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

9th
May 2016
David Young
@wgs4
May 09 2016 01:02
@bewest I’m still working on my alias problem from may 7th. I tried the command you recommended
oref0 alias-helper howdy 'echo howdy $1' | openaps import
it produces error
raise ValueError("No JSON object could be decoded")
ValueError: No JSON object could be decoded
after some lines of instruction scroll
Jarred Yaw
@jyaw
May 09 2016 01:09
Funny story. loaded up the Amazon cart the other day and didn't realize til this morning that I forgot to order... the Pi! Hahaha.
Ben West
@bewest
May 09 2016 01:12
hmm
curious what oref0 alias-helper howdy 'echo howdy $1' does
should print some json stuff
David Young
@wgs4
May 09 2016 01:14
that does nothing
Ben West
@bewest
May 09 2016 01:15
ah, well basically it sounds like you don't have any aliases yet
David Young
@wgs4
May 09 2016 01:15
nothing I can see on screen anyway
Brian Bonner
@itsn0tm3
May 09 2016 01:15
This message was deleted
Ben West
@bewest
May 09 2016 01:15
the oref0 tool can help with this, if you want to get the latest version: sudo npm install -g git://github.com/openaps/oref0.git'#dev'
David Young
@wgs4
May 09 2016 01:18
now I get the following line when I run the previoys oref0 command
{"type": "alias", "name": "howdy", "howdy": {"command": "! bash -c \"echo howdy $1\" --"}}
Ben West
@bewest
May 09 2016 01:23
:-)
add the | openaps alias back on to end again
oref0 alias-helper howdy 'echo howdy $1' | openaps import
David Young
@wgs4
May 09 2016 01:23
howdy ! bash -c "echo howdy $1" --
Ben West
@bewest
May 09 2016 01:28
:-)
openaps alias show should print non-error now
also you can try out your new alias:
openaps howdy wgs4
David Young
@wgs4
May 09 2016 01:28
things are working now
at least with howdy
Following the open doc still is fuzzy though.
it jumps into telling me to update the monitor-pump alias.. I have no clue what my monitor-pump alias should be doing, and I’m still thinking I dont have one.?
Ben West
@bewest
May 09 2016 01:30
not sure which doc you are referring to
David Young
@wgs4
May 09 2016 01:31
visualization and monitoring — OpenAPS 0.0.0 doc
Scott Leibrand
@scottleibrand
May 09 2016 01:31
Try skipping ahead. The viz section still depends on some stuff in later sections.
David Young
@wgs4
May 09 2016 01:31
doh!
Scott Leibrand
@scottleibrand
May 09 2016 01:31
Note what stuff needs moved for us: some PRs are needed to fix.
Viz got moved up without all the dependencies fixed.
David Young
@wgs4
May 09 2016 01:32
I will try to do a PR now. haven’t figured that out yet either.
Ben West
@bewest
May 09 2016 01:33
@wgs4 I actually recommend trying out this area http://openaps.readthedocs.io/en/latest/docs/openaps-guide/overview.html
it goes over a high level over view of how the openaps tools themselves work, outside of the concept of "a loop"
more like the "hello world" thing we just did
David Young
@wgs4
May 09 2016 01:34
will do
Ben West
@bewest
May 09 2016 01:34
will allow trying out alias and report and use and friends without worrying about getting it right for a loop
David Young
@wgs4
May 09 2016 01:37
so the alias feature is specific to OpenAPS? it is different than a basic linux alias? Or it is built on the linux alias system, specifically using OpenAPS commands?
Ben West
@bewest
May 09 2016 01:38
http://openaps.readthedocs.io/en/latest/docs/openaps-guide/core/alias.html it's section of the doc I just sent, basically openaps alias it a feature that allows bringing any linux commands into openaps
David Young
@wgs4
May 09 2016 01:39
oh, it was the a cofiguration of possiblities I did not think of. ;-) thanks.
Ben West
@bewest
May 09 2016 01:39
alias is last, because it builds on top of the other primitives
I usually like to intro people to device and use first
and feel free to check in here as you go
David Young
@wgs4
May 09 2016 01:43
I’ve been tryin not to bother you guys too much, but I have been afraid to skip ahead when I get stuck because I’m afraid I’d forget the steps I skipped and then cause a larger headache trying to troubleshoot down the road.
