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

26th
Dec 2017
Jieseldeep
@Jieseldeep
Dec 26 2017 00:27

Hello again. I'm trying to setup BT tethering and am at the

sudo /usr/local/bin/bluetoothd --experimental &

step and get the result

sudo: unable to resolve host panc

Dave Acklam
@dcacklam
Dec 26 2017 00:39

@scottleibrand
Non-edison bug for you:
The un-checked execution of sudo ~/src/EdisonVoltage/voltage results in oref0-pump-loop results in the monitor/edison_voltage.json file ALWAYS existing even on raspi.

This (as far as I can tell so far) breaks nightscout status uploading (oref0-ns-loop determines whether to try and parse/push edison_voltage based on whether the file exists).

tepidjuice
@tepidjuice
Dec 26 2017 02:06
@danamlewis I've checked the times on my pump/phone and rig and they are all within one minute of each other
I keep getting prompted to update my profile by NS could this be the issue?
Dana Lewis
@danamlewis
Dec 26 2017 02:21
Check to make sure NS is also on the right time zone
Scott Leibrand
@scottleibrand
Dec 26 2017 02:55
@Jieseldeep that usually just means you haven’t rebooted yet since changing the hostname to panc. It doesn’t affect anything.
tepidjuice
@tepidjuice
Dec 26 2017 03:14
@danamlewis yes right time zone still getting no basal appearing.
and there are no future treatments.
Dave Acklam
@dcacklam
Dec 26 2017 03:28
Question: for 'openaps ns-loop' is there a 'verbose' or 'debug' option?
I've managed to get my 'pills' back by switching to the old oref0-ns-loop.sh & getting rid of the 'sudo ~/src/EdisonVoltage/voltage' line that creates an edison_battery.json even if you're not on an edison....
But I can't make 'openaps ns-loop' work on the pi or pi zero, and it fails silently with no info into the log
(the same myopenaps config worked on my old edison rig, the only difference is the hardware platform)
tepidjuice
@tepidjuice
Dec 26 2017 03:40
@dcacklam only thing I could think to do would be run each component of ns-loop to see which is failing?
tepidjuice
@tepidjuice
Dec 26 2017 03:47
is openaps/oref0#868 relevant?
Dave Acklam
@dcacklam
Dec 26 2017 05:09
That is what I had to do to get the .sh working.
@tepidjuice Or more specifically, I had to stop the creation of said empty file. The routines for Edison battery in oref0-loop are not checking for an Edison device OR the functional existence of a valid copy of the Edison voltage executable before they run, which results in a null file if not on Edison hardware or if the executable is absent.
Cameron Chunn
@applehat
Dec 26 2017 06:53
Is there an easy way to clear out some old NS entries without just blasting away all my records? I'm running into an upload error with xDrip, and it seems like if I delete a single entry in MongoDB it, it is allowing 1 more BG reading to upload - So I think I need to clear some out - but I don't want to blast it all away (autotune and such using it, as well as personal reporting)
Cameron Chunn
@applehat
Dec 26 2017 07:24
If anyone else is curious: The answer I ended up going with was manually connecting to the mongodb and deleting records myself. Unsure what error was actually getting thrown in NS (no logs in heroku?) but clearing out some old entries fixed it. Back to looping.
Matthias
@ceben80
Dec 26 2017 07:27
@dcacklam Q regarding your x12 setup, sorry asking again, the full loop is running w/o any issues (e.g. comm w/ the x12 and pump history was fully read?)
RamiLaakso
@RamiLaakso
Dec 26 2017 11:19
Hi!
Been running about oref0 three months without any trouble. Just an occasional future value needed to be deleted that's all. This morning NS does not show the status of OpenAPS, however. I don't see anything special in the logs. Any hints?
RamiLaakso
@RamiLaakso
Dec 26 2017 11:28
image.png
So the loop is working, just the NS is missing the status data.
RamiLaakso
@RamiLaakso
Dec 26 2017 13:24

Papertrail log:

