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

18th
Mar 2016
Tim Howard
@tghoward
Mar 18 2016 00:50
@scottleibrand when ssh-ing into an edison, can you get colors in vim or at bash prompt?
I got so used to color on the RPi :smile:
Scott Leibrand
@scottleibrand
Mar 18 2016 00:52
Yeah, just have to copy over the bashrc.
Tim Howard
@tghoward
Mar 18 2016 00:54
huh, ok. I thought I tried choosing a colorscheme in vim and it stayed b&w. I'll keep futsing. Thanks. Good to know it's fine for you.
Scott Leibrand
@scottleibrand
Mar 18 2016 00:56
Vim is the vimrc
.bashrc and .vimrc in your homedir.
I have lots of example stuff I can share later if you want to go to town with syntax highlighting and fancy bash history stuff etc.
Tim Howard
@tghoward
Mar 18 2016 01:08
I was wrong, just got colors working in vim. Yaa. Will try .bashrc later.
Ben West
@bewest
Mar 18 2016 01:15
got edison booted, had to build dfu-util and xfstk from source, got a technique to insert wifi credentials before first boot, which is nice
dougtakeuti
@dougtakeuti
Mar 18 2016 01:59
@bewest Is you wifi technique accessible to mere mortals or do you have to be initiated? I'd be very interested on that.
Ben West
@bewest
Mar 18 2016 02:04
I mount the image on my machine before flashing/installing
on linux, I do this: sudo mount -t ext4 edison-image-edison.ext4 mnt/
then I can edit mnt/etc/wpa_supplicant/wpa_supplicant.conf in advance
then umount mnt/ to save it, then flash/install as normal
oh mac, it can be done using hmmmm hdiutil mount edison-image-edison.ext4 maybe?
same trick works on rpi images
so before first boot, you can edit the sd card... usually inserting the sd card again will bring up the filesystem
diskutil gui might have a menu for it also
dougtakeuti
@dougtakeuti
Mar 18 2016 03:19
Thanks, I'll save that for my next edison
Brian Bonner
@itsn0tm3
Mar 18 2016 03:55
hmm is the pyserial not in python anymore?
I'm trying to build another aps but the pyserial isn't installing, says it's not found now
I remember it working before
nevermind
got it
I guess the server's acting funny tonight
I'm not wearing enough green apparently
Ben West
@bewest
Mar 18 2016 03:58
hehe
Dana Lewis
@danamlewis
Mar 18 2016 04:44
UK folks, just booked flights :) London for evening of 25 or 26 May! guess I need to figure out how many people are up for a meetup, then we can pick a good location.
Brian Bonner
@itsn0tm3
Mar 18 2016 04:53
sounds like fun!
Brian Bonner
@itsn0tm3
Mar 18 2016 05:12
the range on that TI stick is incredible (compared to the carelink!)
I'm sitting 2 rooms away from it and it's still working
Ben West
@bewest
Mar 18 2016 05:12
yeah, it's hard to get used to
Brian Bonner
@itsn0tm3
Mar 18 2016 05:13
I'm wondering how far it would work at the cube farm at work
if it can poke through my plaster walls this well, I wonder if it'd cover the whole center ha
Ben West
@bewest
Mar 18 2016 05:14
another possibility for that kind of situation is to put several out there
Brian Bonner
@itsn0tm3
Mar 18 2016 05:15
I have really strict IT admins so they'd probably frown at that :(
Ben West
@bewest
Mar 18 2016 05:15
but since it's a fixed location, you can have them "port forward" the serial port
ah
hehehe
you can forward all the sticks onto one machine
and have openaps on that machine just figure out which stick is best
Brian Bonner
@itsn0tm3
Mar 18 2016 05:16
how would that work?
hardwire the TI stick to the serial, then feed that through ethernet to a machine?
Ben West
@bewest
Mar 18 2016 05:16
would test each stick using --port during preflight
there's a tool called socat which can take a serial port and forward it to tcp/ip port
it can also do the reverse
Brian Bonner
@itsn0tm3
Mar 18 2016 05:17
hmm no reason why I couldn't do that here at home too?
