is there a way for me to verify oref0 is functioning?
@gclyon maybe try
oref0-get-profile settings/settings.json settings/bg_targets_raw.json settings/insulin_sensitivities_raw.json settings/basal_profile.json preferences.json
i get a bash command not found error. do i need to put openaps first or...???
Looks like oref0 might not be installed right. Try the install step again?
is the only command needed this one?
openaps device add oref0 process oref0
openaps device add oref0 process oref0
I'm following this:
docs/docs/docs/walkthrough/phase-2/Using-oref0-tools.md
The top is confusing. Not sure if I'm supposed to be registering oref0 somehow?... Basically, I started working at the 'older directions' section
would try
sudo npm install -g git://github.com/openaps/oref0.git#dev
this installs a more recent version of oref0
the quick-packages.sh runs
sudo npm install -g json oref0
looks like you never ran that, or if you did it didn't succeed
re-try that and make sure you can then run
oref0-get-profile
I have a really dumb practical question. If shutting down the pi without properly powering it off can corrupt the sd card, how do you swap from power source to power source. Unless you only do at home after you power down using ssh through wifi....
most people get a battery that can do pass-through charging and then almost never unplug it
Ahh. Ok. I have one of those.
Getting ever closer to the loop... so close
ok, thanks everyone. i think i did not install it properly. that should fix my issues.
also, there's some miscommunication, the way that's framed...
usually the filesystem itself is fine, what gets corrupted are a couple of files in
.git/ subdir which don't get finished correctly
so, usually the git repo is corrupt and the filesystem/sdcard is fine
@bewest so that also happens when you run out of juice in your battery right?
right, if I yank the power (of any kind) while openaps/git is running, it tends to corrupt the git repo
if openaps/git is not running it tends to be fine
in any case, issuing something like
sudo shutdown now will cause it to stop everything nicely
how about the edison? I've seen the same thing happening as I use lipo batteries in the 3.7V input. Is there a way around that to avoid sshing into the rig to shutdown when you are on the go?
Also, is there a time delay needed after
sudo shutdown now or is it really immediate?
Re: the shutdown - one way around having to always SSH in to shutdown (or to leave it always running) would be to use the GPIO pins (on the pi, don't know how similar the Edison is) to trigger a script that shuts down (and halts openaps, or whatever you want first). Requires a little bit of wiring, though, and you'd have to figure out how to keep it from getting pressed accidentally. Not sure what sort of switch would be best for that. Example of the wiring is here, option two ('safe shutdown'): http://www.raspberry-pi-geek.com/Archive/2013/01/Adding-an-On-Off-switch-to-your-Raspberry-Pi
A switch like this could work (flip-up cover): http://www.ebay.com/itm/Pilot-Automotive-Safety-Cover-Push-Button-Switch-With-Blue-LED-PL-SW58B-/161464917010?_trksid=p2352135.l4275
If you install acpi on the Edison, pushing the button should do a clean shutdown
Was just looking at that - looks much nicer on the Edison than pi, with that power button on the breakout board. Also looks like you can customize what the power button does if you want: https://communities.intel.com/thread/61044 But for a basic shutdown probably not needed. Wish the RPi could do that!
@jasoncalabrese (from earlier today) I did the rm -rf .git; cd ..; openaps init your-loop-dir and it appears to have killed my loop directory. It doesnt show up as a directory anymore and it appears to be gone... I ran it from my loop directory and obviously replace your-loop-dir with my real directory.. is that file gone now?
or should i say.. are those files gone now
@jasoncalabrese weird.. i did a reboot and now the directory is full :)
@scottleibrand unfortunatelythe dev oref0 didnt detect any corruption either. still getting the same error :(
can you do:
grep diff `which oref0-reset-git`
returns 'no such file or directory'
same with just
which oref0-reset-git?
hmm.. hasnt returned anything yet... seems like its still trying though because it hasnt given me a $ yet
curious what exactly you're running. can you paste the output as a code block?
yes.. im actually ssh'ing thru my phone right now so i just need to switch connectoins
ok, don't mind my impatience then: just LMK if you get stuck again
@scottleibrand really appreciate it!! i have a feeling im doing something dumb, but its been hours of looking through this and ive got nothing.
did you paste my command or type it?
typed it
and if you typed, did you use ' or ` ?
ah.. i used '
Anyone know what the "Carbs activity / absorption rate" setting means in nightscout profile? The help text makes no sense to me.
