ok, noone can set a basal of .509, seems like autotune should round to the pump's scroll rate (0.025, 0.05, 0.1, ...)
would that affect the other calculations for carb ratio, ISF?
different pumps have different scroll rates, so not sure we have an easy way to do that
profile.json :)
I think people can round up themsevles
@scottleibrand @danamlewis either of you know if canadian pump is on 868 or 915 mhz?
but does it use the basal rate to calculate the other factors?
my rig seems to have stopped pulling glucose values from the pump (Enlite)... any idea what to look for first?
(I'm out now so I'll have to do actual debugging when I get home)
@j4cbo when did you last calibrate
calibrate as in fingerstick? around 10am, and openaps stopped looping around 2
ok. sometimes medtronic is funky about calibrations being what causes it to not get glucose, so that’s usually my first guess
hmm
the pump is still happily showing glucose values though...?
yea, if it’s only been 4 hours that’s probably not it
i just tried autotune on a ops server (Debian), did all steps from docs/autotune.md and getting lots of:
/usr/local/bin/oref0-autotune-recommends-report: line 61: printf: 28.773: invalid number
@ahanuban does the pump model have NA or WW in it?
@tynbendad @RosyRed sorry, I should have asked before tagging you on Fbk, let me know if you’d like me to untag you (from CGMITC post)
Fine by me. If I can do it anyone can!
:+1: you should say that ;) on Fbk
andlater stuff like:
~/src/oref0/bin/oref0-autotune-core.js autotune.1.2017-01-03.json profile.json profile.pump.json > newprofile.1.2017-01-03.json
/root/src/oref0/lib/autotune/index.js:77
for (var i=0; i < basalGlucose.length; ++i) {
^
TypeError: Cannot read property 'length' of undefined
at tuneAllTheThings (/root/src/oref0/lib/autotune/index.js:77:39)
at Object.<anonymous> (/root/src/oref0/bin/oref0-autotune-core.js:59:27)
at Module._compile (module.js:409:26)
at Object.Module._extensions..js (module.js:416:10)
at Module.load (module.js:343:32)
at Function.Module._load (module.js:300:12)
at Function.Module.runMain (module.js:441:10)
at startup (node.js:139:18)
at node.js:974:3
~/src/oref0/bin/oref0-autotune-prep.js ns-treatments.json profile.json ns-entries.2017-01-04.json > autotune.1.2017-01-04.json
Could not parse input data: [SyntaxError: Unexpected end of input]
@RosyRed ps - you today did ~2/3 of what it takes to create openaps. so if you got the hardware, you totally could have an OpenAPS rig, too, if you wanted to have one.
Dexcom is fighting with my insurance company or something. I was hoping to get the G5 this week :(
@scottleibrand where would I find NA or WW, it doesn't have either of those on the back
what does it have for model number?
MMT-522NAL is one of mine
@scottleibrand MMT-754LCMS
and @ceben80 what was that about G5-to-Edison direct BLE?
@danamlewis Good to know. I'm pretty happy with Loop but honestly don't know the differences well enough to know which one is best for us.
@ahanuban dunno then
@RosyRed a lot of people build both and then experiment with them to see what works best when. some people even end up using them both at different times of day.
openaps/docs#567
I have loved this challenge so much I might just build both. The nice people who help me is what has made all the difference. Awesome community effort here.
@scottleibrand I got wlan0 up and running! no idea what was wrong but I restarted from scratch and it's perfect.
thanks for the hlep
help*
great!
@RosyRed :) :confetti_ball:
someone got their explorer board today and got up and running a full OpenAPS setup in 1h50min
Wow!
that's quite fast :D
I have 30 mins and have to run. I'd like to add the extra basals. Should I wait a bit longer - no rush, I can come back tomorrow too
it’s about on average - people with existing rigs have done it in 45 mintues, people with another type of loop are doing it 1h30-2h, some people are taking longer, but once you get through flashing it’s not like it takes days/weeks anymore :)
@RosyRed you can do it now! we’re here for troubleshooting, so..
Scott said he would write up an example.
I just finished and Dana just merged it
yea. refresh ;) should be there
the autotune results seem a bit funky the longer period i use. so, i ran a report for 1 day and then a report for same day+1, and then a report for day thru day+1, the latter seems way off of the former 2 reports...
probably want to do step 3 over with the new example
2017-01-14
Parameter | Current | Autotune
ISF | 43.000 | 49.669
CSF | n/a | 3.018
Carb Ratio | 14.000 | 16.458
Basal Profile | |
00:00 | 0.500 | 0.489
00:30 | |
2017-01-15
Parameter | Current | Autotune
ISF | 43.000 | 51.923
CSF | n/a | 3.030
Carb Ratio | 14.000 | 17.136
Basal Profile | |
00:00 | 0.500 | 0.479
00:30 | |
2017-01-14 thru 2017-01-15
Parameter | Current | Autotune
ISF | 43.000 | 57.929
CSF | n/a | 2.979
Carb Ratio | 14.000 | 19.446
Basal Profile | |
00:00 | 0.500 | 0.469
00:30 |
@tynbendad the more data, the more it will change things. autotune is only "allowed" to make 20% of the changes the data would suggest on any given day
they're all consistently saying your ISF and carb ratio need to be higher. I would try making a small adjustment in that direction and run with it for a week or so, and then re-run autotune on that week's worth of data
so just change 1, 2 , 3 , etc
change all of the <your_whatever> to your values, yes
k
and change the start times to match what you have in your pump
HELLLO and thank you! I’m the 1 hour and 50 min setup girl. Can’t believe it. Board got delivered two days early. I’d even planned to take off work on Monday to sit down and set this all up. Board got here today and wow…that was fast. INCREDIBLE bit of work and I can’t wait to dig into it. The fact that I did it without needing any troubleshooting steps is huge kudos to those instruction writers :clap:
that's counter to what i see though, since today we needed to give extra bolus's for some meals
@Kdisimone :raised_hands: :confetti_ball: :clap:
@RosyRed if you need more than 4 schedules, just copy and paste and keep incrementing the "i" numbers
Trying to upload data from my G5 Clarity site... It is not on Studio... what is the suggested approach?
@RosyRed you'll also need to adjust the "minutes" value if you change "start" - it's the number of minutes since midnight, so just multiply the hour by 60
but at least its saying basal is in line, i was wondering how i was going to figure that out now that we're temp basal'ing all the time
@tynbendad what do you usually see after meal absorption is complete? do you come down hard and require zero-temping on the way down?
@RedCyclist1 upload data from where to what?
@tynbendad if it's recommending you increase ISF, that means that you're falling faster than expected when you have lots of insulin activity and carb absorption is complete
that's the graph for the same 2 days i pasted autotune on
looks like you need to be doing Eating Soon
Trying to establish baseline instructions say"Before getting started, we ask that you store at least 30 days of CGM data. For now, the easiest way to do that is to upload your Dexcom receiver to Dexcom Studio"
@tynbendad it's high-temping the whole way up, and then zero-temping the whole way down
but G5 uploads to Clarity website, not Studio
@tynbendad that looks like an insulin timing mismatch more than anything
leave the brackets on the time or remove?
@RedCyclist1 the main thing there is "get a copy of your data"
@RedCyclist1 ah, that needs updating. clarity is the new stuido
feel free to PR that :)
My pi3 stopped working so I got on putty and tail the logs, getting an error that says"failed to run pack-refs" checked space on sd card, have 42mb free of 62mb it says but its a 32gig stck, any tips?
@RosyRed remove all the <angle brackets>
@scottleibrand he won't prebolus now that he is using pump and loop :(
@scottleibrand Do I do anything with "mintues"
@tynbendad do you have an eating soon mode button in loop (modified workout mode or whatever)?
but isf & carb ratio shouldn't need to both go higher (weaker)?
@RosyRed change minutes to match start (hours * 60)
@scottleibrand I mean minutes that is listed under rate
@tynbendad we'd have to do the math on how much total net insulin is being delivered (by bolus and by loop temp basals) after each meal
Is there any specific format to store data, excel, word?
@RedCyclist1 have you been using NS?
if your data is already in NS, that is actually the ideal
@RosyRed the "minutes" under rate actually go with the "start" above it
I'll update the json to put minutes above rate so that's clearer. the computer doesn't care what order they're in. :)
@scottleibrand we do have workout mode set to a lower target for eating soon, but he doesn't use it regularly. however, as of a few days ago i added regular lower targets around his mealtimes since they are consistent.
k, that should help some
@scottleibrand so if it changes at 5:30 a.m. it would be "i": 1,
"start": "05:30:00",
"rate": <1.0>,
"minutes": 480
"start": "05:30:00",
"rate": <1.0>,
"minutes": 480
"rate": 1.0,
I'll remove the angle brackets too
@scottleibrand I have been using NS for 3 weeks, so that should be good enough?
@RedCyclist1 yep
i guess i'll keep an eye on autotune results, espec. for basal, not going to make any adjustments just yet
:+1:
wondering if my vm will still be usable after i close this session and then open a new one
should be. you can try opening a new session before closing that one
@RosyRed I just updated the docs again to be clearer there
should i stop the vm instance to save quota?
is the error"failed to run pack-refs" an sd card storage space problem?