I will read ahead more, and review this overview you gave me
Ben West
@bewest
May 09 2016 01:45
no worries, it's fine to check in often
diabeticgonewild
@diabeticgonewild
May 09 2016 02:21
@bewest I got this
root@diabeticgonewild3:~/tutorial# openaps vendor add openapscontrib.timezones
No module named recurrent
openapscontrib.timezones doesn't seem to be an importable python module
If it is a python module, try using --path to influence
PYTHONPATH
With regards to openaps/docs#132
David Young
@wgs4
May 09 2016 02:22
how do you post that code in color and background like that in gitter?
diabeticgonewild
@diabeticgonewild
May 09 2016 02:22

"```"

"```"

David Young
@wgs4
May 09 2016 02:23
thx
diabeticgonewild
@diabeticgonewild
May 09 2016 02:23
Exactly like that without the quotes, with a line inbetween
YW
diabeticgonewild
@diabeticgonewild
May 09 2016 02:41
Oh, I missed copying something...
Sleepless in Seattle is on Oxygen Channel tonight....they just showed my great grandmother's house on TV! :D
Jason Calabrese
@jasoncalabrese
May 09 2016 02:44
I saw someone else ran into an issue with recurrent
diabeticgonewild
@diabeticgonewild
May 09 2016 02:44
Yeah, I think the error is somehow linked to this:
sudo pip install -U openaps openaps-contrib openapscontrib.mmhistorytools
Jason Calabrese
@jasoncalabrese
May 09 2016 02:45
Must be some packaging changes, it needs to be install manually
diabeticgonewild
@diabeticgonewild
May 09 2016 02:45
Didn't install openapscontrib.timezones anywhere
Jason Calabrese
@jasoncalabrese
May 09 2016 02:45
Maybe related to the change to use pip over easy install
diabeticgonewild
@diabeticgonewild
May 09 2016 02:45
I tried installing it manually with that command, to correct the error, but it didn't work....
I have no idea.
Jason Calabrese
@jasoncalabrese
May 09 2016 02:46
I'd try the easy install method
diabeticgonewild
@diabeticgonewild
May 09 2016 02:46
Cool
diabeticgonewild
@diabeticgonewild
May 09 2016 03:01
I'm using Yocto so I'm concerned that may be a factor, too...
Hristina Dimova
@xpucuto
May 09 2016 04:11
Good morning everyone. I have a question . Is there a way to increase my temp basal rate inside openaps ? On my pump it was set on 4U/H , i even changed it to 5U/H, but my max basal givem from openaps is 3.1 U/H.
I am comfortable with 4U/H as mornings i am very insulin resistant ( nothing new right ) .
My max_basal.json is set to 4 .
Dana Lewis
@danamlewis
May 09 2016 04:18
What are your normal basal rates? There's a safety cap of 3x Max daily and 4x current basal
Hristina Dimova
@xpucuto
May 09 2016 04:20
before openaps my higher basal rate was 1.2 during the dat but often i used temp with 4u/h when i have a fast spike .
Ben West
@bewest
May 09 2016 04:23
sudo pip install recurrent
I'll have to fix that
Dana Lewis
@danamlewis
May 09 2016 04:23
So if you're talking about meal spikes that need that much, you may need to be considering the insulin timing for meals - ie set a temp target and do "eating soon" before the meal. And also consider trying advanced meal assist, which will ramp up more quickly for meals
diabeticgonewild
@diabeticgonewild
May 09 2016 04:23
Nice!
Ben West
@bewest
May 09 2016 04:23
@diabeticgonewild simply doing sudo pip install recurrent should fix that error
diabeticgonewild
@diabeticgonewild
May 09 2016 04:23
Thanks, @bewest
Dana Lewis
@danamlewis
May 09 2016 04:23
A little different scenario for needing that much basal increase compared to a non-meal situation
Hristina Dimova
@xpucuto
May 09 2016 04:25
i read about this advanced meal assist but how do i enable it. As far as i understood you enter it as a treatment from teh care portal, and openaps starts increasing basal preparing for meal .