Ben West
@bewest
Mar 18 2016 05:17
could work, yeah
Brian Bonner
@itsn0tm3
Mar 18 2016 05:17
that would stop all of the PI's from fighting
Ben West
@bewest
Mar 18 2016 05:17
yeah, it'll take a little elbow grease to make it work
Brian Bonner
@itsn0tm3
Mar 18 2016 05:17
I wouldn't mind making a doc out of it if I can get it to work
I can think of quite a few situations where that would be beneficial
running this one on the machine with the stick plugged in can forward the stick to a tcp/ip port
this one can do the reverse on the remote/server machine https://github.com/bewest/decoding-carelink/blob/master/bin/socat_run_app.sh
it creates a serial port on the remote machine by forwarding from some tcp/ip port
I tested it a few years ago here, https://github.com/bewest/insulaudit-ssh-tools
Brian Bonner
@itsn0tm3
Mar 18 2016 05:20
mind.blown.
I'll def have to take a look at it all and give it a go this weekend!
thanks for sharing!
Ben West
@bewest
Mar 18 2016 05:22
no prob, hope you get a kick out of it
Brian Bonner
@itsn0tm3
Mar 18 2016 05:23
I'll let you know if I get it working!
bedtime for me though
Matt Pressnall
@logichammer
Mar 18 2016 06:03

hmmmm...been looping just fine and in the middle of everything work and me NOT testing / messing with the pi when I run openaps report invoke enact/suggested.json I get this back

Traceback (most recent call last):
  File "/usr/local/bin/openaps-report", line 4, in <module>
    __import__('pkg_resources').run_script('openaps==0.0.9', 'openaps-report')
  File "build/bdist.linux-armv7l/egg/pkg_resources/__init__.py", line 726, in run_script
  File "build/bdist.linux-armv7l/egg/pkg_resources/__init__.py", line 1651, in run_script
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/EGG-INFO/scripts/openaps-report", line 82, in <module>
    app( )
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/openaps/cli/__init__.py", line 44, in __call__
    self.run(self.args)
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/EGG-INFO/scripts/openaps-report", line 75, in run
    output = app(args, self)
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/openaps/cli/subcommand.py", line 50, in __call__
    return self.method.main(args, app)
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/openaps/reports/invoke.py", line 23, in main
    device = app.devices[report.fields['device']]
KeyError: None

Needed to reset git as it had gotten corrupted but I haven't messed with any core files...

Scott Leibrand
@scottleibrand
Mar 18 2016 06:04
Probably openaps.ini or one of its include files got corrupted.
Matt Pressnall
@logichammer
Mar 18 2016 06:05
Dah! Is that common? How do folks recover from that in an automated way / protect against it?
Matt Pressnall
@logichammer
Mar 18 2016 06:12
something wacko jacko about my bg_targets perhaps
/usr/local/bin/openaps get-settings
pump://JSON/read_bg_targets/settings/bg_targets.json
settings/bg_targets.json  raised  bytearray index out of range
Matt Pressnall
@logichammer
Mar 18 2016 06:27
Thank goodness for git...all running again!
Scott Leibrand
@scottleibrand
Mar 18 2016 06:32
What was it?
Jon Hughes
@JonnyRox
Mar 18 2016 09:25
@danamlewis count me and Sophie in for the London meetup. We'll make sure we're there!
Hichame Yessou
@hichameyessou
Mar 18 2016 12:01
Are the 713,715 & 723 (firmware <2.4) FULLY supported by openaps?
Oskar Pearson
@oskarpearson
Mar 18 2016 12:33
@bewest I actually tested something very similar to you with socat
I didn't get it to work with the subg_rfspy firmware - did you?
The timing issue that I'm currently having with the subg firmware meant that I missed messages
It might work if we can change the mmeowlink code to use the "send packet and then wait for response" syntax instead of "send message" and "now wait for response"
The issue was the pump would respond before the "now wait for response" message crossed the network and chatted to the subg code
BTW: If you've any ideas for how to refactor the state machine interaction with the command object in mmeowlink it'd be appreciated. Perhaps let's take that to the mmeowlink gitter channel though.
@danamlewis That's great news
(the London visit)
If you keep me up to date with how big the meetup is going to be I can help find somewhere to host it
I've contacts at a couple universities and companies that would be open to that sort of thing - just depends on the size
Colin Lennon
@colinlennon
Mar 18 2016 12:45
@danamlewis Good news! I might try and get down to London if a meetup is on.
Oskar Pearson
@oskarpearson
Mar 18 2016 14:10
On the note of London: I'm registered to go to the NHS hack day Mid-May
They run out of tickets super fast, so make sure to book yours if you're thinking of coming
Eric Luhrs
@eluhrs
Mar 18 2016 15:23
I build a pi-based rig and have been using it for about a week and a half. I followed the web guide, which uses the production version of OpenAPS.
I've been really pleased with the GB results so now im ready to commit.
want to build a smaller Edison version using one of the newer code basses.
basses = bases.