@tynbendad your call. not sure what the quotas are, but it wouldn't hurt to shut it down. You can also do
sudo halt from the VM command line to shut it down
@ejoe132 not sure, but you can do
df -h to see how much disk free space you have
@scottleibrand The minutes thing is not sinking in. If the basal changes at 5:30 a.m. What do I put under minutes.
@RosyRed how many minutes are there between midnight and 5:30 am?
5.5*60
:+1:
ah! Ok, got it. I was making that too hard.
:-)
signing off - thanks for autotune and all the help @scottleibrand and @danamlewis !!!
@tynbendad thanks for helping us test it out! :)
Gotta run. Thanks for helping so much. I'll be analyzing number tomorrow for sure!
@RosyRed thanks for the help improving the docs & testing it out!
really needed it at this point, closed loop harder to keep #'s in check. good timing!
@scottleibrand I ended up with the same error again and wlan0 is broken.
My attempt to update openaps failed because of lack of disk space. however
df -h showsroot@juba2:~# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/root 1.4G 754M 563M 58% /
devtmpfs 482M 0 482M 0% /dev
tmpfs 482M 0 482M 0% /dev/shm
tmpfs 482M 55M 428M 12% /run
tmpfs 5.0M 0 5.0M 0% /run/lock
tmpfs 482M 0 482M 0% /sys/fs/cgroup
tmpfs 482M 482M 0 100% /tmp
/dev/mmcblk0p7 32M 3.8M 29M 12% /boot
/dev/mmcblk0p10 1.3G 782M 510M 61% /home
tmpfs 97M 0 97M 0% /run/user/0
@tynbendad we don't notice that with oref0 - you might want to consider building yourself an OpenAPS rig and comparing that to Loop - the setup is not much more work than we just did to run autotune
@dougtakeuti maybe
df -hi to check inodes?
@ahanuban ugh. does it still show the 0 mac address after a full powercycle?
What are inodes?
root@juba2:~# df -hi
Filesystem Inodes IUsed IFree IUse% Mounted on
/dev/root 384K 19K 366K 5% /
devtmpfs 121K 409 121K 1% /dev
tmpfs 121K 1 121K 1% /dev/shm
tmpfs 121K 547 120K 1% /run
tmpfs 121K 3 121K 1% /run/lock
tmpfs 121K 10 121K 1% /sys/fs/cgroup
tmpfs 121K 86K 35K 72% /tmp
/dev/mmcblk0p7 0 0 0 - /boot
/dev/mmcblk0p10 86K 40K 46K 47% /home
tmpfs 121K 4 121K 1% /run/user/0
@scottleibrand yes, oddly enough it does. the edison crashed while I was editing /etc/logrotate.conf and when I powered it up again wlan0 was broken.
@ahanuban sounds like a hardware thing. is the Edison firmly seated and screwed down?
@dougtakeuti that looks good. https://www.google.com/search?q=what+are+inodes
firmly seated, yes: screwed down, no, because my screws are currently stuck with USPS :worried: but it's definitely secure
@dougtakeuti maybe it's your /tmp being full that is causing your issue
I think the Edison cleans out /tmp on every reboot, if you want to try that
ok , I'll try that, thanks
@ahanuban hmm, not sure then. if you have a second Edison, I would see if it has the same issue or not...
@scottleibrand weird thing, I ran sudo ifconfig wlan0 up and now it's working fine
well, not working fine, but it has a proper HWaddr
ifup wlan0 still says
wpa_supplicant: /sbin/wpa_supplicant daemon failed to start
run-parts: /etc/network/if-pre-up.d/wpasupplicant exited with return code 1
Failed to bring up wlan0.
on boot I get
Failed to start Hostname Service.
Failed to start Hostname Service is normal
dunno why it happens, but it doesn't hurt anything
@scottleibrand so I figured out it's another line that I added to my wpa_supplicant.conf file.
ah, what's that?
I tried to add a second network
This message was deleted
so what was the syntax error?
when I do this, each network should have its own entry of network={
ssid="ssid"
psk="pw"
}
correct?
yeah, like:
network={
key_mgmt=NONE
priority=-999
}
network={
ssid="ARIN MEETING"
key_mgmt=NONE
}
yep, that's exactly what I did
it didn't like my second network, which is the wi-fi hotspot on my mobile phone
is it named something like
ahanuban’s iPhone?
with a curly
’ ?
the code was
```
network={
ssid="FBI_Surveillance_HQ"
psk="PWHERE"
}
network={
ssid="FBI Surveillance Van"
psk="PWHERE"
}
```
network={
ssid="FBI_Surveillance_HQ"
psk="PWHERE"
}
network={
ssid="FBI Surveillance Van"
psk="PWHERE"
}
oops
ah, you’re one of those people ;)
yes lol but I don't know why it disliked that
when I remove that, it works
when I add it, it breaks
any idea?
curly quotes are evil. :)
but I didn't put any curly quotes....
oh, sorry, misread you
no worries... at first I though maybe it's the spaces but it's all in proper quotes
can you paste working vs. not-working?
working:
ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=netdev
network={
ssid="FBI_Surveillance_HQ"
psk="PWHERE"
}
@scottleibrand Rebooting worked! Thank you!
@dougtakeuti woot!
broken:
ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=netdev
network={
ssid="FBI_Surveillance_HQ"
psk="PWHERE"
}
network={
ssid="FBI Surveillance Van"
psk="PWHERE"
}
@danamlewis Trying to complete first update.. stuck at the comparing changes... In the compare changes, I get base master and compare master, not dev like in the instructions...
yes, target master. (clearly, the first-pr also needs a PR :))
but it should be comparing your branch master, to openaps master
@scottleibrand it only breaks after a reboot
which is why I didn't catch it last time
@ahanuban weird. this is not really an openaps thing, so you might get better help on a more general forum? not sure. maybe start with something like https://www.google.com/search?q=wpa+supplicant+syntax+error ?
I'm out of ideas
alright. for now I'll do it with only one network and at least I've identified the issue
thanks
good luck
@danamlewis so I see the section prior to my update and then the section below shows with my update...sounds like I should just click on "Create Pull Request"?
yup
congrats on your first PR! :)
@danamlewis Had to make some contribution however small.... LOL ;)
@RedCyclist1 that’s exactly the point :) small contributions from everyone add up! :confetti_ball: :)
and they start getting addictive: keep this up and before long you'll be making a lot more bigger ones. :)
:+1: once you pop you can’t stop…? :)
@danamlewis OK, sorry to be soo literal, but the Edison board comes in 2 pieces... I assume we need to bolt them together before we get started...
There’s the Edison, and the Explorer Board. the Edison just pops on, but yes you can also screw it on to be secure.
@danamlewis quick question... one of our local cycling mags reached out to me to do a quick article... can I mention what we are doing or is there some risk in discussing?
@scottleibrand what would be the result in autotune of carbs not registered in ns and the gaps in bg caused by sensor warmup?
the gaps in BG might affect its estimate of CSF and IC ratio
@RedCyclist1 by all means share! :smile: let me know if you need any background links to point the reporter to.
@dougtakeuti particularly if the gaps occur when most of the carb absorption is occurring, it might think that not all of the carbs ended up absorbing, and therefore that your CSF needs to be lower
We usually don't register carb corrections
if you fail to enter carbs, then it will think the deviations from those carbs are attributable to basal, and will try to raise the basal levels to compensate
@danamlewis Not sure of the intent or discussion but I assume it is around being a T1D cyclist, I figured I would talk about how I am working to get better controle with Open APS... I will be sure to direct him your way if he wants details.
so you should ignore autotune telling you to raise basals for the 3h or so prior to whenever you tend to do non-entered carbs
ok, we need a couple of weeks of proper data logging then
yeah, that'd be best. you can still look at overnights, though, if you don't have to correct much then
and if there's any time of day it wants a lower basal, then you know that's not due to missing data
(This is also going to be an issue for anyone using Loop who logs fake carbs....I wouldn't use data from any time periods where you're entering fake data)
true. someone should write up and PR these caveats into the autotune page.
@Kdisimone I'm also thinking about edits to the Loop guide about that (fake carbs, and using autotune in general since we had several Loop-ers test today)
YES definitely need that. If I can get it stopped as suggestion it on FB group, it might get some traction. ESPECIALLY with an acutal reason now in place to use the data so meaningfully.
I would differentiate between fake carbs used to trigger a correction, vs. carbs entered in place of fats/proteins. if there is something that will cause a positive BG deviation, entering an estimate of that is probably a good thing. but just entering fake carbs instead of a temporary target is probably not so good.
Yea, talking about the latter (fake instead of adjusting absorption or actual original carb count or using a temp target to adjust Loop outcomes)
autotune will compare how much BG actually rose after a meal to the carbs entered, and use that to calculate CSF. so the meal entry should reflect, as accurately as possible, the estimate of how many carbs' worth of BG rise are expected
yes…the problem is that there’s been…oh never mind ^^that’s what I was goign to say. FB has had a lot of fake carbs for corrections of stubborn highs
k
I just tried to use autotune on a couple of days and at the very end it showed
Autotune pump profile recommendations:
---------------------------------------------------------
Recommendations Log File: /root/myaps/autotune/autotune_recommendations.log
/usr/local/bin/oref0-autotune-recommends-report: line 43: jq: command not found
/usr/local/bin/oref0-autotune-recommends-report: line 44: jq: command not found
@dougtakeuti in https://github.com/openaps/docs/blob/master/docs/docs/walkthrough/phase-4/autotune.md#phase-c-wip-running-autotune-more-easily-as-an-average-user step 2, 2nd bullet is the fix for that
:+1:
@SandraK82 did you get your autotune error figured out?