Dec 26 15:15:09 localhost subg_rfspy: status: OK
Dec 26 15:15:09 localhost Version: subg_rfspy 0.8
Dec 26 15:15:53 localhost root: Uploaded nightscout/recent-missing-entries.json to https://dddddd.herokuapp.com
Dec 26 15:15:58 localhost root: Uploaded upload/ns-status.json to https://dddddd.herokuapp.com
Dec 26 15:16:38 localhost SubgRfspyLink: opening serial port
Dec 26 15:16:38 localhost subg_rfspy: status: OK
Dec 26 15:16:38 localhost Version: subg_rfspy 0.8
Dec 26 15:16:38 localhost Timed: out or other comms error - Received an error response Timeout - retrying: 1 of 3
Dec 26 15:16:39 localhost Timed: out or other comms error - Received an error response Timeout - retrying: 2 of 3
Dec 26 15:16:40 localhost Timed: out or other comms error - Received an error response Timeout - retrying: 3 of 3
Dec 26 15:16:42 localhost BEGIN: POWER CONTROL 562354
Dec 26 15:18:24 localhost root: Unable to upload to https://dddddd.herokuapp.com
Dec 26 15:18:28 localhost root: Uploaded upload/ns-status.json to https://dddddd.herokuapp.com
Dec 26 15:18:36 localhost SubgRfspyLink: opening serial port
Dec 26 15:18:37 localhost subg_rfspy: status: OK
Dec 26 15:18:37 localhost Version: subg_rfspy 0.8
Dec 26 15:19:32 localhost SubgRfspyLink: opening serial port
Dec 26 15:19:32 localhost subg_rfspy: status: OK
Dec 26 15:19:32 localhost Version: subg_rfspy 0.8
Dec 26 15:19:33 localhost Pump: 562354 is already responding. Not sending wakeup messages
Dec 26 15:19:35 localhost Timed: out or other comms error - Received an error response Timeout - retrying: 1 of 3
Dec 26 15:19:36 localhost Timed: out or other comms error - Received an error response Timeout - retrying: 2 of 3
Dec 26 15:19:54 localhost root: Uploaded nightscout/recent-missing-entries.json to https://dddddd.herokuapp.com
Dec 26 15:19:58 localhost root: Uploaded upload/ns-status.json to https://dddddd.herokuapp.com
Dec 26 15:20:11 localhost SubgRfspyLink: opening serial port
Dec 26 15:20:11 localhost subg_rfspy: status: OK
Dec 26 15:20:11 localhost Version: subg_rfspy 0.8
Dec 26 15:20:11 localhost Pump: 562354 is already responding. Not sending wakeup messages
Dec 26 15:20:12 localhost SubgRfspyLink: opening serial port
Dec 26 15:20:12 localhost subg_rfspy: status: OK
Dec 26 15:20:12 localhost Version: subg_rfspy 0.8
Dec 26 15:20:25 localhost SubgRfspyLink: opening serial port
Dec 26 15:20:25 localhost subg_rfspy: status: OK
Dec 26 15:20:25 localhost Version: subg_rfspy 0.8
Dec 26 15:21:08 localhost SubgRfspyLink: opening serial port
Dec 26 15:21:08 localhost subg_rfspy: status: OK
Dec 26 15:21:08 localhost Version: subg_rfspy 0.8
Dec 26 15:21:17 localhost SubgRfspyLink: opening serial port
Dec 26 15:21:17 localhost subg_rfspy: status: OK
Dec 26 15:21:17 localhost Version: subg_rfspy 0.8

tepidjuice
@tepidjuice
Dec 26 2017 14:42
@danamlewis turned out my mlab database was full. Just needed to clear that.
maybe we should put something in the docs about that issue?
alimhassam
@alimhassam
Dec 26 2017 14:54
i think @danamlewis was suggesting to donate your data before clearing anything.
i forget the details.
Jacob H
@jdhigh
Dec 26 2017 15:48
Could a hotspot autologin tool like this be used with Jubilinux? https://github.com/johnboiles/hotspot_autologin
alimhassam
@alimhassam
Dec 26 2017 17:07
i don't see why that wouldn't work.
Zakm0n
@Zakm0n
Dec 26 2017 17:40
OK, so I updated to the newest oref version, and it didn't work. Here is what the setup script gave me,

The following extra packages will be installed:
dns-root-data dnsmasq-base libmnl0 libnetfilter-conntrack3
Suggested packages:
resolvconf
The following NEW packages will be installed:
dns-root-data dnsmasq dnsmasq-base libmnl0 libnetfilter-conntrack3
0 upgraded, 5 newly installed, 0 to remove and 0 not upgraded.
Need to get 425 kB/501 kB of archives.
After this operation, 1,232 kB of additional disk space will be used.
Err http://http.debian.net/debian/ jessie/main dnsmasq-base i386 2.72-3+deb8u1
404 Not Found
Err http://security.debian.org/ jessie/updates/main dnsmasq-base i386 2.72-3+deb8u1
404 Not Found [IP: 128.61.240.73 80]
Err http://security.debian.org/ jessie/updates/main dnsmasq all 2.72-3+deb8u1
404 Not Found [IP: 128.61.240.73 80]
E: Failed to fetch http://security.debian.org/pool/updates/main/d/dnsmasq/dnsmasq-base_2.72-3+deb8u1_i386.deb 404 Not Found [IP: 128.61.240.73 80]