Is there a document that explains which branches have which new features?
And perhaps explaines a little about what the new features do?
im particularly interested in auto sensitivity.
Stephan
@MosiGitHub
Mar 18 2016 15:39
really interesting timestamp in my pumhistory.json ;-)
[
{
"_type": "SelectBasalProfile",
"_description": "SelectBasalProfile 2060-01-19T16:18:16 head[2], body[0] op[0x14]",
"timestamp": "2060-01-19T16:18:16",
"_body": "",
"appended": [
Oskar Pearson
@oskarpearson
Mar 18 2016 15:41
@MosiGitHub That's pretty interesting. Does it happen repeatedly if you fetch history?
Possibly related: bewest/decoding-carelink#145
Stephan
@MosiGitHub
Mar 18 2016 15:57
@oskarpearson was see first time I saw it. I checked the time of the pump and the Edison. Both are matching to the current date. Now it seems to be okay again....If it happens again I 'll let you now
Oskar Pearson
@oskarpearson
Mar 18 2016 16:00
Ok. Might be random corruption of the radio in-transit that happens to coincidentally pass CRC checks. Not great - but not sure how we'd avoid it.
Let us know if you see it again - and try follow the same process as described in that support ticket
We had it for about 30-40 minutes and then it resolved
Stephan
@MosiGitHub
Mar 18 2016 16:58
okay, thanks for your support
Jim Matheson
@jmatheson
Mar 18 2016 17:03
@itsn0tm3 interesting about range. my TI range is pretty poor. I had it on the middle of my nightstand behind my ipad last night and it was not working. had to drag it on the very corner right next to me in order for it to work. mine is maybe 50% better than carelink. I juts got 2 new TI’s in the mail today. will be interesting to see if it is the stick itself, something with my apartment or setup that is causing me to have results not as good as you or others.
Stephan
@MosiGitHub
Mar 18 2016 17:05
@jmatheson I have presently two TI sticks. One has an amazing range the other one is really poor - with the same setup.
Jason Calabrese
@jasoncalabrese
Mar 18 2016 17:07
I've heard that there is a lot of variance between to TI sticks
Wouldn't be a big deal if they weren't as expensive
Bill Stackpole
@billstackpole
Mar 18 2016 17:19
hi all - have some questions regarding scripting the enact/suggested.json - anyone around to offer advice?
'''pi@openaps-raspi:~/billopenaps $ cat enact/suggested.json | openaps use pump set_temp_basal -
{
"bg": 122,
"temp": "absolute",
"snoozeBG": 120,
"timestamp": "2016-03-18T13:18:38.001383",
"missing": [
"duration",
"rate"
],
"reason": "Eventual BG 120>100 but Avg. Delta -0.67 < Exp. Delta -0.2; no temp to cancel",
"eventualBG": 120,
"error": "missing required input fields",
"input": {
"reason": "Eventual BG 120>100 but Avg. Delta -0.67 < Exp. Delta -0.2; no temp to cancel",
"bg": 122,
"eventualBG": 120,
"temp": "absolute",
"snoozeBG": 120,
"tick": "+0"
},
"tick": "+0"
}pi@openaps-raspi:~/billopenaps $
first, not sure if my syntax to enact this setting is correct - thoughts?
Looks like duration and rate are missing - does this make sense to you?
is this because the system is suggesting me NOT to make a change?
Scott Leibrand
@scottleibrand
Mar 18 2016 17:24
In that case oref0 is recommending no action. Correct.
Bill Stackpole
@billstackpole
Mar 18 2016 17:25
so does it look to you like the command I'm sending is correct? (all I did was pass the "suggested.json" into enact and pipe to the openaps use command pointing at my pump?)
do I need to set up a cron job for the looping part?
Scott Leibrand
@scottleibrand
Mar 18 2016 17:25
@eluhrs the only good docs on branch differences are the oref0/issues and pull requests. Kinda low-level.
Bill Stackpole
@billstackpole
Mar 18 2016 17:26
(On break next week from school - want to start looping and debug on the Pi, then move to start building an Edison/TI based system)
also understand that Dana has some survey instrument she is passing around.
<smile>
Scott Leibrand
@scottleibrand
Mar 18 2016 17:27
@billstackpole that command will work, if there is a temp to enact. But for regular use, you want to set up an enacted.json report that does the same thing and puts the results into a file you can upload to NS.
You don't want any use commands in your automated (cron) loop. Should be all reports and aliases that call them.