@scottleibrand I just made a PR for nightscout dev, disconnected deployment source in azure, chose source and re-deployed, and now my nightscout site is showing the dashes and no data. did I do something wrong?
I’m having a devil of a time getting the cgm-remote-monitor dev branch to work too all day. i’m on master now instead
I can switch back to master, was wondering if I had done something wrong
@danamlewis what does it mean to flash the Edison?
flashing the Edison means installing a new operating system on it
@ahanuban I'm not an expert on NS. and I use a non-standard setup on my own VM
lots of others here will be more help on that
Master NS is fine for everyone these days though, dev not needed unless you're testing something particular
thanks @danamlewis
here's a video of starting flash....I'm about to publish the finish flashing video https://www.youtube.com/watch?v=kCFO3W0IYxQ
@Aaybob - Wow thanks!
On the Autotune stuff @scottleibrand @danamlewis , even with OpenAPS I'm a massive user of carb equivalents to manage protein using AMA. I go too high without it. I'll run an Autotune with the last weeks data, but aside from the last couple of days, a lot of the carbs are handling protein.
I haven't found a way around it.
(And with AMA I have to run at 0.75 protein instead of the 0.5 I can use with dualbwave boluses )
I'm not sure how Autotune will handle that so I haven't trialled it yet.
ran oref0-setup and got the following error message:
Command python setup.py egg_info failed with error code 1 in /tmp/pip-tVsjGA-build
Storing debug log for failure in /root/.pip/pip.log
Couldn't install spi_serial
@tim2000s once we get it finished you might like the supermicrobolus stuff for dealing with those meals.
Actually, being a little dumb, I'd forgotten that I'm running dev and it's there anyway. It's quite interesting looking at it as the recommendations, given that I'm approximating carbs for protein, are pretty much spot on with where I am already.
SMB basically allows AMA to get the same amount of insulin in 15m that a temp basal would do over 30.
:+1:
@tim2000s :+1:
There's obviously a couple of areas where work and weekend days differ and I can see that showing up as a change suggestion that I was already aware of and thinking about.
Nice
Would be interesting to maintain different weekday and weekend profiles and autotune them separately.
Should be a really minor tweak (wrapper script probably) to do that.
:+1:
Would be great for a lot of people
Yes, it's something I've thought pumps should have for a long time.
I would love that. Anna’s school days are markedly different for about 2 hours in the morning and 2 hours in the later afternoon.
I was just thinking about your two Loops @Kdisimone
Someone want to create an issue?
Someone want to open an issue on that?
@scottleibrand and I are out to dinner
Jinx. ;-)
And jinxing but eating ;)
Lol
@tim2000s I bet you could do the PR for that one too.
:+1:
I already have that blog post in my head if someone wants to make it happen ;)
Thinking about it, it's a bit cleverer than that. Using a basic learning algo, you could determine the change in getting up, etc, look at glucose levels, then adjust the basal off the back of it.
I would start simple and iterate.
It might take a week or two to automate, but once you've figured the learning algo, it can be us d for a lot more based on glucose changes.
Autotuning different days of week differently doesn't require any changes to the underlying tuning algorithms, so that is low hanging fruit.
That would also give you better idea of what kinds of differences your ML algo should be looking for.
Scheduling basal profile changes should be fairly straightforward. Taking it further and combining autotuning with a basic learning calendar is where it gets interesting.
You win at whatever gets done ;)
But seriously - would be great to log an issue and outline various ways to approach it
here's the finish flashing video https://www.youtube.com/watch?v=btU771DIpB8
Let me get pancreabble fixed then I'll take a look at basal scheduling.
:+1:
@ahanuban That's the same pump model we've got. mmtune reports it at 915Mhz.
@jmcrawford THANKS
On the Autotune data, is there a required source? If you can pull an extract from Diasend, can that be used? Thinking about wider use for people in Europe using Libre who perhaps aren't configured with Nightscout. What format for data is required? It's the kind of tool that HCPs would find really helpful.
(Just twitter shared your blog post to two of the more forward thinking HCPs on twitter in the UK - think they'll be interested!)
https://twitter.com/danamlewis/status/822994632715816961 @tim2000s we are on the same page ;) just needs someone to do plumbing probably to confirm what structure of file needed
Thx for sharing @tim2000s :smile:
Alright, here at step 1. I like Digital Ocean - anyone already using?
@danamlewis just asked diasend/gluko the same question 👍
nice :)
Time for bed though. Goodnight!
@kellyisworking we’re on DO, but someone today spun up w/ google cloud and seemed happy with it
@tim2000s night!
nice. thanks!!!
:+1:
openaps/oref0#340 <— please add discussion there for the multiple day/different type of buckets for tuning w/ autotune discussion.
@Aaybob IF we don't have a grounding mat, is there anything else we can do to be careful of static electricity?
live somewhere with decent humidity? ;-)
(static isn't much of a problem in Seattle for some reason) :-P
per https://github.com/openaps/oref0/issues/340#issuecomment-274306255, I just wrote a quick script to test day-of-week autotuning, if anyone wants to do the honors. :-)
:+1:
Later tonight 👍
@ahanuban I just attempted the install with oref0-setup.sh in dev and got the same error about spi_serial install... did you figure this out?
@jyaw I installed it manually and then re-ran the script; it works now. But it would be better not to have to do that...
ah, installed spi_serial manually?
let's troubleshoot please
I don't want to lose the chance to fix it
what does /root/.pip/pip.log show?
.... you talking to me?
yes, if you got the same error @ahanuban did
cool
@jyaw it would have to be you right now because I'm not near my testing rig unfortunately
his said
Storing debug log for failure in /root/.pip/pip.log
checking now...
hey, this is @jyaw... just on my buddy's computer now. He's getting started tonight :smile:
heh k
scottleibrand @scottleibrand waits to see if @jyaw starts arguing with himself
it says....
------------------------------------------------------------
/usr/bin/pip run on Sat Jan 21 22:21:52 2017
Downloading/unpacking git+https://github.com/EnhancedRadioDevices/915MHzEdisonExplorer_SW.git@master
Cloning https://github.com/EnhancedRadioDevices/915MHzEdisonExplorer_SW.git (to master) to /tmp/pip-NwISGL-build
Found command 'git' at '/usr/bin/git'
Running command /usr/bin/git clone -q https://github.com/EnhancedRadioDevices/915MHzEdisonExplorer_SW.git /tmp/pip-NwISGL-build
Running command /usr/bin/git show-ref
de609d222ef77607874d06ec605fa53f745ac8d2 refs/heads/master
de609d222ef77607874d06ec605fa53f745ac8d2 refs/remotes/origin/HEAD
52707359257c097ef46b438d74cf210ce74d4f01 refs/remotes/origin/ResetLED
de609d222ef77607874d06ec605fa53f745ac8d2 refs/remotes/origin/master
Running command /usr/bin/git rev-parse HEAD
de609d222ef77607874d06ec605fa53f745ac8d2
Running setup.py (path:/tmp/pip-NwISGL-build/setup.py) egg_info for package from git+https://github.com/EnhancedRadioDevices/915MHzEdisonExplorer_SW.git@master
Traceback (most recent call last):
File "<string>", line 17, in <module>
IOError: [Errno 2] No such file or directory: '/tmp/pip-NwISGL-build/setup.py'
Complete output from command python setup.py egg_info:
Traceback (most recent call last):
File "<string>", line 17, in <module>
IOError: [Errno 2] No such file or directory: '/tmp/pip-NwISGL-build/setup.py'
----------------------------------------
Cleaning up...
Command python setup.py egg_info failed with error code 1 in /tmp/pip-NwISGL-build
Exception information:
Traceback (most recent call last):
File "/usr/lib/python2.7/dist-packages/pip/basecommand.py", line 122, in main
status = self.run(options, args)
File "/usr/lib/python2.7/dist-packages/pip/commands/install.py", line 290, in run
requirement_set.prepare_files(finder, force_root_egg_info=self.bundle, bundle=self.bundle)
File "/usr/lib/python2.7/dist-packages/pip/req.py", line 1230, in prepare_files
req_to_install.run_egg_info()
File "/usr/lib/python2.7/dist-packages/pip/req.py", line 326, in run_egg_info
command_desc='python setup.py egg_info')
File "/usr/lib/python2.7/dist-packages/pip/util.py", line 716, in call_subprocess
% (command_desc, proc.returncode, cwd))
InstallationError: Command python setup.py egg_info failed with error code 1 in /tmp/pip-NwISGL-build
haha
I think setup.py is just down in the spi_serial directory, but its not looking there...
yep, that's what I'm trying to figure out how to work around
could just do a link?
hmmm.... so I'm installing the ERD git repo with the spi_serial subdirectory then
let's try something like
pip install -e 'git+https://github.com/EnhancedRadioDevices/915MHzEdisonExplorer_SW.git#egg=spi_serial&subdirectory=spi_serial'
will that know where I want to put it the repo with that syntax?
or does that not matter
I don't think that syntax is quite right: still trying to figure it out
maybe it would be easier to put a setup.py in the root of 915MHzEdisonExplorer_SW
I think so... do you have control of that?