The popup text for that form section says:
grams per unit time. Represents both the change in COB per unit of time, as well as the amount of carbs that should take effect over that time. Carb absorption / activity curves are less well understood than insulin activity, but can be approximated using an initial delay followed by a constant rate of absorption (g/hr).
And more importantly, does openaps use this value when calculating basal rates?
Openaps does not use that number for calculating basal rates.
In NS, that number is used to estimate how quickly carbs enter your blood after eating. So if you set it at 30g/hr, it is saying that when you enter a meal, it will take one hour for the last of those carbs to appear in your blood.
Wish I knew what was up with my BG. OpenAPS has been pretty much at the max basal rate for the last 48 hours, even after an insulin vial / pump site change
Ibuprofen seems to make my blood sugar go down/increases my ISF
What is max basal rate for your aps @cjo20 ?
Under 2u
@johners97 I could see ibuprofen doing that if the resistance was caused by some sort of inflammation response
best way to start over? as in remove openaps without reinstalling rapbsberry pi??
I'm a bit worried about setting up this loop...NHS Scotland technically own my insulin pump. They might think i'm abusing it and take it off me!
yeah, i'm thinking we should use sudo names on here! i checked into getting the dexcom, and was afraid rep might look to see if i was on here! ha!!
psudo
pseudo
ahh
pseudonym is a fake name
I don't think anything bad will happen as a result of medtronic or dexcom knowing you're trying openaps
Why would the reps care how you're using the pump/Dexcom? You're not modifying it or altering it in any way.
suggestions on uninstalling openaps?
Well, could the NHS take it off me for developing a system to stop me having as many hypoglycemic seizures? Probably. They took a pump off my friend for not checking their blood sugar for 2 days out of the 7. Imagine how they'd react to this
@gclyon You could just remove the openaps directory and start over.
^
johners97: I suspect they aren't going to just straight out say "You used openaps, we're taking your pump off you". I'd hope at least there would be a discussion about whether you should be using it or not
@scottleibrand i think it was a rookie mistake trying to use the corruption fix from dev.. I just created the the new one in my current project without updating everything. I was hoping that would have worked, once i brought over the entire dev branch and ran it, it fixed the corruption. (Now i just need to figure out what else is different in dev since my determine-basal process is now erroring out, but a JSON error is a lot less intimidating than that git diff corruption!)
@cjo20 I wouldn't put it past them
mariano42: post the error here :)
i have tried every way i know of...
$ rm myopenaps
rm: cannot remove ‘myopenaps’: Is a directoryrm -r myopenaps
@cjo20 .. and change my streak of stubbornly spending hours trying to troubleshoot bc I am convinced I can figure it out myself?? :) I have a feeling it's going to be an easy one, as soon as i saw my git repo was back in business I chose to end on a high note and call it a night last night.
is there a way to say yes to all for each line it's asking me if i want to remove?
Can you set max basal rate to 12 units?
rm -rf myopenaps
be careful with
-rf though, you can delete more than you expect if you aren't careful
too late! ha
so i deleted the myopenaps directory. then tried running the "recommended way to install" using package-mgr
curl -s https://raw.githubusercontent.com/openaps/docs/master/scripts/quick-packages.sh | bash -and i still get error about no openaps script...
......openaps installed
Traceback (most recent call last):
File "/usr/local/bin/openaps", line 4, in <module>
__import__('pkg_resources').run_script('openaps==0.1.5', 'openaps')
File "build/bdist.linux-armv7l/egg/pkg_resources/__init__.py", line 719, in run_script
File "build/bdist.linux-armv7l/egg/pkg_resources/__init__.py", line 1496, in run_script
pkg_resources.ResolutionError: No script named 'openaps'
i'm not sure i understand which install to use... what's diff between package manager and installing from source?
I installed from the source
What should net basal insulin be?
What do you mean?
cat monitor/iob.json
{
"netbasalinsulin": 0,
"activity": 0,
"basaliob": 0,
"hightempinsulin": 0,
"bolussnooze": 0,
"iob": 0
johners97: I think it's the amount of insulin you're getting at that time, I could be wrong though
if your basal rate is 0.5u/hr, and it's set a temp basal of 0.6u/hr, I think the netbasalinsulin is 0.1u/hr
[report "monitor/iob.json"]
profile = settings/profile.json
use = shell
clock = monitor/clock.json
reporter = JSON
json_default = True
pumphistory = monitor/pumphistory.json
device = calculate-iob
remainder =
oref0-profile = settings/profile.json
It isn't working as it should
Any idea why? :)
It doesn't seem to be determining Iob in monitor/iob.json
Despite having a temporary basal rate active
I used
openaps report invoke followed by cat for the respective output file and iob doesn't seem to be ringing true as it is 0 ! (not 1)
check the input files to that report
I wish medtronic sent a copy of what you submitted to their online "contact us" form
The input files are json yes?
johners97: yes
I have checked them
are they correct?
have you zoned your pumphistory?
zoned?
