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

25th
Sep 2015
Chris Oattes
@cjo20
Sep 25 2015 00:02
currently I'm playing about with nano on PuTTY, which is fine for modifying things quickly
but that isn't good for larger scale modification of code
Scott Leibrand
@scottleibrand
Sep 25 2015 00:03
vim is pretty powerful if you're willing to learn it
that's what I use for everything
Ben West
@bewest
Sep 25 2015 00:03
I use vim for everything
ok
it helps to give the rpi a wifi dongle
Scott Leibrand
@scottleibrand
Sep 25 2015 00:04
heh yes
Chris Oattes
@cjo20
Sep 25 2015 00:04
http://www.sublimetext.com/ is what I like to use
and yeah, I've got a dongle for the pi
so just 3 things plugged in to it, dongle, carelink and power
Scott Leibrand
@scottleibrand
Sep 25 2015 00:06
Sublime Text would be nice if all the stuff I needed to edit were on my Mac, but it's usually on a server somewhere
might be able to use it with git repos though
not sure if it's worth retraining my fingers from vim though
Chris Oattes
@cjo20
Sep 25 2015 00:07
not sure aobut using it with git
Scott Leibrand
@scottleibrand
Sep 25 2015 00:07
if I were you I would probably edit stuff locally, then git push, and git pull on the pi
you can do those outside of sublime from the command line if it's not integrated
Ben West
@bewest
Sep 25 2015 00:14
blob
just barely caught it
nice
Chris Oattes
@cjo20
Sep 25 2015 00:21
@scottleibrand yeah, the problem I have with going via git is the time it takes to make small changes while tweaking stuff
Scott Leibrand
@scottleibrand
Sep 25 2015 00:28
yeah, for those I use vim on the pi
although I also do a lot of small changes via the edit, git push, git pull workflow
Chris Oattes
@cjo20
Sep 25 2015 00:30
I hate having all of the small "yup, that was a stupid mistake to make" commits in the history :P
Scott Leibrand
@scottleibrand
Sep 25 2015 00:30
git commit --amend :)
Ben West
@bewest
Sep 25 2015 00:30
git rebase -i :-)
Scott Leibrand
@scottleibrand
Sep 25 2015 00:31
although once you've pushed you have to push -f if you do commit --amend
Ali Mazaheri
@amazaheri
Sep 25 2015 00:32
@bewest running a loop! w00t!
Ben West
@bewest
Sep 25 2015 00:32
just stepping manually :-)
Scott Leibrand
@scottleibrand
Sep 25 2015 00:33
hmm, I'll have to try git rebase -i. when do you find that most useful?
Ben West
@bewest
Sep 25 2015 00:33
usually when you want to re-arrange entire branch onto a new branch
Ali Mazaheri
@amazaheri
Sep 25 2015 00:33
@msrcgm I looked at your repo, can you run the commands so we have some new jsons
Ben West
@bewest
Sep 25 2015 00:34
oops wasn't quite enough
need to change basal rate for afternoons
Ali Mazaheri
@amazaheri
Sep 25 2015 00:34
:)
Scott Leibrand
@scottleibrand
Sep 25 2015 00:34
have you been continuing to run the loop since that initial temp?
Ben West
@bewest
Sep 25 2015 00:35
yeah, re-temping
openaps gather-and-predict && json -f 02.diyps/diyps.json
Scott Leibrand
@scottleibrand
Sep 25 2015 00:35
heh, using diyps in there is gonna get really confusing for us
Ali Mazaheri
@amazaheri
Sep 25 2015 00:36
lol
(c)
Scott Leibrand
@scottleibrand
Sep 25 2015 00:36
(TM)
Ali Mazaheri
@amazaheri
Sep 25 2015 00:36
lol
Ben West
@bewest
Sep 25 2015 00:37
then openaps use pump set_temp_basal 02.diyps/diyps.json
this is the "diyp" algorithm, correct?
maybe diyps-recommendation.json might be better
if you have 523 or something, you might send it via notifications same way as diyps classic
Jason Calabrese
@jasoncalabrese
Sep 25 2015 00:50
Got my pi
Chris Oattes
@cjo20
Sep 25 2015 00:57

@bewest how attached are you to

    if not os.path.exists(cfg_file):
      print "Not an openaps environment, run: openaps init"
      sys.exit(1)

in cli/__init__.py?