E: Failed to fetch http://security.debian.org/pool/updates/main/d/dnsmasq/dnsmasq_2.72-3+deb8u1_all.deb 404 Not Found [IP: 128.61.240.73 80]

E: Unable to fetch some archives, maybe run apt-get update or try with --fix-missing?
Couldn't install hostapd dnsmasq

Dana Lewis
@danamlewis
Dec 26 2017 17:46
Do you get the same thing if you run it a second time?
Zakm0n
@Zakm0n
Dec 26 2017 17:51
yup, this paste is actually from the second time
Zakm0n
@Zakm0n
Dec 26 2017 18:07
third time fails with the same info as well
Zakm0n
@Zakm0n
Dec 26 2017 18:18
following those links in my browser gives me a 404 as well, did debian remove those files or something?
Scott Leibrand
@scottleibrand
Dec 26 2017 18:57
Try another apt-get update?
Zakm0n
@Zakm0n
Dec 26 2017 19:06
Looks like that did the trick
Brian Cadre
@bcadre
Dec 26 2017 20:43
Question for everyone: I updated NS to enable use of the BW - What I cannot figure out is if the BW on NS actually administers insulin based on the entries we make in NS or is it simply a tool to track/guide boluses given via the BW on the pump?
Jacob H
@jdhigh
Dec 26 2017 20:43
It doesn't administer.
Brian Cadre
@bcadre
Dec 26 2017 20:44
Thanks so much @jdhigh.
Then, here's another question for those who use the BW in NS: Looking at NS, it appears as if she received 2 boluses; 1 from the pump (given) and one recorded on NS (not given). Also, the COB pill thinks she has 2x the amt of carbs on board, due to the "multiple entries." How do people use the BW on NS? Is it only for doing calculations but not for entering treatments? Any advice would be most helpful. Thanks in advance.
Here's the NS screenshot: Note the duplicated boluses at 1:17, 1:18.
Brian Cadre
@bcadre
Dec 26 2017 20:49
Screen Shot 2017-12-26 at 2.48.31 PM.png
Dana Lewis
@danamlewis
Dec 26 2017 20:51
You might want to delete the bw NS entry
So OpenAPS doesn't dose off those carbs if they weren't eaten
Brian Cadre
@bcadre
Dec 26 2017 20:55
@danamlewis thanks. deleted. Waiting for it to update. Much appreciated...
philipgo
@philipgo
Dec 26 2017 20:57
I think you need to reboot the rig to make OpenAPS forget those carbs
Brian Cadre
@bcadre
Dec 26 2017 21:00
Will do.
Brian Cadre
@bcadre
Dec 26 2017 21:07
@danamlewis Rebooted but carbs still there; it's ok tho since we actually gave some carb and didn't log it. We'll give a manual bolus to cover. Thanks again for your help!
Brian Cadre
@bcadre
Dec 26 2017 21:36
@drnoname82 if you use the BW, perhaps you can guide me on my previous question: How is it intended to be used? It seems as if one administers a bolus with the pump (automatically imported into NS) and then logs a “treatment” in NS using the BW, the carbs are doubled and it appears as if the person got 2x the insulin. What am I doing wrong/missing? Please help me understand. Thanks.
philipgo
@philipgo
Dec 26 2017 21:41
@bcadre I do not use the BW, sorry. We just enter the carbs via Nightscout and do a partial bolus and let OpenAPS handle the rest.
Dana Lewis
@danamlewis
Dec 26 2017 21:42
@bcadre BW was designed pre-looping for non-loopers to help them make better dosing decisions
it's not intended to be a looper tool; and I don't think anyone has tested it to see what info flows into (or not) the BW
but it will combine into the COB pill and potentially feed carbs into your rig, as you've discovered. so, I wouldn't use it without some careful testing (and small amounts) to see what it does/doesn't do
it does NOT necessarily mean you got the insulin "bolus" the BW recommended; but the carbs going into the rig may cause OpenAPS to decide to dose more insulin. So follow the netIOB only in the OpenAPS pill
Brian Cadre
@bcadre
Dec 26 2017 21:59
Got it. Thanks for the clarification. Again, much appreciated!
Brian Cadre
@bcadre
Dec 26 2017 22:10
@drnoname82 If I may ask you another question: we are trying to ameliorate the carb spike by doing a “manual advanced meal assist” as outlined in the docs. You say you enter the carbs in NS and then give a partial bolus and let OpenAPS handle the rest; are you using the BW on the pump to dose or are you bolusing manually? And, you say you’re giving a partial bolus; are you cutting back by an amount already on board (as in an advanced meal assist situation) or some other situation-dependent amount? Again, I’m learning and I thank you for your patience and guidance.
Dana Lewis
@danamlewis
Dec 26 2017 22:12