I guess that means 'no' :P
It's in the documentation
I'm adding oref0 devices
At calculate iob
read all of that page
hmm, you're using enlite or dexcom?
enlite
I still don't know what I've missed
:)
have you finished reading the page?
I haven't reached "determine basal" yet
:)
hmm, actually, it shouldn't need it at that point
I am at calculate iob
what does the command line say when you invoke the iob report
~/my_openaps $ openaps report invoke monitor/iob.json
calculate-iob://JSON/shell/monitor/iob.json
reporting monitor/iob.json
and what is in monitor/iob.json?
:~/my_openaps $ cat monitor/iob.json
{
"netbasalinsulin": 0,
"activity": 0,
"basaliob": 0,
"hightempinsulin": 0,
"bolussnooze": 0,
"iob": 0
when was the last time you did a bolus?
16:07
today
what does monitor/clock.json contain?
0.300 units
~/my_openaps $ cat monitor/clock.json
"2016-07-27T16:01:06"
when was the last time you updated pumphistory.json?
Was it, by any chance, 6 minutes before you did the bolus? :P
Assuming you did the bolus about 40 minutes ago
yassss
You're a legend
:P
forgot to invoke the clock
to be in line with now
Net basal insulin ain't working
though
~/my_openaps $ cat monitor/iob.json
{
"netbasalinsulin": 0,
"activity": 0.0022,
"basaliob": 0,
"hightempinsulin": 0,
"bolussnooze": 0.098,
"iob": 0.24
what's your net basal insulin rate at the moment?
135% or 1.400 units
(in reality )
you mean you've got a 135% temp basal set at the moment?
Yes
~/my_openaps $ cat monitor/iob.json
{
"netbasalinsulin": 0,
"activity": 0.0023,
"basaliob": 0,
"hightempinsulin": 0,
"bolussnooze": 0.082,
"iob": 0.232
when did you set that basal rate? Is it actually a percentage basal rate?
when did you last update pumphistory.json?
I set it 7 minutes ago
It is an actual % basal rate
Should I consider switching away from that to the numerical?
Percentage basal rates aren't supported by oref0
I thought that!!
Thats why
I just wasnt sure
I thought it would take the numbers in
hmm create a % conversion? like units conversion ?
It's easier to just work with raw units/hr
Right it works now :) Thanks
~/my_openaps $ cat monitor/iob.json
{
"netbasalinsulin": -0.05,
"activity": 0.0028,
"basaliob": -0.05,
"hightempinsulin": 0,
"bolussnooze": 0.
Is net basal insulin a proportion?
0.05 = 50% for example?
no
Hmmmmm
@johners97 that means you currently have net -.05u less insulin in your body than normal, so less than your normal basal rate activity
@cjo20 the netbasalinsulin is a cumulative total of how much +/- basal insulin you've gotten (over DIA I think)
@danamlewis if you were to rate aps out of 10 on promoting blood sugar control what would you rate it?:)
Curious - have any of the low-level hardware hackers here attempted to read data directly from a G4 without having to use the receiver?
that's what xdrip does, it reads the signal from the transmitter
I am wholly unfamiliar with xDrip. Oops.
@johners97 Not quite sure what you’re asking. It’s a great tool to help provide nudges to keep my BGs in line. But like anything else, it depends on how much you’re willing to put in, to get out of it.
The reason I was asking was I was wondering if we could make the amount of things we carry around even smaller. Building something like that into my rig directly so I could remove the receiver from the equation would be nice.
lots of name overloading with xdrip, the code that runs wixel is xbridge, except for the od version called xdrip like the android app and the ios xdrip g5 library
there's @oschumac's https://github.com/oschumac/python-uart-pi-xbridge
I haven't tried it yet, but really want to
Yeah, I was always a bit confused with what xdrip actually is. Need to go the Reddit route, and call one of them "truexdrip" or something...