I can do it on my fork and test
and then if it works we can PR
I went ahead and did a pip install . in the spi_serial folder
k. I think I can test from here myself
Resetting spi_serial
Traceback (most recent call last):
File "/usr/local/bin/reset_spi_serial.py", line 3, in <module>
s = spi_serial.SpiSerial()
File "/usr/local/lib/python2.7/dist-packages/spi_serial/spi_serial.py", line 11, in __init__
self.dev = m.spiFromDesc("spi-raw-5-1")
AttributeError: 'module' object has no attribute 'spiFromDesc'
Attempting to communicate with pump:
mmtune: Usage: grep [OPTION]... PATTERN [FILE]...
Try 'grep --help' for more information.
Just before running the mmtune call in oref0-setup.sh I got this error
that means your attempted spi_serial install failed I think
try
sudo pip install git+https://github.com/scottleibrand/915MHzEdisonExplorer_SW.git@master
I'm back. Trying to add my additional basals and am now getting this error: -bash: oref0-autotune: command not found
@rosyred is this the same VM you were on earlier?
@scottleibrand yes
Downloading/unpacking git+https://github.com/scottleibrand/915MHzEdisonExplorer_SW.git@master
Cloning https://github.com/scottleibrand/915MHzEdisonExplorer_SW.git (to master) to /tmp/pip-s9SnPe-build
Running setup.py (path:/tmp/pip-s9SnPe-build/setup.py) egg_info for package from git+https://github.com/scottleibrand/915MHzEdisonExplorer_SW.git@master
package init file 'spi_serial/__init__.py' not found (or not a regular file)
Requirement already satisfied (use --upgrade to upgrade): spi-serial==0.0 from git+https://github.com/scottleibrand/915MHzEdisonExplorer_SW.git@master in /home/.rootfs/usr/local/lib/python2.7/dist-packages
Cleaning up...
@RosyRed the fix should be to re-run step 2, but not sure why it's doing that
ok
perhaps setup.py needs to be modified?
yeah. trying something else
so I just copied spi_serial into a top-level repo, and confirmed that
sudo pip install git+https://github.com/scottleibrand/spi_serial.git works...
@scottleibrand It didn't fix it.
@RosyRed can you paste the output of the
sudo pip install git+https://github.com/openaps/openaps.git@dev command please?
@scottleibrand
-bash: oref0-autotune: command not found
g17thomptes@instance-1:~$ sudo pip install git+https://github.com/openaps/openaps.git@dev
Downloading/unpacking git+https://github.com/openaps/openaps.git@dev
Cloning https://github.com/openaps/openaps.git (to dev) to /tmp/pip-zEgMDU-build
Running setup.py (path:/tmp/pip-zEgMDU-build/setup.py) egg_info for package from git+https://github.com/openaps/o
penaps.git@dev
Requirement already satisfied (use --upgrade to upgrade): openaps==0.2.0-dev from git+https://github.com/openaps/
openaps.git@dev in /usr/local/lib/python2.7/dist-packages
Requirement already satisfied (use --upgrade to upgrade): pyserial in /usr/local/lib/python2.7/dist-packages (from
openaps==0.2.0-dev)
Requirement already satisfied (use --upgrade to upgrade): python-dateutil in /usr/local/lib/python2.7/dist-packages
(from openaps==0.2.0-dev)
Requirement already satisfied (use --upgrade to upgrade): argcomplete in /usr/local/lib/python2.7/dist-packages (fr
om openaps==0.2.0-dev)
Requirement already satisfied (use --upgrade to upgrade): gitpython in /usr/local/lib/python2.7/dist-packages (from
openaps==0.2.0-dev)
Requirement already satisfied (use --upgrade to upgrade): mock in /usr/local/lib/python2.7/dist-packages (from open
aps==0.2.0-dev)
Requirement already satisfied (use --upgrade to upgrade): nose in /usr/local/lib/python2.7/dist-packages (from open
aps==0.2.0-dev)
Requirement already satisfied (use --upgrade to upgrade): decocare>0.0.26 in /usr/local/lib/python2.7/dist-packages
(from openaps==0.2.0-dev)
Requirement already satisfied (use --upgrade to upgrade): dexcom-reader>=0.1.8 in /usr/local/lib/python2.7/dist-pac
kages (from openaps==0.2.0-dev)
Cleaning up...
g17thomptes@instance-1:~$
ok, the problem here is that I'm an idiot. ;-)
lemme fix the docs
@scottleibrand We are at the place to set up the Type/edit the following:
myedisonhostname=<thehostname-you-want>
And then paste the following to rename your Edison accordingly:
echo $myedisonhostname > /etc/hostname
sed -i"" "s/localhost$/localhost $myedisonhostname/" /etc/hosts
Run these commands to set secure passwords:
passwd root
passwd edison
having trouble setting up the initial Edison set up. We are at Type/edit the following:
myedisonhostname=<thehostname-you-want>
And then paste the following to rename your Edison accordingly:
echo $myedisonhostname > /etc/hostname
sed -i"" "s/localhost$/localhost $myedisonhostname/" /etc/hosts
Run these commands to set secure passwords:
passwd root
@jyaw @fowlerbenjamin I just updated oref0 dev branch with that workaround for now
Resetting spi_serial
Traceback (most recent call last):
File "/usr/local/bin/reset_spi_serial.py", line 3, in <module>
s = spi_serial.SpiSerial()
File "/usr/local/lib/python2.7/dist-packages/spi_serial/spi_serial.py", line 11, in __init__
self.dev = m.spiFromDesc("spi-raw-5-1")
AttributeError: 'module' object has no attribute 'spiFromDesc'
Attempting to communicate with pump:
mmtune: Usage: grep [OPTION]... PATTERN [FILE]...
Try 'grep --help' for more information.
Sorry for the duplicate messaging...
I still got that same error after the last command you posted
hmm, odd
is there a way to uninstall/re-install?
pip has an uninstall command
cool... I think I did another manual install and updated the dev checkout, so I'll run the script here in a seconds
just to confirm, you are using oref0 dev, right?
yes
k. I think this only affects people who went straight to dev, because master still installs it the old way (cloning)
so we either need to confirm this fixes it, or go back to the old cloning method
thx for being the guinea pig there. :)
nice late night in the huntsville
nothing much else to do haha
(I created a google form (https://goo.gl/forms/Cxbkt9H2z05F93Mg2) so we can track general experiences about autotune, and linked it to the autotune setup doc page)
Resetting spi_serial
Traceback (most recent call last):
File "/usr/local/bin/reset_spi_serial.py", line 3, in <module>
s = spi_serial.SpiSerial()
File "/usr/local/lib/python2.7/dist-packages/spi_serial/spi_serial.py", line 11, in __init__
self.dev = m.spiFromDesc("spi-raw-5-1")
AttributeError: 'module' object has no attribute 'spiFromDesc'
Attempting to communicate with pump:
mmtune: Usage: grep [OPTION]... PATTERN [FILE]...
Try 'grep --help' for more information.
:worried:
@scottleibrand so I think we'll regress to master and then setup and then update to dev
or you can copy and paste the spi_serial setup commands out of the master oref0-setup
ah, hmmm... sounds better
I did the command from master oref0-setup.sh for the spi_serial stuff and then ran oref0-setup.sh from dev
and I got the same "resetting spi_serial" error from above
rebooting for fun here
just in case
@scottleibrand ? Sorry, you look busy
oh yeah, let me get back to that, thx for the reminder
I don't know how you keep up with all these convo's at once.
not well, apparently. :)
Much better than I would
reboot :smile: :fireworks: :beers:
@RosyRed ok, updated the docs
please refresh and try that step again
Thank you.
I guess that's what needed to happen to get past the error in the first place, so I think your fix works but for some reason it needed the reboot
@fowlerbenjamin k
@scottleibrand Back in Business. Thanks
What's the best way to print these off to review?
there is an --xlsx option that is supposed to generate data for Excel, but I haven't gotten it to work yet
or you could just copy and paste the contents of the ~/myopenaps/autotune/autotune_recommendations.log file
ok
I do
cat ~/myopenaps/autotune/autotune_recommendations.log | egrep "\." to make it easier to read
(that just removes basal lines that are empty)
you could even copy and paste that output into Excel, and then use the Text to Columns feature to play with it there
Ok, ready for bed. Do I just close this? Does it save all that I've done so I just open and run again when needed?
just had someone fill out the form (yay) and commented that they too got the “command not found” error for step 4 - that should be resolved now, for whoever that was if you want to run it again
@RosyRed Yeah, you should be able to just leave the VM running. It's also fine to shut down the VM if you know how to start it back up again from Google Cloud.
and with that in mind, I added an optional email address so if anyone shares bugs and wants us to follow up directly, we’ll be able to.
@scottleibrand Sounds good. Thank you
Night all
night!
thx for all your help testing @RosyRed!
we're getting temp basal recommendations not enacted... what should we check?
We're getting the "received": false field/value in the enacted.json file as well
hoping that doesn't mean the pump isn't loopable
% or absolute temp basals?
That was me. Is there a different page? I was looking at the FB page link.
it was percent
changed to rate
hope that fixes!