Dana Lewis
@danamlewis
May 09 2016 04:27
Two different things. Eating soon mode is prior to the meal. AMA does require knowing the carbs from a meal, either from care portal or from the pump. You have to be willing to enter carbs consistently if you enable it. (Re: how to, requires some configuration that's not in the main docs yet, we're working on getting that better written up.) in either case, you still do your meal bolus.
Hristina Dimova
@xpucuto
May 09 2016 04:32
thanks for the clarification . i guess for mornings i will have to enter carbs :(
Jason Calabrese
@jasoncalabrese
May 09 2016 04:33
Eating Soon doesn't need carbs, and has made a really big difference for us
Hristina Dimova
@xpucuto
May 09 2016 04:34
ok but how to enable it so i can see in my treatments list?
Jason Calabrese
@jasoncalabrese
May 09 2016 04:35
It needs the ama branch of oref0 and you need to pull the temp targets from NS and use them when creating the profile
The NS side is already in the dev branch
Mark Wilson
@mddub
May 09 2016 04:37
had my first experience with my SD card on the pi getting corrupted (been ~5 months)
Hristina Dimova
@xpucuto
May 09 2016 04:37
new things on the list. first i have to setup my rig to pull data from nightscout because at the moment i am uploading the cgm data with cable from the receiver trough pi.
Mark Wilson
@mddub
May 09 2016 04:38
is this generally a death sentence for the card, i.e. i shouldn’t bother trying to reformat?
Dana Lewis
@danamlewis
May 09 2016 04:38
@mddub you can re-image usually
Ben West
@bewest
May 09 2016 04:39
our usage of the term "corrupt" has been very imprecise
do you mean it won't boot now?
Mark Wilson
@mddub
May 09 2016 04:39
Dana Lewis
@danamlewis
May 09 2016 04:39
Yea. Depends on what's corrupted.
Ben West
@bewest
May 09 2016 04:39
or your openaps instance went crazy?
git status fails?
usually means the git instance itself is corrupted
Mark Wilson
@mddub
May 09 2016 04:39
prelude was a few days of errors indicating a corrupted python install
at least, errors consistent with that based on googling
Ben West
@bewest
May 09 2016 04:39
I see, this error lw to mooks ne
to me
what happens with git status?
Mark Wilson
@mddub
May 09 2016 04:40
i was able to fix by rebooting the first few times, but now it’s persistent
git status, log, etc. work fine
Ben West
@bewest
May 09 2016 04:40
ah I see
what happns with hmmm pydoc openaps
interesting, you have openaps 0.0.10
latest is 0.1.0-dev0
ah, ok
Mark Wilson
@mddub
May 09 2016 04:41
then i’ll probably end up on the latter when i reflash :)
Ben West
@bewest
May 09 2016 04:42
no need to reflash
so that error happens when there are outdated .pyc files
looks like solution is to try and delete any .pyc files
I would probably try installing/updating openaps
Mark Wilson
@mddub
May 09 2016 04:46
@bewest so i’ve rebooted maybe 3 times today and the error was still there, just rebooted a 4th to try your pydoc suggestion and it works again
(had the new SD card in already)
i wonder if that means the problem is worse than just stale *.pyc files
Aaron Michelson
@Aaybob
May 09 2016 06:08
Just jumping in----Has anybody seen calculate-iob to stop working. I'm guessing its just a random place in the code that electrically failed. I'm reformatting the memory.
Scott Leibrand
@scottleibrand
May 09 2016 06:11
You'll have to be more specific about identifying exactly what you saw if you want to figure out what it was.
Aaron Michelson
@Aaybob
May 09 2016 06:12
Thanks but I didn't save any error message. But does the RPi2 have issues with corrupted memory?
I'm sure every turn off of power has followed a shutdown
Maybe a jiggle of the plug end and it went off without my noticing
ddaniels1
@ddaniels1
May 09 2016 07:00
Need Cron help. I had a friend over who was helping me the other day and can't get a hold of him. He made a cron file to run the NS download every minute once I got that up and running. Now when I type '''crontab -e''' I can see the cron file but can't edit or get out. neither Q nor esc works.