Bill Stackpole
@billstackpole
Mar 18 2016 17:28
Not sure how NS and OpenAPS interact. Have a NS system in place now (with the watch and such). Can you point me to the readings to get myself up to speed?
missing the report generation for the enact.json
Scott Leibrand
@scottleibrand
Mar 18 2016 17:29
The survey thing was probably the one for the ADA abstract. That has now been submitted. There is also the form to fill out after 3 days looping. That is in the docs.
Eric Luhrs
@eluhrs
Mar 18 2016 17:29
Thanks, @scottleibrand . I think I've read that the meal assist branch includes auto sensitivity. Is that correct? And can you point me to repo?
Bill Stackpole
@billstackpole
Mar 18 2016 17:30
so I'm confused about the docs - is that the gitbook doc or simply the github repo docs section?
Scott Leibrand
@scottleibrand
Mar 18 2016 17:31
The latter. We're trying to deprecate gitbooks. It's supposed to point to the openaps/docs, but doesn't include everything.
Bill Stackpole
@billstackpole
Mar 18 2016 17:33
Ah - hence my confusion. K. I'll chase down the doc that dana has posted. Should I be using the branch specified in the gitbooks approach or re-read the docs directly from the github repo to be sure the branch I'm using is the branch I should be using?
Eric Luhrs
@eluhrs
Mar 18 2016 17:34
@scottleibrand Thanks. I was looking at openaps branches, not oref0. So do I start with a copy of my working openaps install, then just replace oref0 with the dev version?
Jim Matheson
@jmatheson
Mar 18 2016 17:35
Speaking of London, it’s looking like I am probably going to be in town for several days in Mid April if @oskarpearson and others are interested I’d love to meet up.
Bill Stackpole
@billstackpole
Mar 18 2016 17:35
So is the NS project included in the loop control in any way? It appears that NS manages displaying data for me and allowing remote access to that data stream but I wasn't under the impression that the decisions being made by OpenAPS were related to the NS data stream.
Oskar Pearson
@oskarpearson
Mar 18 2016 17:38
@jmatheson Oh good! Sounds great
Scott Leibrand
@scottleibrand
Mar 18 2016 17:47
@billstackpole for your first loop you should use master via the packages (quick-package.sh). All the docs should be identical on that part.
NS doesn't have to be in the loop, but it can optionally be your source of glucose data if you don't have a CGM plugged in.
@eluhrs yeah that's easiest. At this point you could skip dev and go straight to meal-assist if you're interested in the new features. We're almost to the point of merging meal-assist to dev I think.
Eric Luhrs
@eluhrs
Mar 18 2016 18:05
@scottleibrand okay, I've got it now. Thanks for all of your help. I do appreciate it! 👍
Bill Stackpole
@billstackpole
Mar 18 2016 18:27
@scottleibrand thank you!
Jim Matheson
@jmatheson
Mar 18 2016 18:37
does anyone have an alias setup that checks for either a carelink or TI and moves forward appropriately depending on what is avail?
Scott Leibrand
@scottleibrand
Mar 18 2016 18:43
No, I have different aliases on different rigs.
Dana Lewis
@danamlewis
Mar 18 2016 18:49
@hichameyessou not sure how you define "FULLY supported". but yes, those models (with <2.4A) work with OpenAPS. You can read data AND send the necessary temp basal commands.
Greg Scull
@komby
Mar 18 2016 18:53
So while I am waiting to get a pump that works with OpenAPS am I correct in assuming that I will be able to setup OpenAPS to get data from nightscout and use it for reccomendations to send manually to the Omnipod?
Dana Lewis
@danamlewis
Mar 18 2016 18:55
@komby Omnipod is not currently compatible with OpenAPS
you CAN get data from Nightscout for OpenAPS, if you're online. You can also plug a receiver in to make sure it gets offline.
But if you're looking for recommendations to manually enact, then you want something like HAPP, which will give you manual recommenations you can then do by hand on an Omnipod
Greg Scull
@komby
Mar 18 2016 19:01
I am following the Ominpod group closely and hope to be able to help on the dev side soon. We have a ping (and im following marius's work closely) but even though I am a programmer by day I am not brave enough to write that loop in isolation. I was looking for a stop gap where I can start getting my guesstimates vetted by the algorithm on how to bring bg back into range.
I will look at HAPP now
Dana Lewis
@danamlewis
Mar 18 2016 19:02
:+1:
Ben West
@bewest
Mar 18 2016 19:21
for someone using NS classic, it's probably easiest to run everything on your laptop
just to learn how everything works/mechanics