Step 4 thing
the magic!
huntsville +1
hey @flutelynne_twitter
we fixed the docs to fix that for @RosyRed - can you refresh your browser page and re-run step 2?
specifically the last two bullets (pull/clone and install oref0 dev)
@danamlewis thanks!
Yes Shut things down. But wil try again
@fowlerbenjamin ;) welcome
@flutelynne_twitter ohh, good - was worried it would be someone who wasn’t on here!
Wow! Amazing!
it worked? yay!
Yes! I ran it with my OpenAPS NS, then also did the Loop NS. Very nice. Now to figure out the carb readouts. But I will do that another day. Hey, if I can figure it out, anyone can… Have a great night!
:) :+1: :clap: that’s what I say! night!
Does anyone know the size of the Phillips screwdriver needed for the M2 screws That attach to the Edison/Explorer board?
I believe it is a Phillips #0 screwdriver but I don't want to waste any money.
If anyone has any clue or knows what screwdriver or screwdriver set they purchased, it would be appreciated.
@diabeticgonewild I just used my fingers but a 0 should work fine
Fingers?! Wow!!! 😜😂
I'm going to give fingernails a trial, Hahahaha! I don't bite my fingernails, so I should be good, even though I don't keep my fingernails long like most chicks, hahaha!
Eh, seems tight enough, you can't get it that tight anyway unless you use pliers or a small socket to hold the nut
Going to give it a trial right now!!!!
Maybe the tip of an edge of a steak knife?
IDK...
But that's awesome! Totally made my day! 😊
Trying it right now...but high off of Lyrica...whoops...totally affects my depth perception at higher dosages so I don't know how it is going to go. Sucks.
@walker0 I tried using my fingernail and also the tip of the edge of a steak knife. Both didn't work. With the fingernail technique my hand/fingers shake to much to hold and twist the screw hole in the grove without losing position in the groove.
Do you have to hold the shaft of the screw on while twisting and loosening the screw?
I just pressed on the screw with my finger and screwed on the nut
Because I can't even hold the screw shaft without shaking uncontrollably
So you didn't hold the shaft, at all?
You were basically able to do it with one hand?
I'm concerned because I don't think I can loosen the screws off of the Edison.
And everyone living at my house has essential tremor at least as bad as me. We all have dexterity problems. It's genetic.
If that's the case, after my family fails to do so, I may have to get one of my friends to help me out.
Or I could just order the Sparkfun Hardware kit for like $3 and pretty much be done.
Cause putting on a screw is easier than loosening a tight small screw
@walker0 So you basically loosened the screw with 1 hand???
No I just grab the nut in one hand and press against the screw head
Screwdriver makes it easy though. Maybe go find one, even one that is a bit too big or a bit too small or a flathead will probably work.
Yeah I will. I will stop by Home Depot tomorrow. Thank you so much for your detailed and solid answers! 😍😊
Can you connect an
xDrip directly to the Explorer board (offline--without using Nightscout)?
I can't find the answer in OpenAPS docs.
Yes, and it's on the list for half a dozen people to document ;)
Hell yeah! Will do!!!!!
I just haven't felt well the past month or so and I am getting back to not being dysfunctional
So going to do it tomorrow
You can hold me to my word
@j4cbo don't know, I'm running G4 :wink:
ah ok
G4 --(???)--> xDrip board --(BLE)--> Edison is a separate thing from G5 --(BLE)--> Edison
i'm planning on trying to get G5-direct-to-Edison working... once my G5 shows up
@danamlewis Thank you for the answer regarding being able to connect the
xDrip to the Explorer board, too! 😊
:+1:
now going live, see twitter :wink:
The #0 is needed for the screw that come with the mini breakout board intel Edison kit. M2 screws is a question to be answered.
@diabeticgonewild I have a great number of screwdrivers. For those screws, if they're tight you need the #0. If they're loose you can follow everybody's suggestions.
@diabeticgonewild I just put the screws in today. Used a set of tweezers to hold the nut, and just a random small Phillips head screwdriver to tighten. It doesn't need to be as small as an eyeglasses kit. But, like @walker0 said, I could get it pretty finger tightened so long as I could hold the opposite site with tweezers.
@RedCyclist1 Sweet yeah great question. You can use a metal rod, I use a medium size screwdriver. Put the point into an electrical outlet (the ground plug), for example a plug strip on the desk. It must be plugged in but can be turned off. Whatever you are comfortable with. The ground plug is the round one, the longer flat one is the neutral ( which is connected to ground at a single point in the house), the shorter flat one is hot. The metal rod is now a ground rod. When you touch it all static electricity will be discharged. That is what I clip my wrist strap and ground mat to.
I wouldn't advise sticking anything into an electrical outlet: too easy to miss and stick it in the wrong hole.
to ground yourself, touching something metal that's plugged in, like your computer, is usually good enough
@Aaybob You gave me the precise info I was specifically asking for initially. Now I know definitively what to get tomorrow morning at Home Depot! 😊
@kdisimone I really appreciate the specific details and techniques you used to get the screws and nuts loosened and disassembled. I am extremely bad with my hands and this will be really helpful. This will save me tomorrow! Thank you!!! 😃
@scottleibrand that's a really good idea @RedCyclist1 Your computer has a metal rod already plugged in to the correct hole. The power chord connects that rod to the metal in the computer frame. If its case is metal too then touching the computer case will even your body voltage with ground. The problem is forgetting. That's why I have a mat. Lets say I pick up the board forgetting, my hand usually touches the mat by accident, good accident.
@scottleibrand @danamlewis Just ran autotune on a google cloud instance. Ran so much faster than my edison. I ran it with a few different compromise 24/7 sensitivities and the suggestions seem much more reasonable than my edison run.
:+1:
what was different about how you ran it on Google cloud vs. your edison, besides the speed of course?
Different sensitivity and a different time frame. I think there might be some weird data in my NS from a while ago that stuffed up the edison run. I'll have to go looking through my NS data. It was hard to monitor it though as I set it running on ~6 weeks of data I think and it took a couple of days to complete the 10 runs I had set it for haha
Also I made a point of following the new instructions and they seem like they work and will be easy enough for most people to follow
Somebody asked on Facebook how much cheaper BASAGLAR, the biosimilar to Lantus, costs versus Lantus. The answer isn't straightforward as the pens are distributed in quantities of 300 units while Lantus is usually in a 10 mL bottle. Also if you "convert" everything into 10 mL, it's easier to compare to other insulins sold. Anyway, it's a ripoff and we don't
end up saving much. I created a Gist and answered their question. Take a look: https://gist.github.com/diabeticgonewild/70e43b21de75f28640592df8aa117a47
@walker0 sounds good. Dunno if you noticed, but we removed all reference to running multiple runs in the new docs. Seems better to have everyone just run it over more days if they want to see more change.
Yep I noticed and it seems like a good idea to me.
Hi! I'm trying to install oref0 to get to autotune. I think something went wrong with some update recently, I get this following message: "Errors were encountered while processing:
mysql-common"
and eventually can't install anything after that. Any idea how I can fix this?
mysql-common"
and eventually can't install anything after that. Any idea how I can fix this?
This was after running the following command:
@scottleibrand no had no time yet to figure it out
@danamlewis @diabeticgonewild I was wondering the same thing about an xDrip. As I watch the raw data, my son's pediatric G4 algorithm seems woefully inadequate for making treatment decisions, especially during rapid changes. I couldn't find anything in the docs regarding xdrip direct connection. And are we talking with no phone, straight to the Pi or Edison? Also- can the xDrip be powered from the Pi or edison without the need for yet another microUSB cable in the mix?
@vincepf @danamlewis @scottleibrand If the Pi can connect via BLE (I am not sure it can) then yes. I don't think the Pi is configured in a way that it can handle BLE connections, even though it can have Bluetooth, plus run the script that corresponds to BLE, as the driver may not be set up in that specific script. If Dana, Scott, or somebody else can answer
and attest to that, it would be appreciated.
As far as I know, ONLY the Edison and Explorer board can handle BLE with
xdrip
I also wouldn't recommend powering the
xdrip with anything else except the embedded lipo that is typically used with the vast majority of xdrip setups. It is not smart to mess around with that, although the Pi MAY be able to power it. The Edison/Exporer can't be used as a power source for the xDrip, no matter what!
The above description is for offline use ONLY
@vincepf if you are not using the serial port of the pi otherwise you could connect the wixel to the serial port of the py. You would still have the problem that you neey the calibration data
@vincepf Seriously don't bother with an offline
xdrip with a raspberry Pi. I seriously don't think it will work anyway. And if you troubleshoot your failure, you will torture yourself with an infinite loop of insanity. You will not succeed.
@C-Ville I was thinking the same thing about calibration data. It would be great to eliminate the receiver, but Transmitter -> xDrip -> Phone -> Nightscout -> openAPS seems like just too many moving parts, with lots to go wrong.
Yes, that sounds mean, but I have learned several things the hard way. I am one of the most stubborn miles you will ever meet and I have done dumb things just to try them out over and over again.
@diabeticgonewild One who has never done dumb things has never tried anything difficult.
I wouldn't recommend powering off of the raspberry Pi, even though you can do that. You are causing MULTIPLE sources of failure. If the raspberry Pi loses power, so does your
xdrip and your CGM data is lost so when you power back up, affecting your OpenAPS loop adversely, guaranteed.