Ben West
@bewest
May 09 2016 07:04
when you press esc a bunch of times,
are there any keyboard hints displayed on the edges of the screen, or does it just kind of blink?
or maybe do nothing even
ddaniels1
@ddaniels1
May 09 2016 07:07
only way out is to close the terminal and log back in to the RPI
Ben West
@bewest
May 09 2016 07:08
ok, I think it's vim
if you press esc a few times, and type :help
it should do new and interesting things
to get out of that,
:q
should be the file again
now we can write it out to disk and quit doing: :wq
ddaniels1
@ddaniels1
May 09 2016 07:14
Cool. you got me going, thanks
Ben West
@bewest
May 09 2016 07:21
vim can be... an acquired taste
ddaniels1
@ddaniels1
May 09 2016 07:26
wow, why would anyone use this? I'm sure there is a reason
is there not a simple text editor for cron>
?
Ben West
@bewest
May 09 2016 07:31
you can change it
you don't have to use it
it uses what ever is in EDITOR environment vriables
so you can try nano
EDITOR=nano crontab -e
if you like it you can make it permanent
a lot of people like joe also
you can make it permanent with http://raspberrypi.stackexchange.com/a/5368
ddaniels1
@ddaniels1
May 09 2016 07:34
Great thanks @bewest
Ben West
@bewest
May 09 2016 07:36
as for why people like vim, I actually find it very hard to use a computer without vim anymore because I love it so much... it's very good at rewarding long term use with short commands that demand very little of your joints/muscles
ie, you can do a lot of things without "leaving the home row" of the keyboard
ddaniels1
@ddaniels1
May 09 2016 07:37
Got it, knew there must be a good reason
Ben West
@bewest
May 09 2016 07:51
over a lifetime/career, the slight difference in burden can add up, I literally know of several emacs users who have had to take time off of work to let their joints and muscles heal, but I'm also very biased ;)
Pete Schwamb
@ps2
May 09 2016 11:35
With vi It's just your left pinky that will fall off, eventually, from try to reach ESC. ;)
Colin Lennon
@colinlennon
May 09 2016 11:44
@andrew-warrington @danamlewis @oskarpearson @xpucuto My flight is booked :smile: Looking forward to meeting you all on the 26th.
Oskar Pearson
@oskarpearson
May 09 2016 11:44
:) Great!
Hristina Dimova
@xpucuto
May 09 2016 11:48
:dancer: can't wait to meet you all
Toby Canning
@TC2013
May 09 2016 13:12
@scottleibrand I'm not sure why yours isn't charging fully and ours is. We average 15 hours of unplugged edison per day. We plug in every night around 10pm and have an automatic reboot at 11:59pm. I don't think we ever get that "blinky green charge light."
Dan Piessens
@dpiessens
May 09 2016 15:58
@bewest I've been doing some digging though groups about the G5 bluetooth protocol and noticed a lot of chatter about it late last year. Unfortunately I upgraded to my G5 before learning about OpenAPS, so now I'm bent on getting it working. Is there anyone trying to decode it, and/or did Dexcom ever come back with the protocol? If not I have an iOS dev friend that was going to help me decompile the app...
Lori
@LoriLori
May 09 2016 16:18
@dpiessens you could take a look at https://github.com/loudnate/xDripG5
Mark Wilson
@mddub
May 09 2016 16:48
@bewest after a few days of that python import error coming and going when i reboot my pi: https://gist.github.com/mddub/617504b084185d532ef4d8c6f996cc85
for the record, i’m cavalier about unplugging my pi without shutting it down safely (at least once a day for the past ~5 months), but until now, haven’t had issues beyond git (which can be recovered from automatically)
Andrew Warrington
@andrew-warrington
May 09 2016 17:59
@mddub I got a lot of segmentation faults for completely random commands (ls was one of them) when my /dev/root partition got full.