Ben West
@bewest
Sep 25 2015 00:57
hehehe
well
somewhat?
many openaps operations invoke git
so if someone tries running openaps inside the openaps source code directory, that could easily go quite badly
Chris Oattes
@cjo20
Sep 25 2015 00:59
it might be better to rely on an empty file called .aps or something to do that job
Ben West
@bewest
Sep 25 2015 01:00
it's testing for existence of openaps.ini
Chris Oattes
@cjo20
Sep 25 2015 01:01
I know it is
Ben West
@bewest
Sep 25 2015 01:02
is empty unrelated file better somehow?
this one actually needed
Scott Leibrand
@scottleibrand
Sep 25 2015 01:02
@jasoncalabrese yay! :)
Chris Oattes
@cjo20
Sep 25 2015 01:03
because it doesn't tie the config file to being in the openaps directory (there is already functionality in the reader to check in the cwd, ~/ and /etc/openaps)
Ali Mazaheri
@amazaheri
Sep 25 2015 01:03
nice 2 for price of one, @bewest and @jasoncalabrese looping
Chris Oattes
@cjo20
Sep 25 2015 01:04
I don't know if there is a use case for having multiple openaps environments with the same .ini
Jason Calabrese
@jasoncalabrese
Sep 25 2015 01:04
First need to get it up and ssh to it
Chris Oattes
@cjo20
Sep 25 2015 01:04
that you'd only have to edit once to propogate to both environments
Ali Mazaheri
@amazaheri
Sep 25 2015 01:04
:)
Ed Raskin
@msrcgm
Sep 25 2015 01:09
@amazaheri which commands do you want run? The full progression or just calculate iob
Any thoughts as to what is going wrong?
Ali Mazaheri
@amazaheri
Sep 25 2015 01:10
not yet, please add a bolus
and couple of temps then run your magic alias
Ed Raskin
@msrcgm
Sep 25 2015 01:12
Ok. The boy is back from baseball then dinner. Wanna come over and play pi with me over something Amber?
Scott Leibrand
@scottleibrand
Sep 25 2015 01:12
what? bolus?
Ali Mazaheri
@amazaheri
Sep 25 2015 01:13
Pump is not connected
Scott Leibrand
@scottleibrand
Sep 25 2015 01:13
why are we adding boluses to anythign?
still
we shouldn't be configuring boluses in reports at all
Ben West
@bewest
Sep 25 2015 01:13
it's to get history into an empty pump
Ali Mazaheri
@amazaheri
Sep 25 2015 01:13
not report
Ben West
@bewest
Sep 25 2015 01:13
it's not a report
Ali Mazaheri
@amazaheri
Sep 25 2015 01:13
its IOB calc
Scott Leibrand
@scottleibrand
Sep 25 2015 01:13
ok, so just manual use commands?
Ben West
@bewest
Sep 25 2015 01:13
some people are switching pumps
Ali Mazaheri
@amazaheri
Sep 25 2015 01:13
yes
Scott Leibrand
@scottleibrand
Sep 25 2015 01:14
k
Ben West
@bewest
Sep 25 2015 01:14
with buttons :-)
Scott Leibrand
@scottleibrand
Sep 25 2015 01:14
even better
Ali Mazaheri
@amazaheri
Sep 25 2015 01:14
lol
or ...
@msrcgm we have guests tonight but I should be able to go over them a bit later tonight with couple shots of Espresso
@jasoncalabrese my temp basal treatments magically disappeard from NightScout
:) like someone with eraser wiped them out
Ed Raskin
@msrcgm
Sep 25 2015 01:18
Don't worry @scottleibrand not blousing, just "catching the Medtronic up" to the t:slim bolus for dinner due to pump switch.
