasal and TempBasalDuration records to be adjacent. In this case it looks like the BasalProfileStart is getting decoded at the exact same time as the TempBasal record, such that there is no adjacent TempBasalDuration record. Do you have any idea what's happening to make your pump log a BasalProfileStart record at such an odd time? Are you changing basal profiles manually on the pump?
Marco
@CaptainBalou
Oct 01 2017 22:00
No I am absolutely not knowing why that happens. I never changed the profile. Never! It is midnight here and I try to figure out more tomorrow. First I try to find out what I did at exactly that timestamp. I think this would help to reproduce things. Good night Scott!
PopcornPam
@PopcornPam_twitter
Oct 01 2017 23:57
Hey @scottleibrand, you mentioned that if you are running dev you should update it pretty regularly. If I run cd ~/src/oref0 && git checkout dev && git pull and it says it is up to date, am I good, or still go through the whole thing again?