I fixed it by running apt-get autoremove and apt-get clean
plus deleting some old zipped logs from /var/log
that was on an Edison though
Mark Wilson
@mddub
May 09 2016 18:14
@andrew-warrington oh, good point, i hadn’t thought to check that
hmmm, looks like i’m fine on space, and the error manifests differently now: https://gist.github.com/mddub/279ece79c5ec97d41a743f1f7ecf6459
corrupted stdlib files? that’s consistent with the “reinstall python” advice i saw on some forum threads
but at that point i think flashing a new card is the easier path
Alex Kuzmanov
@alexkuzmanov
May 09 2016 20:04
I have got already produced upload-status.json but cannot upload it to the nightscout and there are no entries in mongo. Suppose to have problems with the timezones. "device": "openaps://raspberrypi", "created_at": "2016-05-09T19:42:47.620Z", "pump": { "battery": { "status": "normal", "voltage": 1.29 }, "status": { "status": "normal", "timestamp": "2016-05-09T19:42:14.000Z", "bolusing": false, "suspended": false }, "reservoir": 132.45, "clock": "2016-05-09T22:41:48+03:00"
Ben West
@bewest
May 09 2016 20:13
no problem there, it's all zoned
Alex Kuzmanov
@alexkuzmanov
May 09 2016 20:17
should I check the nightscout site for "correct" timezone?
Ben West
@bewest
May 09 2016 20:19
nothing wrong with timezones in this picture, you mentioned something about entries though?
is there an error when you try to upload to NS?
Alex Kuzmanov
@alexkuzmanov
May 09 2016 20:21
no errors. the IOB is visible but nothing more and there are no new entries in mongo db
Ben West
@bewest
May 09 2016 20:21
this doesn't include entries, that's a separate upload
Alex Kuzmanov
@alexkuzmanov
May 09 2016 20:22
and I don't have the "respond" file devicestatus.json
Ben West
@bewest
May 09 2016 20:23
need to get glucose, format for NS and then upload that to entries
if you're using dexcom, in openaps dev there's a new oref0_glucose use for dexcom that does the formatting nicely for NS
Alex Kuzmanov
@alexkuzmanov
May 09 2016 20:24
i'm using Enlite
so as I understand after having the upload-status.json I need to format it for NS before upload?
Ben West
@bewest
May 09 2016 20:35
hmm, no it sounds like upload-status itself is probably ok
the trick here is there are several different things to upload, separately
there is status, treatments, and entries, they are all separate
you've got status, which is a high level summary
there's also treatments, which is basically pumphistory transformed to look like NS treatments/careportal API
and also entries, which is the glucose values
they are different data sets, and have to be handled individually
I wrote the medtronic data parsing, but I actually don't know if I have a good example of entries/glucose from enlite to be uploaded to NS
if someone could get a bunch of example history glucose history from enlite or similar they'd like to upload to NS entries, and assign it in an issue to me or add it to here: https://github.com/openaps/docs/issues/129#issuecomment-216394218
it'd be very helpful, I can provide an equivalent oref0_glucose for medtronic glucose
David Young
@wgs4
May 09 2016 20:39
Question on Pull Requests.. I am trying to make my first PR on the dev docs. when I change the branch from master to dev, the button New Pull Request disappears and becomes a button that says View #135. So I am confused on what/where to click, as the instructions I’m following tell me I should change the branch from master to dev...
Ben West
@bewest
May 09 2016 20:58
from your fork, you mean?
for PRs from contribs, usually it's from your branch on your fork (on the right) to our dev branch (on the left)
I had to use the "compare across forks" feature
then selected your name and dev branch on the right
then selected our dev branch on the elft
left
David Young
@wgs4
May 09 2016 21:00
i see. I will have to pull request to fix those instructions as well. :-)
Ben West
@bewest
May 09 2016 21:00
:-D
blob
blob
blob
Dana Lewis
@danamlewis
May 09 2016 21:01
(kudos and thanks @wgs4 for working through that :))
David Young
@wgs4
May 09 2016 21:02
but you do want me to apply this to the DEV branch, not the master ?