Thanks @amazaheri
Chris Oattes
@cjo20
Sep 25 2015 01:18
@bewest https://github.com/openaps/openaps/compare/master...cjo20:configfile that should split 'serial' in to a separate config file, only done limited testing though
Ali Mazaheri
@amazaheri
Sep 25 2015 01:18
@jasoncalabrese actually none of treatments showing up
Scott Leibrand
@scottleibrand
Sep 25 2015 01:23
blousing I'm much less concerned about that bolusing anyhow. :)
Chris Oattes
@cjo20
Sep 25 2015 01:28
@scottleibrand I found this http://www.martinrowan.co.uk/2015/07/live-editing-raspberry-pi-files-remotely-windows-pc-using-sublime-text-rsub-putty/ it seems to work quite well actually. Not sure if it would work with vim though :P
Ali Mazaheri
@amazaheri
Sep 25 2015 01:29
WinScp works great for me
it comes with my buddy putty too :)
Ben West
@bewest
Sep 25 2015 01:45
long live putty, tab tab tab
Ali Mazaheri
@amazaheri
Sep 25 2015 01:45
:+1:
Ben West
@bewest
Sep 25 2015 01:53
@cjo20 yeah, interesting
Chris Oattes
@cjo20
Sep 25 2015 01:56
@bewest you do end up with a bunch of empty sections in the secret.ini, but I don't think that is particularly problematic
Ben West
@bewest
Sep 25 2015 02:04
I'll try to see if there's a way to optimize it somehow
Chris Oattes
@cjo20
Sep 25 2015 02:13
Ben West
@bewest
Sep 25 2015 02:13
yeah, something like that
also want to see about discovering secrets somehow depending on context
Chris Oattes
@cjo20
Sep 25 2015 02:16
@bewest dependant on the section it is in? or just configurable by other parts of the program?
Chris Oattes
@cjo20
Sep 25 2015 02:25
updated the gist with something that should do the latter
Ben West
@bewest
Sep 25 2015 02:55
each "section" is usually a single "high level object"
Ben West
@bewest
Sep 25 2015 03:03
what sections are allowed to include stuff?
Chris Oattes
@cjo20
Sep 25 2015 03:04
what do you mean?
Ben West
@bewest
Sep 25 2015 03:04
include = pump-secrets.ini
the include would apply to that section only
hmmm, doesn't do the filtering you have that way though :-(
Chris Oattes
@cjo20
Sep 25 2015 03:05
I didn't see anything that mentioned a way of doing includes automatically, and the problem would be redirecting things
Ben West
@bewest
Sep 25 2015 03:05
I'm thinking about how to handle api-secrets/tokens for plugins
maybe all devices get their own ini
Chris Oattes
@cjo20
Sep 25 2015 03:07
what would be the benefit of having multiple ini files?
Ben West
@bewest
Sep 25 2015 03:07
that's what you're doing
Chris Oattes
@cjo20
Sep 25 2015 03:07
*multiple secret ini files
Ben West
@bewest
Sep 25 2015 03:07
you can keep some of them with secrets out of git
nightscout, dexcom share and others all require credentials which are secret
I can't predict all future secrets
so allowing each "device" to have it's own extended config
that solves the problem
all the secret stuff goes in the extended config in it's own ini
so when you add pump with serial, it'll add ini with same name
and I'll use like a configured = pump.ini or whatever instead of the extra params
it's really there for "extra/extended" params anyway
Chris Oattes
@cjo20
Sep 25 2015 03:09
did you see this? cjo20/openaps@6ed7f0e
Ben West
@bewest
Sep 25 2015 03:10
yeah
well
no
I have now
Chris Oattes
@cjo20
Sep 25 2015 03:10
:P
Ben West
@bewest
Sep 25 2015 03:10
thing is other plugins get added
eg web api device gets added
and it needs api credentials
eg the 'share bridge
scott and dana log into dexcom share server with user/pass
Chris Oattes
@cjo20
Sep 25 2015 03:11
the plugins would have access to the Config class though, wouldn't they?
Ben West
@bewest
Sep 25 2015 03:11
hmm, but it would depend on carefully knowing and adding just right I think
each device already has list of extended settings
might as well put them all in config specific to extended settings for that device
then it's easy to keep it out of git
Chris Oattes
@cjo20
Sep 25 2015 03:12
if they're able to access the config data, they can call .add_secret. And you could change it from just being a list of strings to being a dictionary of lists, where the key in the dictionary is the section type
and the value is a list of things you want kept secret for that section type
Ben West
@bewest
Sep 25 2015 03:13
will play around with it
I think it makes it bit harder to write the plugin
if it's a simple bash script or something...
or node js script
it can't change the python config object
Chris Oattes
@cjo20
Sep 25 2015 03:13
are they able to access the python config object at all?