@vincepf the other offline solution that i know of without the g4 receiver apart from colins nice setup would be the repo from @oschumac
@vincepf This is not smart. I have done dumb things and I have learned the hard way on this project. I have been in this particular chat room since January 2015 working on this. I can tell you if you haven't already set up multiple iterations of OpenAPS on your own, following the instructions like HUnDREDS of people successfully have—and choose to deviate
from the norm on the instructions—you will fail and ultimately you will be the loser. You will not win. There are too many peculiar intricacies within this project to win out at trying something completely new—without even having multiple setups created off of the basic OpenAPS doc instructions
@C-Ville I'm sorry, but I don't know who colin is or what exactly his setup involves.
@colinlennon did the implementation of xdrip cellphone data to the edison/pi
@C-Ville Thanks. Looks like right now the only xdrip solution without a phone may be @oschumac 's work.
Does anyone know if you can initiate a calibration for
xDrip directly though SSHing into the Explorer board and entering in some sort of "calibration command" (with Calibration BG value) in the SSH terminal? Or do you somehow have to connect and configure the Explorer board to pull the calibration BG values from Nightscout?
Thank you in advance for an answer. 😃
@diabeticgonewild as far as i know @oschumac solution is based on ssh. But based on the intel breakout board, the explorer board does not have a serial port to connect the wixel directly
Awesome! Thank you so much! I will take a look at his Github page! 😍😃👍
OK so I went through @oschumssc repos which are purely amazing. He does not have a setup for retrieving the calibration data. I wonder if I could "spoof" and inject Code with "fake Calibration data from NS" or inject the embedded
xDrip code/program on the Explorer board.
I guess I need to work on that.
Actually, I haven't been able to read to save my life lately and I really have a problem. Sucks.
Can the edison connect directly to the NS uploader phone (via bluetooth) and read BGs from the xdrip app for openaps purposes? Our uploader phone/service won't support bluetooth tethering, so we can't go that route.
So trying to run autotune on a linux VM - it runs beautifully until creating the recommendations and falls over when it reaches 04:30 this text:
ubuntu@ip-172-31-6-162:~$ ./src/oref0/bin/oref0-autotune-recommends-report.sh ~/myopenaps
./src/oref0/bin/oref0-autotune-recommends-report.sh: line 95: 2
3 - 1: syntax error in expression (error token is "3 - 1")
suggestions?
Ignore me - it is important to set the minutes in profile.json to minutes from midnight, probably worth pointing out the ISF should be relative to mg/dL rather than mmol/L :-)
@C-Ville @C-Ville @tim2000s I'm trying to update the cron entry for xDripAPS, so that you can switch between pulling CGM data from NS and from xDripAPS easily. For some reason though, my rig isn't pulling CGM data from NS - I keep getting
JSON object could not be decoded errors...
It's not an issue for me as I'm using xDripAPS 24/7, but would one of you be able to test my change? The new cron entry looks like this -
* * * * * cd /root/myopenaps && ps aux | grep -v grep | grep -q 'openaps monitor-xdrip' || ( date; cp -rf xdrip/glucose.json xdrip/last-glucose.json; openaps monitor-xdrip) | tee -a /var/log/openaps/xdrip-loop.log; cmp --silent xdrip/glucose.json xdrip/last-glucose.json || cp -up xdrip/glucose.json monitor/glucose.json
The change should mean that the glucose data pulled from xDripAPS will only be copied to
WiFi, if you turn off BT tethering/WiFi hotspot, the file won't be updated by xDripAPS which should mean that the CGM data pulled from NS is used.
monitor/glucose.json when it has changed. So, when your Edison is connected to the internet viaWiFi, if you turn off BT tethering/WiFi hotspot, the file won't be updated by xDripAPS which should mean that the CGM data pulled from NS is used.
@colinlennon As an alternative, could you run xDripAPS all the time, and upload the data to NS when a network connection was available?
Sorry - meant to tag @ceben80 in rather then @C-Ville twice!
@jmcrawford Yip, that's exactly what I do now. The only reason you might want to switch is to save phone battery. BT tethering isn't too demanding, but running a WiFi hotspot drains battery pretty rapidly.
@colinlennon great Colin will update my cron later :wink: 1st my big Eddie :smile: later his tiny brother
Thanks @ceben80 ! Strange problem with my rig. It seems to fail to pull CGM data from NS when it's on a WiFi network and not BT tethered to my phone. Treatments and temp-targets seem to work, it's just
get-ns-glucosethat fails. If I turn on BT tethering on my phone, when the rig reconnects, get-ns-glucose and by extension get-ns-bg works! Weird.
just updated
will see tomorrow how it's works @colinlennon
as u know I'm running with my huawei mobil router so alway WiFi connected to the same IP net
@colinlennon I now have WiFi connection (thus NS connection) and
monitor/glucose.jsonis up to date :+1:xdrip/glucose.jsonis now 7 min old, think it's okay?!
@ceben80 Sounds good! Did you switch off the REST upload the xDripAPS in the Android app?
@alternateal can you PR in that change please?
@scarral is this an existing Linux box, or a new install?
@scottleibrand PR there as #572
Thx. One comment, then it looks good to merge.
@scottleibrand try that version on #572
@colinlennon that sounds awesome
@colinlennon no didn't do that, will make it today later :wink:
@bdr1177 There are apps that can be used for BT tethering that bypass the native tethering methods. Helpful if your provider doesn't allow tethering.
@colinlennon REST switched off and same positive result as posted above
https://gitter.im/nightscout/intend-to-bolus?at=5884f641e836bf70107909e3
https://gitter.im/nightscout/intend-to-bolus?at=5884f641e836bf70107909e3
@colinlennon are you still using yocto or do you BT tether using jubilinux?
I am running through the current dev documents and now I have the question, what the difference is between
and
mkdir -p ~/src; cd ~/src && git clone -b dev git://github.com/openaps/oref0.git || (cd oref0 && git checkout dev && git pull)and
mkdir -p ~/src; cd ~/src && git clone git://github.com/openaps/oref0.git || (cd oref0 && git checkout dev && git pull)
both should get to the dev version of ore0?
the former is correct
the latter won't checkout the dev branch the first time, only if you already have it cloned
where is the latter exactly?
@scottleibrand no need to worry, the dev docs point to master instead of dev and I made the change myself, and was not sure, why there are differences
I'm trying to setup Autotune on a Google Cloud VM, but I get an error "E: Unable to locate package nodejs-legacy" as the last line. I setup an Ubuntu 1204 image
@scottleibrand so the original in the docs says
Step 1: Pull/clone oref0
in phase 2
Pull/clone the latest oref0 master by running:mkdir -p ~/src; cd ~/src && git clone git://github.com/openaps/oref0.git || (cd oref0 && git checkout master && git pull)
@C-Ville wait, the dev docs? we don't use those except for testing out major structural changes
@scottleibrand ah okay I wanted to use dev oref0 for xdrip and thought the dev docs would be the way to go
the dev branch of oref0, yes
but there are no different docs for dev
@jlaes which command did you run that gave you the nodejs-legacy error?
@scottleibrand thanks for explaining, my bad
might be worth PR'ing in a note about that
@C-Ville thanks for working on docs for that :D
@scottleibrand curl -s https://raw.githubusercontent.com/openaps/docs/master/scripts/quick-packages.sh | bash -
the first one from the noob directions
I don't know how to install node-js on my VM
ok. try just
sudo apt-get install nodejs
node has lots of inconsistent stuff with their package naming, unfortunately
I installed nodejs, ran the curl again and got the same error, retrying
oh, yeah, that won't work sorry
you'll have to open up https://raw.githubusercontent.com/openaps/docs/master/scripts/quick-packages.sh and run the commands manually
changing
nodejs-legacy to just nodejs
yeah, doesn't work, same error
ok
I wonder what's different about your VM vs. the Google cloud VMs that did work yesterday
@C-Ville my Explorer rig is Jubilinux, BT tethering working on it.
maybe I should setup a new one?
rather than trying to do things manually
any specific setup instructions for GCloud?
let's see if @tynbendad or @RosyRed or anyone else can tell us which Google cloud VM worked for them
or you can try different ones and see if they work better
I have no idea about the config/options when creating the VM, what to choose and what not to choose and what is relevant and what can be left untrouched
I haven't seen them myself. is there a newer version of Ubuntu you could try?
@scottleibrand maybe i have missed something somewhere, had no json cmd... did a
npm install -g json and now i'm waiting
Ubuntu 16.10
amd64 yakkety image built on 2017-01-03
amd64 yakkety image built on 2017-01-03
@SandraK82 waiting for what?
would that work?
@jlaes worth a shot
waiting for
oref0-autotune --dir=~/myopenaps --ns-host=http://localhost:1337 --start-date=2017-01-22
ah, k. I assume it's running through and giving you good output?