Ben West
@bewest
May 09 2016 21:02
blob
David Young
@wgs4
May 09 2016 21:02
branch
Dana Lewis
@danamlewis
May 09 2016 21:02
correct
PR to DEV branch of docs
Ben West
@bewest
May 09 2016 21:02
blob
David Young
@wgs4
May 09 2016 21:02
ok
Dana Lewis
@danamlewis
May 09 2016 21:02
we congregate changes there (dev), then move them all over to master in chunks.
Ben West
@bewest
May 09 2016 21:03
yup, everything gets extruded via dev
think of it as dev is peer-review
David Young
@wgs4
May 09 2016 21:04
so it might make more sense for me to be following the dev version of the tutorials for my own setup then as well.?
Scott Leibrand
@scottleibrand
May 09 2016 21:05
if you see something confusing/wrong/unclear, doesn't hurt to check dev
at the moment there aren't a ton of big differences. sometimes we start pulling together not-quite-ready-for-primetime changes in dev, though, so I wouldn't want to have people always do that.
Dana Lewis
@danamlewis
May 09 2016 21:08
just merged @wgs4 first PR. :confetti_ball: :fireworks:
David Young
@wgs4
May 09 2016 21:14
:-)
Alex Kuzmanov
@alexkuzmanov
May 09 2016 21:16
@bewest I got the point. I already upload the status and treatments. About the BG - I will upload an example of files - suppose glucose.json/glucose_raw.json to the. Thanks a lot.
Lori
@LoriLori
May 09 2016 21:37
do you think this would make sense ?
"bg":108.3624008657518,"tick":"+0.43847485597342484"
Scott Leibrand
@scottleibrand
May 09 2016 21:46
those should be whole numbers
Lori
@LoriLori
May 09 2016 21:46
I am thinking about rather putting number as I get them not rounded
Scott Leibrand
@scottleibrand
May 09 2016 21:47
extra arbitrary precision is generally bad
and not sure how you'd get BG readings from the CGM other than on whole number intervals
Lori
@LoriLori
May 09 2016 21:48
cgm number is for example 158678 that translates to 108.3624008657518 using calibration
Scott Leibrand
@scottleibrand
May 09 2016 21:49
ah. if you want to make it 108.4 that'd be fine, but full arbitrary precision just makes it hard to read and doesn't add any value
the CGM is taking as calibration input meter values that are precise to +/- 20%, so you really only get two significant figures of precision
Lori
@LoriLori
May 09 2016 21:51
kind of delta is then non whole number and it can mach activity a bit more
Scott Leibrand
@scottleibrand
May 09 2016 21:51
yeah, two significant figures on the delta is usually a resolution of 0.1 or so
Lori
@LoriLori
May 09 2016 21:53
example would be 109, 109, 109, 108, 108, 108 where it would stop temp
Scott Leibrand
@scottleibrand
May 09 2016 21:57
only in really old versions of the algorithm
modern versions of determine-basal don't have those binary edge conditions any more
Lori
@LoriLori
May 09 2016 21:57
I am running old version, thanks
Scott Leibrand
@scottleibrand
May 09 2016 21:58
we're hoping to release dev to master soon. If you'd like to test that, I think you'll like it a lot better. :)
Lori
@LoriLori
May 09 2016 22:00
thanks for great work you are doing :)
jaylagorio
@jaylagorio
May 09 2016 22:47
@bewest On a Dexcom receiver, how do you correlate an EGV record (with BG) with a Sensor Data record (with filtered, unfiltered, RSSI)? Just comparing system time on each record?
jaylagorio
@jaylagorio
May 09 2016 23:25
That's what I was thinking, thanks!
David Young
@wgs4
May 09 2016 23:49
tried my first real PR to the Dev documentation. If it looks good, then I will spend more time with further documentation PR updates that have caused me confusion along the way.
Dana Lewis
@danamlewis
May 09 2016 23:50
looks good, I merged it :+1:
keep ‘em coming! :)
(you can do them piece by piece, or a huge swath of changes to multiple pages all at once, totally up to you)
David Young
@wgs4
May 09 2016 23:52
You bet.
something is wrong with the image I linked.
does the image have to be approved separate ?
hmm.. the image is in the correct directory on the repository. It must be my link? But my link worked properly in preview on my repository page. wonder if there is some other bug or linking information I need.