Ben West
@bewest
Sep 25 2015 03:13
if it's not written in python, then no
or if someone writes a new python tool, they might call it out of order or something
I'll play with it
Scott Leibrand
@scottleibrand
Sep 25 2015 03:16
We currently use environment variables for our secrets, set in crontab.
Not ideal, but it works.
Chris Oattes
@cjo20
Sep 25 2015 03:19
the problem with other languages and the config file, is that they're going to have to work out how to parse / write it anyway if they want to use the .ini
Ben West
@bewest
Sep 25 2015 03:20
the way it works is arguments get curried
so the prefixen generally get saved in the ini
then added in order to some command
so the node/ruby/bash command gets called with the arguments saved prior
not sure, maybe environment variables are enough/better
Chris Oattes
@cjo20
Sep 25 2015 03:23
I imagine that gets quite messy when you start having things you can have multiple instances of each requiring a secret
Ben West
@bewest
Sep 25 2015 03:23
that would be more weight towards separate INI for each
if it's environment variable it can be defined in cron job or /etc/default/openaps or something
if environment variable is no good, it has to be passed as arg somewhere
so it has to go in INI somewhere
preferably not in git
Chris Oattes
@cjo20
Sep 25 2015 03:30
I don't know if its because its 0430, but I'm not understanding what the scenario is that it's a problem to have only 2 .inis, a 'public' one and a 'secret' one
Ben West
@bewest
Sep 25 2015 03:33
not a problem, per se
Chris Oattes
@cjo20
Sep 25 2015 03:33
if something is written in python, it uses the config object from openaps, and it's fine. There may be a requirement for people developing plugins to prefix their section names with the name of the plugin to prevent naming conflicts, but the worst that happens if there is a conflict is that a setting ends up not being stored in git, which probably isn't the end of the world
Ben West
@bewest
Sep 25 2015 03:34
I'm thinking of someting like openaps device add ns-api process --require use ns-api-cli https://api-secret@my-url
would love to coordinate that with eg
err
This message was deleted
hah
with device.extend_config(...)
that way if a device has an extended config, like medtronic, like ns-api, like dexcom-share
it's probably sensitive....
and is stored separarely, universally
then I can run openaps use ns-api shell insert pump-history.json or similar
err
and the api secret and endpoint are stored securely outside git, and not in openaps.ini
Chris Oattes
@cjo20
Sep 25 2015 03:40
Anything using python can use the secret mechanism to keep things out of git. Any plugins not using python could have their own config file that is outside of the python system, and the author can recommend that those files get ignored too, but I don't think you need to start having "include"s in the .ini for that
Ben West
@bewest
Sep 25 2015 03:40
basically I want to handle this in set_config somehow from device interface
where set_config interacts with the config
great place to pass message to config that this thing is secret
all the extra "device" config should be secret
add_secret_option
Chris Oattes
@cjo20
Sep 25 2015 03:43
or as a flag in add_option
Chris Oattes
@cjo20
Sep 25 2015 03:43
def add_option (self, k, v, secret=false):
Ben West
@bewest
Sep 25 2015 03:44
def set_config (args, device):
  device.add_option('serial', args.serial)
def set_config (args, device):
  device.add_option('cmd', args.cmd)
  device.add_option('args', ' '.join(args.args))
  device.add_option('fields', ' '.join(args.require))
second one is how process does it
fields should also be secret
this is where device "advertises" to openaps what options it needs
so makes sense to associate secret options with secretive status
seems simpler to me almost
to say "if the device as extended/secret config, put it in configured additional ini per device"
that way ns api doesn't need to know about medtronic's serial number
we already have one additional ini
Chris Oattes
@cjo20
Sep 25 2015 03:46
is it a problem if ns api does know about medtronic's serial number?