most people doing that have SMB enabled, btw. the docs may need updating if that's not clear
Brian Cadre
@bcadre
Dec 26 2017 22:19
We are not using SMB (yet) but the principle should still be similar as it pertains to an “advanced meal assist,” correct? I’m trying to gauge how much to cut back the actual bolus dose when she’s been on a temp basal for a temp target, again in an attempt to ameliorate the post-carb spike she gets with certain foods, etc.
Dana Lewis
@danamlewis
Dec 26 2017 22:20
AMA works no matter what
but I would do a full meal bolus if you're not on Fiasp and/or not using SMB yet
Instead, I would consider a) eating soon temp target prior to meals and b) if you need it later, another lower temp target if you want the rig to do "more", safely
Brian Cadre
@bcadre
Dec 26 2017 22:23
Sounds good. We’ll experiment. Again, thanks.
Dana Lewis
@danamlewis
Dec 26 2017 22:23
:+1:
Brian Cadre
@bcadre
Dec 26 2017 23:01
What is “MinGuardBG” in the OpenAPS pill in NS?
PieterGit
@PieterGit
Dec 26 2017 23:06
I managed to fix two Explorer v1.2 boards with broken micro usb connector for the UART port today with solder, glue, desolderen braid solder remover and a lot of patience. It even works :smile: . Did others ever had problems with broken micro usb connectors on an explorer board? The EU dexcom needs USB OTG and USB power and does not work with bluetooth. I will try to design a case which protects the USB cables better, because with daily use the connectors others seem to get stuck and break.
image.png
alimhassam
@alimhassam
Dec 26 2017 23:15
@bcadre minguardBG is the weighted average of the different oref0 prediction curve. If minGuardBG < threshold, then it'll disable SMB features, and might set a zero temp basal to avoid going too low.
Brian Cadre
@bcadre
Dec 26 2017 23:17
@alimhassam Thanks so much for the info; I was unable to find this in the docs.
Paul Dickens
@thebookins
Dec 26 2017 23:18
Hi All, looking for a bit of advice around what happens around CR transition times. We currently have a CR of 13 from midnight to 9:30 am, which changes to 16 for the rest of the day (despite trying a single CR, breakfast always seems to need more). ISF is 5.0 between 7 am and 9:30 am; 6.0 the rest of the day. We've noticed that at 9:30 the predictions all drop suddenly, as any COB left from breakfast is now being considered against a CR of 16 and the IOB is judged to be too much. The loop low-temps as a result and we can get stuck high. Would it be possible (or even make sense) to stick with the CR at the time the carbs were given until all decayed? If not, any suggestions about how to handle this? We could of course extend the breakfast CR out a few more hours, but this would be too strong for morning tea.
PieterGit
@PieterGit
Dec 26 2017 23:21
@bcadre @alimhassam : it's not a weighted average anymore, now is based on the COB BG and UAM BG based on fractionCarbsLeft. Simply said it's the mimimal expected bg value. As @alimhassam indicates: it will be used to set a zero temp if you will otherwise go low. See https://github.com/openaps/oref0/blob/174ef43c7f5dfb7a9c3f578bf38a939d4bccb034/lib/determine-basal/determine-basal.js#L697-L709
alimhassam
@alimhassam
Dec 26 2017 23:22
@bcadre to get into the nitty gritty details of the calculation, if you're comfortable with it, is to read the code for determine-basal which starts here. https://github.com/openaps/oref0/blob/dev/lib/determine-basal/determine-basal.js#L51
ah ^^^ yes better answer from @PieterGit
i guess, that still counts as a weighted average.
PieterGit
@PieterGit
Dec 26 2017 23:24
weighted average based on the fractionsCarbsLeft yes
alimhassam
@alimhassam
Dec 26 2017 23:24
:+1:
PieterGit
@PieterGit
Dec 26 2017 23:24
but not weighted on IOB BG for example
@thebookins : wouldn't splitting up the CR and ISF in smaller bits help to avoid this problem or make it smaller?
e.g. CR 13 to 9:30am, CR 14 to 10am, CR 15 to 11am, CR 16 to 12am for example
i assume you are aware of the the autotune limitations of using multiple CR and ISF with autotune? (it does not work, only the first segment is tuned)
alimhassam
@alimhassam
Dec 26 2017 23:28
@thebookins IMO, it would make a lot of sense to use the CR at the time of the carb entry rather than the time of calculation.
(But i don't think mutiple CR is something well supported right now). There are also other things which might make the COB calculation jump between ticks that i noticed, but it's not easy to change.