@cjo20 Thanks for the info, that helps. Does the Carbs Activity/Absorption Rate affect the COB display of NS? If not, any idea how it's used?
after updating to the dev branch I am getting errors when running my enact/suggested.json report. Error : date field not found: cannot calculate avgdelta Error: date field not found : cannot calculate two delta Could not parse input data : [ReferenceError: twoavg is not defined] .... I agree that two avg is not defined, so that makes sense, but I am not sure what it even is. Trying to look through the dev branch in github for some ideas, but missing anything.
your data doesn't have a
date field, which some of my new code in dev expects. there is a workaround that involves using the json tool to add a date field to your glucose data, or you can go back to master for now. either way, we should also open an issue to make oref0 handle your date format. can you open one at https://github.com/openaps/oref0/issues/new and include a sample from your zoned glucose.json?
@bfaloona the COB in NS is currently using the the absorption rate , but I'd like it to get the COB from the device status like we do for IOB so it matches what the loop calculated
@scottleibrand ok thanks
@jasoncalabrese That makes sense, thank you!
I'm often confused looking at nightscout where individual 'pills' are getting their info. I'm not sure if this is normal for those new to looping, or indicates that I'm woefully outmatched by the complexity involved! :worried:
As an example, IOB pill reports three different values (IOB, Basal IOB, and Treatments IOB), but I'm not 100% sure how to validate them.
And specifically, I'm not sure what 'IOB' means exactly. Is it supposed to be basal iob? Sometimes it jumps up by a few units for no reason and then drops back down an hour later (?!?)
This message was deleted
This message was deleted
@scottleibrand create issue #180, first time i have done this.. let me know if that works
in the IOB case the Treatments IOB is what NS calculates based only on boluses
(The image I just posted is supposed to be a nightscout screenshot. It shows the wrong image in my gitter client. ?)
@mariano42 looks like that glucose.json is not zoned. do you have a -zoned version?
@jasoncalabrese That makes sense. And Basal IOB would be how 'ahead' or 'behind' your normal rate you are. Right?
the main IOB displayed when you openaps is the Net IOB
that included the bolus IOB and the effect of the temp basals
Ah! The two numbers combined?
almost
Did my screenshot come through? or does it show something else?
gets complicated, but there's a timing issue too
it was a putty screenshot
yeah, very strange. Not what I pasted in.
@jasoncalabrese Have you seen the issue where the Net IOB is reported WAY too high? Yesterday we saw basaliob 1.04, treatmentiob 0.09, and netiob 3.31
@bfaloona What does your basal profile look like?
it should match what's shown in the OpenAPS pill and your openaps logs
@cjo20 yesterday we had three basal rates: 0.7, 0.8, and 0.675. Did that answer your profile question?
I'm trying to reconfigure a device to pull glucose entries from Nightscout. I've always pulled from dexcom before. As I understand the directions I should be able to use
<head><title>Document Moved</title></head>
^
No JSON object could be decoded
openaps use ns shell get but I'm getting the following error message: json: error: input is not JSON: Unexpected '<' at line 1, column 1:<head><title>Document Moved</title></head>
^
No JSON object could be decoded
Ideas?
make sure you're using https and not http
http will redirect to https
ok i'll check that
@bfaloona I can't remember exactly which value it effects, but there is a potential issue where the amount of IOB is calculated assuming that the baseline basal rate was always the same as it is now. If you have a default basal of 1u/hr from midnight to 10am, and run a 2u/hr temp basal, then switch to a default basal of 0.5u/hr, it'll think it was giving you 1.5u/hr net basal, not 1u/hr
@cjo20 interesting.
Two nights in a row I've needed to clone from github again because of git corruption. My do-everything alias starts off with
oref0 fix-git-corruption and usually reports Git repo does not appear to be corrupt
Any suggestions for resolving this?
On a related note, it worries me that THOUSANDS of commits are being created. Should I just not worry about that?
My instinct is to reduce the number of commits by a factor of 10. What's the best practice here?
Also, should I be pushing commits occasionally, or in my loop, or not unless I'm making changes?
@jasoncalabrese Thank you so much. That was it. Easy fix
OMG I have so many questions. :smile: :fire: :smile:
How do people monitor the loop errors? I'm considering pushing logs to Loggly so I can investigate problems more efficiently. What do others do here?
I create a new repo when there's an unrecoverable corruption and push that to github instead of trying to reuse the same
I have a bunch of repos
@jasoncalabrese Do you push from inside your loop?