Ben West
@bewest
Sep 25 2015 03:46
probably not, but it keeps design doors open for future
dunno
it'd be nice if fields there could also go in secret
seems like nice place for message passing of some sort
device.add_extension('fields', ' '.join(args.require))
device.add_extension('serial', args.serial)
I'm thinking for the $n ini, that openaps device add --ini ../outside/foo.ini new-device process --require operation $somesecret $someothersecret would put something like extended = ../outside/foo.ini in openaps.ini
so it'd be super easy to keep it out of git because the path itself is optional
so openaps device add --ini ../secret.ini pump medtronic 665455
the serial = 665455 line would be in ../secret.ini
Chris Oattes
@cjo20
Sep 25 2015 03:51
that starts making command lines more complicated, with potentially a ton of .ini files each with one line in them
Ben West
@bewest
Sep 25 2015 03:51
openaps device add pump medtronic 665455 would put it in ./pump.ini
no, only thin would be extended config
that's it
the serial
the things explicitly marked
so it'd only be the ones with extended configs
like medtronic, and ... some future things
also you could re-use same secret.ini
Ben West
@bewest
Sep 25 2015 03:58
hmmm yeah
Chris Oattes
@cjo20
Sep 25 2015 03:58
which keeps most of the mechanics of it hidden from the user
the person developing the plugin should have a good idea of what should be private
Ben West
@bewest
Sep 25 2015 04:00
yeah
treat secret_fields like required
maybe rename secret_fields to sensitive or private
Chris Oattes
@cjo20
Sep 25 2015 04:00
yeah
I've not coded it up properly, I'll probably do that tomorrow
well, today now
Ben West
@bewest
Sep 25 2015 04:01
:-)
Chris Oattes
@cjo20
Sep 25 2015 04:01
I definitely wasn't going to stay up until 5am again today
yet...
leave me a message if you have any more ideas about this :) I'm going to try and get some sleep before I have to get up for work
Ben West
@bewest
Sep 25 2015 04:03
yeah get some sleep
:-)
:zzz:
Chris Oattes
@cjo20
Sep 25 2015 04:03
g'night :)
Ben West
@bewest
Sep 25 2015 04:09
interesting, openaps/diyps assigned a a temp rate, but it's still a big high and I want it down faster
but don't want to over do it
so doing against the eventualBG, which is lower, figure the temp will do it's thing and that "iob" minus the correction would be the remainder I want
Scott Leibrand
@scottleibrand
Sep 25 2015 04:10
Sorry, more details pls?
paste of determine-basal stderr output would be ideal
determine-basal://text/shell/02.diyps/diyps.json
IOB: 0.81, Bolus IOB: 0.00
Avg. Delta: -1.0, BGI: -1.7
15m deviation: 2
BG: 195-4 -> 160-160 (Unadjusted: 158-158)
reporting 02.diyps/diyps.json
{
  "temp": "absolute",
  "bg": 195,
  "tick": -4,
  "eventualBG": 160,
  "snoozeBG": 160,
  "duration": 0,
  "rate": 0,
  "reason": "-4 and 160"
}
160 is still bit high
I plan on eating, so would rather dose down closer to lower end of target
Scott Leibrand
@scottleibrand
Sep 25 2015 04:13
Ah, yeah it doesn't know you're eating soon.
The -4 delta is why it canceled the high temp.
Ben West
@bewest
Sep 25 2015 04:14
right, I didn't actually send that to my pump
so I'm still high temped
Scott Leibrand
@scottleibrand
Sep 25 2015 04:14
You should be dropping about -2 / 5m
Ben West
@bewest
Sep 25 2015 04:14
might cut it by half though I guess
or kill it in another 10 minutes
Scott Leibrand
@scottleibrand
Sep 25 2015 04:14
how high is the temp it has running?
Ben West
@bewest
Sep 25 2015 04:14
max, 2.0
Scott Leibrand
@scottleibrand
Sep 25 2015 04:15
and how much would you prebolus?
Ben West
@bewest
Sep 25 2015 04:15
bolused 0.7
Scott Leibrand
@scottleibrand
Sep 25 2015 04:15
re-run and see what it thinks of that. :-)
Ben West
@bewest
Sep 25 2015 04:15
yeah
was just thinking
temp will run for another 18 minutes
Scott Leibrand
@scottleibrand
Sep 25 2015 04:16
bolus IOB will make it more interesting.