@scottleibrand do I need to select anything for the Application images? (SQL Servers)
@SandraK82 the json install is the third to last line of https://raw.githubusercontent.com/openaps/docs/master/scripts/quick-packages.sh FWIW
OS Image Ubuntu 16.10
amd64 yakkety image built on 2017-01-03
amd64 yakkety image built on 2017-01-03
@jlaes you don't want SQL
ok
yeah did not install openaps, only oref0, missed right the first step :/
anything need to be changed in the Identity & API Access, Management, Firewall etc? Or just create it with defaults?
defaults are probably fine there
any kernel hackers around? i managed booting 4.10rc4 to initramfs, but get stuck mounting the mmc root partition
Ok, I'll try again with this instance
and spidev5.1 seems to work, but i want my mmc root device
Linux VM built on AWS free servers: ubuntu/images/hvm-ssd/ubuntu-xenial-16.04-amd64-server-20161214 (ami-57eae033)
Took 94 minutes to run 22 days worth of data - I'll live with that. Interestingly it has moved my Basal by more than 20% so I guess I missed that in the settings - when I had run it previously (during the holidays) on my RPi it had limited change - this seemed unconstrained.
@scottleibrand this one worked, moving forward. Thanks!
@alternateal your preferences.json has an autosens_min and max setting that control that
if you have the default of 0.7, it can do up to 30% more sensitive, vs. only 20% (1.2) more resistant
@jlaes can you PR in a note about which Ubuntu version worked for you?
@scottleibrand - just me playing around and not taking note of everything again :-)
heh, a lot of this stuff is non-obvious. :)
@danamlewis so, I hope that worked with my first PR
@scottleibrand there isn't a preferences.json when running stand-alone on a VM or if there is I can't find it - agree when running on the looping device - so nothing to constrain.
I wonder on the benefit of running Autotune on more than about a week's worth of data. We know how people's basals can change fairly quickly, and if your data is different today from a week ago, is there benefit to looking at the data that's a month old?
@tim2000s - alternatively using more than one week allows you to dampen out some of the potential one off events that happened in a single week??
^ yes
@alternateal the stuff from preferences.json all goes into profile.json, so it should be there. I'll put it back into the example.
which one to use for decimal for basal rates in the profile? "," or "."
I would assume
.
thanks
it might be possible to configure your VM with a non-US locale to support
,, but I'm not sure
the example doesn't have any values so hard to tell
my VM is not in the US
the locale would be a default from the ubuntu image, so it probably has US
physical location doesn't matter
I'm in Finland, VM is in Europe-West and my decimal separators are always mixed up due to mostly operating in US English
heh, fun. :)
yeah
Excel is the worst
Just thinking it through though, if you are looking at a month of data, say, while you dampen out anomalies, what is the rate of change of basal, ISF, etc? I know mine are fairly slow moving, but we also see those who find them very fast moving.
we try to capture fast-moving changes with autosens and slow-moving changes with autotune
(issue created for the MDI users interested in autotune - will point volunteers there: openaps/oref0#341)
@tim2000s day of week might also address some of those changes, too, for in-between autosens and autotune
I understand that aspect, but I wonder how slow moving we're really talking. Many of the discussions I've had about basal rate changes have seen them being observed over a one or two week period. There's some piece of the statistical regression when looking back over longer than that that irks me I think!
I can see the benefit of day or two days a week as well.
e.g. comparing weekend with weekday
if you haven't already, might want to start testing openaps/oref0#340, as it will take a week to kick in
short term: if you want to review on shorter data, you can. But in general, until we get feedback on wider groups of people testing, I'm not going to feel comfortable making any blanket recommendations at this point about what is right.
(I like where you're going in general w thinking about this, I just think it's premature until we get more time under belts with this. It really was only coded 3 weeks ago :smile: )
yeah, I'm not sure that our ISF recommendations don't have some bias (toward higher more sensitive values), for example. need more people testing autotune and providing feedback.
And there may always be some debate on this..will be interesting to see once we get people to donate data into Open Humans if we can do better tests and compare results from using 1 wk vs 3 was of data?
@danamlewis thinking about your openhuman data and autotune (amongst other random things flitting through my mind as I try to get other work done) - are there (do you think there will be) enough data points to separate out pediatric from adult users - I suspect there will be differences between those groups also.....
All makes sense. I was having a conversation with a couple of HCPs this morning on twitter, and having had a chance to cogitate on it, there's something about variance over the longer term that feels as though you're introducing flutter. As you say, it's worth getting the data and running the tests. It's kind of a gut feel at the moment for me.\
@JELCRAWFORD we'll ask age, so we can run separate age batches if we want to compare with and with peds mixed in, or even multiple kids age groups ..but really comes down to who donates data :+1:
@tim2000s :+1:
@/all for all interested: ti-usb over serial to raspi: https://github.com/oskarpearson/mmeowlink/wiki/TI-USB-Stick-over-Serial
I am getting the following error, when running
curl -s https://raw.githubusercontent.com/openaps/docs/master/scripts/quick-packages.sh | bash -Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:
The following packages have unmet dependencies:
nodejs-legacy : Depends: nodejs (>= 0.6.19~dfsg1-3~) but it is not going to be installed
npm : Depends: nodejs but it is not going to be installed
Depends: node-abbrev (>= 1.0.4) but it is not going to be installed
Depends: node-ansi (>= 0.3.0-2) but it is not going to be installed
Depends: node-ansi-color-table but it is not going to be installed
Depends: node-archy but it is not going to be installed
Depends: node-block-stream but it is not going to be installed
Depends: node-fstream (>= 0.1.22) but it is not going to be installed
Depends: node-fstream-ignore but it is not going to be installed
Depends: node-github-url-from-git but it is not going to be installed
Depends: node-glob (>= 3.1.21) but it is not going to be installed
Depends: node-graceful-fs (>= 2.0.0) but it is not going to be installed
Depends: node-inherits but it is not going to be installed
Depends: node-ini (>= 1.1.0) but it is not going to be installed
Depends: node-lockfile but it is not going to be installed
Depends: node-lru-cache (>= 2.3.0) but it is not going to be installed
Depends: node-minimatch (>= 0.2.11) but it is not going to be installed
Depends: node-mkdirp (>= 0.3.3) but it is not going to be installed
Depends: node-gyp (>= 0.10.9) but it is not going to be installed
Depends: node-nopt (>= 3.0.1) but it is not going to be installed
Depends: node-npmlog but it is not going to be installed
Depends: node-once but it is not going to be installed
Depends: node-osenv but it is not going to be installed
Depends: node-read but it is not going to be installed
Depends: node-read-package-json (>= 1.1.0) but it is not going to be installed
Depends: node-request (>= 2.25.0) but it is not going to be installed
Depends: node-retry but it is not going to be installed
Depends: node-rimraf (>= 2.2.2) but it is not going to be installed
Depends: node-semver (>= 2.1.0) but it is not going to be installed
Depends: node-sha but it is not going to be installed
Depends: node-slide but it is not going to be installed
Depends: node-tar (>= 0.1.18) but it is not going to be installed
Depends: node-underscore but it is not going to be installed
Depends: node-which but it is not going to be installed
E: Unable to correct problems, you have held broken packages.
any ideas?
could this be the problem?
edison@AP-Server:~$ nodejs -v
v4.7.2
edison@AP-Server:~$ node -v
v4.7.2
you might need to sudo
@tim2000s you mean to sudo for the curl ?
yes
Is there a missing
. after the bash -
I think it should be
bash - .
@scottleibrand a couple hardware questions for openomni - openaps: 1- to start getting bolus+basal data from Omni, since that shouldn't depend on solving the nonce problem, what hardware should I start with (Edison + ?). 2- Is anyone already working on this, esp with xdrip+ or HAPP? 3- I'm in Seattle (LA for a couple days next week), in case there's a chance to have lunch or coffee with someone. Thanks!
@tim2000s it is copied from the docs and I think the same script did fine on my other rig. However, I was hoping to get my old rig updated as it holds the xdrip and NS and mongoDB all in the same housing
yes, you're right
@tim2000s I needed to upgrade node to a recent version to be able to run NS, and I wonder if this screws up the curl script somehow
@seattlebrighton you probably want to join the Omni slack and see the work that's being done there and ask hardware Qs. https://omniapsslack.azurewebsites.net
my other rig has
edison@miniaps:~$ node -v
v0.10.29
edison@miniaps:~$ nodejs -v
v0.10.29
edison@miniaps:~$there the docs could be followed easily
@danamlewis thanks. I'm already there (I think maybe one of the first). I didn't think people there were yet using Edison software but I'll ask.
@seattlebrighton ask there, that group is the most active on it
@scottleibrand it's an existing linux desktop.
stupid question, if I stop my instance, is it back to normal doing another autotune later when I just start it again? Or should I just leave it on since it's free for now on Google Cloud first year anyway?
@jlaes doesn't matter.
@scottleibrand thanks, thought so too, but this is all new to me.
@scarral you'll have to figure it out on your own then, or spin up a new clean VM.
there are lots of online resources for resolving apt-get errors
just got tripped up on the dev version of oref0 install script.
line 340
if ! python -c "import openxshareble" 2>/dev/null; then
echo Installing openxshareble && sudo pip install git+https://github.com/openaps/openxshareble.git@dev || die "Couldn't install openxshareble"
fidoesn't install repository, just installs package. But then a few lines down it is assumed repository is installed. Will work on it ...
Interesting to read the recommendations, we had a bad first week with a miss behaving RileyLink that got replaced and I get significant differences for ISF, CSF and CR if I run a 2 week report or a 3 week report - basals are quite good on both. Our problems with RL were mostly missed boluses and not high temping after meals at all, so that might have an impact on that. Need to run another normal week at least to get more trending data to make proper conclusions. But an awesome tool, thank you @scottleibrand and @danamlewis !