I have a separate cron to push every 5 mins
so if that fails it doesn't get in the way
great
@jasoncalabrese How do you monitor errors? Nightscout does not show them (feature request?), and looking at logs via ssh is very inefficient for me.
funny that you have a repo named Indy, that's what we named our OpenAPS rigs
I have lots of indy-* repos
:smile:
This message was deleted
This message was deleted
This message was deleted
sorry for the deleted messages. I was trying to paste an image into the chat. keeps posting the wrong one... ugh
finally... here is a pic of my pi zero rig with wifi
@mariano42 I was having a similar problem with the avgdelta and it was corrected by pulling data from my cgm with a different command. When pulling CGM directly from Dexcom G4, change your command to oref0_glucose from iter_glucose or iter_glucose_hours and you may find it fixes things
@jbwittmer thanks... ill give that try
@baughaw :+1: :)
@baughaw So what is all that I'm seeing? Looks like a zero with something on top which has a USB connector, with the TI stick on top of that? I'm trying to come up with a PiZero rig also but was just going to use a 2-port USB hub with Wifi to pull the CGM data from Nightscout and a TI-stick on the other port. What did you do?
Also what size battery do you anticipate needing or do you have any experience with battery draw? Are you updating every 5 minutes? Sorry for all the questions
yes that is a pi zero with a hubpixed board on top and then a TI stick on top of that. Here is a link to the wifi/usb hub board. http://referencedesigner.com/blog/hubpixed-raspberry-pi-zero-hub-without-cable-or-connector/3017/
@baughaw very cool. Thanks.
nice how that connects with the pins
@baughaw how did you get the photo to work (I had the same problem this morning)
@baughaw Congratulations. I'm moving to an Edison next, but I'm jealous because the pi seems so much simpler. Please post more photos of it in action. :smile:
@bfaloona i had to drag and drop the photo to get it to work
@jasoncalabrese The pins are such a cool idea. It works really well
@jasoncalabrese The pins are such a cool idea. It works really well
I had nothing but trouble getting the carelink to work on the pi zero, which I always assumed was related to the hub or the power supply.
Ok, drag and drop worked. This is the screenshot of Net IOB not matching up with basal IOB and treatment IOB. Anyone seen this. Also, sometimes the mouse over text contains the pictured details and sometimes it just shows the net iob.
here is a picture of it in a ghetto case I made. wish I was better at CAD so I could 3D print one.
looks legit to me ;)
:smiley_cat:
Blood sugars have been super low today. Averaging 3.7 mmol/L or 67 mg/dL ! I have rapidly changing insulin requirements during the day that often results in severe episodes.
Well, at least we can eat loads of cake in those circumstances;)?
:cake:
as long as the cake is not a lie
Iborprofen this morning has made my blood sugars flatline...today is cake day indeed
:fire:
wait, I thought ibuprofen was ok
it’s Tylenol (acetaminophen) that’s the bad one
the cake is always a lie
idek
the warning that dexcom gives is for acetaminophen
im not on my cgms at the moment, the same goes with medtronic. Many reported cases of increased incidence of hypoglycemia in type 1 diabetics while taking ibuprofen
interesting...
so solution is just never require pain relievers
:laughing:
think it has something to do with it influencing the electrophysiology of the body. The CGM sensors measure plasma blood glucose based on electrical resistance i believe so anything that effects the electricity in our body from chemicals, ions, drugs may effect the readings???
not that in the known about it tbh
I was looking to pick up some silicone covered wire from Adafruit. What gauge is recommended for the Edison breakout board? Was it 26AWG?
@Spazholio https://www.proto-pic.co.uk/ribbon-cable-10-wire-3ft.html That's what I'm using at the moment
Well, I have 9 spare crimped ribbon cables already, so I don’t need that right now. Also - do you have a 10-pin connector crimper? Nice.
I didn't use a crimper for the connector, I just squeezed the parts together
I've got a crimper for these: https://www.proto-pic.co.uk/female-crimp-pins-for-01-housings-100-pack.html
Ah, ok. Are the wires thicker than the ones that come with the CC debugger? Those wires are scary flimsy. =)
And I still need some individual cables for the Edison, but can’t recall the gauge that’s optimal.
I'm concerned my loop script is introducing timing issues. The loop runs perfectly for hours on end, but then will error out or will enact a temp basal based on 20 min old data.
Should I start over with a new loop script and do-everything alias, or try to reproduce the timing issues and tackle them one by one?
The 'old data' issue I mention APPEARS to be determine-basal succeeding on ~10 minute old data, and THEN enact succeeding ~10 minutes later. Thus, the temp basal was set based on 20 minute old BG levels.
Does that sound like a loop problem on my end, or some other dependency I have wrong? My instinct is that the system should prevent action based on data older than ~10 minutes.