@scottleibrand found another error in my setup. As I use a vps server i already have from a german Hosting Company for autotune it has german locales preselected and that's gives me number format errors (separator is , not . In german)
@tghoward my bad. Thanks.
@scottleibrand, suggest best way to handle? Install repository?
We were hoping to move away from that.
@jlaes :+1: thx for the feedback!
What does it need out of it?
it wants to copy bluetoothd.conf
sudo cp $HOME/src/openxshareble/bluetoothd.conf /etc/dbus-1/system.d/bluetooth.conf
Can we just wget that?
I'm not sure what the changes are, but I can look into it. I'd assume we could just tweak the existing, yes.
That's true, or just wget from the repository online, is that what you mean?
@scottleibrand now it works for me :heart:
:+1:
@tghoward Yeah
@scottleibrand uh new error:
~/src/oref0/bin/oref0-autotune-prep.js ns-treatments.json profile.json ns-entries.2017-01-03.json > autotune.1.2017-01-03.json
Error: carb_ratio 2.727 out of bounds
The good news is... I am running kernel 4.10.0-rc4 on an Explorer board. The bad news is... I get a problem with mraa.
# uname -a
Linux openaps 4.10.0-rc4+ #2 SMP Sun Jan 22 10:33:38 CET 2017 x86_64 x86_64 x86_64 GNU/Linux
# reset_spi_serial.py
Traceback (most recent call last):
File "/usr/local/bin/reset_spi_serial.py", line 6, in <module>
exec(compile(open(__file__).read(), __file__, 'exec'))
File "/root/src/915MHzEdisonExplorer_SW/spi_serial/scripts/reset_spi_serial.py", line 3, in <module>
s = spi_serial.SpiSerial()
File "/root/src/915MHzEdisonExplorer_SW/spi_serial/spi_serial/spi_serial.py", line 7, in __init__
self.cs0 = m.Gpio(23)
File "/usr/local/lib/python2.7/dist-packages/mraa.py", line 995, in __init__
this = _mraa.new_Gpio(pin, owner, raw)
ValueError: Invalid GPIO pin specifiedAny clue how to fix/debug this?
Thanks @ecc1 for helping me with mounting the root filessytem with the new kernel
@SandraK82 i had similar errors. when i run the oref0-autotune shell script it went fine
@PieterGit what do you mean? It's the output from the oref0-autotune script with just more days then I tried with my first verification
@SandraK82 : i had similar carb ratio errors when i tried the commands on https://github.com/openaps/oref0/issues/261#issuecomment-269230720 a while back. when i used the oref0-autotune shell script later i didn't get those errors anymore. i don't use the generated files automatically , so that might be a bug
@PieterGit i use the autotune instructions from the docs https://github.com/openaps/docs/blob/master/docs/docs/walkthrough/phase-4/autotune.md and have provided my current profile from my running rig
do you run at as phase A or phase C
i used the phase A method
C
ok, didn't try that. the A approach worked for me
i still have a libmraa problem with explorerboard with kernel 4.10rc4.
i noticed the following in the
i noticed the following in the
/var/log/syslog:
Jan 22 22:14:45 foobar libmraa[3233]: libmraa version v1.5.1-18-ge386300 initialised by user 'root' with EUID 0
Jan 22 22:14:45 foobar libmraa[3233]: gpio214: init: Failed to open 'export' for writing: No such file or directory
Jan 22 22:14:45 foobar libmraa[3233]: edison: Arduino board not detected, assuming Intel Edison Miniboard
Jan 22 22:14:45 foobar libmraa[3233]: libmraa initialised for platform 'Intel Edison' of type 2
Jan 22 22:14:45 foobar libmraa[3233]: gpio110: init: Failed to open 'export' for writing: No such file or directory
Jan 22 22:14:48 foobar SubgRfspyLink opening serial port
Jan 22 22:15:40 foobar SubgRfspyLink message repeated 2 times: [ opening serial port]
@ecc1 or @oskarpearson or others: any clue?
@PieterGit Glad to know you've got a later kernel up and running - that's pretty neat
It looks to me like it doesn't have drivers for the Edison GPIO - either they aren't enabled or they aren't included in the later kernel by default
@scottleibrand just noticed your request for a PR, where do you want me to PR the Ubuntu version?
Let me find some details:
The libmraa drivers will try and write a '1' or '19' or something like that to /sys/class/gpio/export
That 1/19/whatever correlates with a gpio pin on the hardware board. By writing the number to that export file, it makes the kernel aware that the user-land code (the programs you're running) want to use that gpio pin.
The kernel then makes that pin available through a series of files in /sys/class/gpio/gpioN/
Once done you can write the same pin number to the 'unexport' file and it'll then remove the /sys/class/gpio/gpioN/ directory
Looks like the defconfig for the new kernel might not have everything you need. I think you need some or all of CONFIG_I2C_GPIO, CONFIG_SPI_GPIO, CONFIG_GPIO_SYSFS enabled
@oskarpearson : thanks for the quick answer. i noticed that i don't have the /sys/class/gpio files. the gpio got recently added, see https://communities.intel.com/thread/75472?start=60&tstart=0 but I think it misses either something in the kernel, or something on my system
My bet is this one: CONFIG_GPIO_SYSFS
CONFIG_GPIO_SYSFS is not set
Things like libmraa hide that from the user - but it looks to me like the export file's not present - which means it either doesn't have gpio enabled at all (see Eric's suggestion) or it hasn't detected available gpio hardware
Looks like it's that, then
i got a
/dev/spidev5.1 file, so I was happy, but i think it also needs the config_gpio_sysfs=y
i only have
$ grep -i GPI x86_64_defconfig
CONFIG_KEYBOARD_GPIO=y
CONFIG_GPIOLIB=y
CONFIG_GPIO_MERRIFIELD=y
CONFIG_GPIO_INTEL_MID=y
enabled
kernel make edsion_defconfig also adds
CONFIG_SPI_BITBANG=y to the list
new kernel is compiling.i will let you know tomorrow what the results are. this might make my dex usb on explorer board stable with usb otg. Is anybody using using Dexcom USB with USB OTG (not bluetooth) on a daily basis? I didn't get USB OTG stable on a 3.10 jubuntu kernel with a edison (tried explorer board and intel baseboard)
thanks @oskarpearson and @ecc1
Depending on the breakout board, you either have to power it with >7V (Intel mini-breakout) or it currently is broken (Explorer), or ??? (Sparkfun), I think
i tried with >7V on intel mini breakout, didn't get it working/stable on an explorer board (once it showed the dexcom, otherwise lots of
lsusb -99 errors). Don't have a Sparkfun board
@PieterGit I got Dex USB working on an Explorer board but it stopped working after a while. I'm unsure if it's something I did or it just stopped after a time but I mentioned it to Scott and opened an issue on the Explorer board repo. Once it stopped working I was getting lsusb -99 errors nonstop.
PRing about
xDrip directly connected via BLE to the Edison/Explorer for an offline OpenAPS in lieu of a configuration with Nightscout is not going to happen today for me. I am sorry I cannot keep my word. I had an extremely physically demanding day and I can't push myself any further. I have a demyelinating disease and exhausted--which I feel--literally becomes an understatement to describe how I actually feel with such a medical problem. Sorry.
@PieterGit EnhancedRadioDevices/915MHzEdisonExplorer#14
I’m trying to set up BT tethering. I have an second edison that I have put on the explorer to try this. I have the dev branch downloaded into src dir. when I run
cd ~/src/oref0/ && npm run global-install i get -bash: npm: command not found. Should I keep going or try to resolve this?
anyone got mongodb and NS running on an Edison with the current oref0 and current OpenAPS?
@PieterGit Not sure if it's been mentioned yet, were you running the spi command as root?
@scottleibrand that bluetoothd line adds permissions to the dialout group. I'm having success running openxshareble without those lines, perhaps because we are running as root? At any rate, it might be ok not to even copy that file over. Will keep exploring.
@RussGG You need to consult Google by putting the returned response from console, copied and pasted, in quotes into Google to look for an answer that works. This is not an esoteric issue that ONLY is related to OpenAPS. Hence Google.
@RussGG Just a second. I am posting a possible solution. I just want to do it neatly withing the chatroom.
@RussGG Found this possible solution here npm/npm#3125 https://www.irccloud.com/pastebin/tCoUPPyo/RusGG_help.txt
Please open the text file link.
Also the red higlighted issue/link
Thanks! I appreciate the patience!
You're welcome! :smile: Patience is a virtue to live by! :+1:
@cjo20 : yes i run everything on the explorer as root
@jaylagorio : i don't believe it's a hardware issue. with a recent kernel i don't have
@jaylagorio : i don't believe it's a hardware issue. with a recent kernel i don't have
lsusb -99 errors
@RussGG if you don't have npm installed, an earlier step failed. I would go back and re-do the dependencies.
@tghoward true, we don't need dialout if running as root.
I got as far as creating a google cloud account. Now I have no clue where to enter anything.
@kmoran02 for autotune you mean? Are you on Mac or PC?
PC
Do you have putty? (Command line tool?)
I think I'm going to give up though. Google cloud shows $300 a month after a one month free trial.
@PieterGit I hope you're right! (@scottleibrand) I'd be happy to flash a new image if it would eliminate that, I'm on the stock Jubilinux linked to from the docs.
It should be 1 year free w 5$/month after that