Ben West
@bewest
Sep 25 2015 04:16
yeah, figured I'd have some fun with it today
earlier
I need to cut basal I think
in early afternoon
no bolus all morning
just basal...
started step looping around 4:15
was too fast/furious
so at the bottom there I ate 53 grams of carbs :-)
Scott Leibrand
@scottleibrand
Sep 25 2015 04:17
You should set up ns-upload and turn on IOB pill
Ben West
@bewest
Sep 25 2015 04:17
ok eventual bg 122
Scott Leibrand
@scottleibrand
Sep 25 2015 04:17
oh wait, that doesn't work yet without DIYPS, nm
Ben West
@bewest
Sep 25 2015 04:17
going to go ahead and cancel that temp
Scott Leibrand
@scottleibrand
Sep 25 2015 04:18
Is it still trying to cancel it too?
Ben West
@bewest
Sep 25 2015 04:18
yeah
just sending what it recommended
bewest@hither:~/Documents/foo$ openaps gather-and-predict && json    -f 02.diyps/diyps.json
curl://text/shell/monitor/glucose.json
reporting monitor/glucose.json
pump://JSON/read_clock/monitor/clock.json
reporting monitor/clock.json
pump://JSON/iter_pump/monitor/pump-history.json
reporting monitor/pump-history.json
pump://JSON/read_selected_basal_profile/monitor/active-basal-profile.json
reporting monitor/active-basal-profile.json
pump://JSON/reservoir/monitor/reservoir.json
reporting monitor/reservoir.json
pump://JSON/settings/monitor/settings.json
reporting monitor/settings.json
pump://JSON/read_status/monitor/status.json
reporting monitor/status.json
pump://JSON/read_battery_status/monitor/battery.json
reporting monitor/battery.json
pump://JSON/read_bg_targets/monitor/bg-targets.json
reporting monitor/bg-targets.json
pump://JSON/read_insulin_sensitivies/monitor/insulin-sensitivities.json
reporting monitor/insulin-sensitivities.json
pump://JSON/read_carb_ratios/monitor/carb-ratios.json
reporting monitor/carb-ratios.json
pump://JSON/read_temp_basal/monitor/temp-basal-status.json
reporting monitor/temp-basal-status.json
munge://JSON/clean/01.prepare/mm-cleaned.json
reporting 01.prepare/mm-cleaned.json
munge://JSON/reconcile/01.prepare/mm-reconciled.json
reporting 01.prepare/mm-reconciled.json
munge://JSON/resolve/01.prepare/mm-resolved.json
reporting 01.prepare/mm-resolved.json
munge://JSON/normalize/01.prepare/mm-normalized.json
reporting 01.prepare/mm-normalized.json
get-profile://text/shell/02.diyps/profile.json
reporting 02.diyps/profile.json
calculate-iob://text/shell/02.diyps/iob.json
reporting 02.diyps/iob.json
determine-basal://text/shell/02.diyps/diyps.json
IOB: 1.61, Bolus IOB: 0.70
Avg. Delta: -1.7, BGI: -2.0
15m deviation: 1
BG: 193-2 -> 122-153 (Unadjusted: 121-152)
reporting 02.diyps/diyps.json
{ 
  "temp": "absolute",
  "bg": 193,
  "tick": -2,
  "eventualBG": 122,
  "snoozeBG": 153,
  "duration": 0,
  "rate": 0,
  "reason": "-2 and 122"
}
bewest@hither:~/Documents/foo$ openaps use pump  set_temp_basal 02.diyps/diyps.json
{ 
  "bg": 193,
  "temp": "absolute",
  "snoozeBG": 153,
  "recieved": true,
  "reason": "-2 and 122",
  "rate": 0.0,
  "eventualBG": 122,
  "timestamp": "2015-09-24T21:17:40.644606",
  "duration": 0,
  "tick": -2
}bewest@hither:~/Documents/foo$
synced with openaps-example
Scott Leibrand
@scottleibrand
Sep 25 2015 04:21
You might want to run openap-js dev branch.
Better logic comparing delta to BGI/2 instead of 0.
Ben West
@bewest
Sep 25 2015 04:23
interesting
on dev
$ openaps analyze-and-predict 
munge://JSON/clean/01.prepare/mm-cleaned.json
reporting 01.prepare/mm-cleaned.json
munge://JSON/reconcile/01.prepare/mm-reconciled.json
reporting 01.prepare/mm-reconciled.json
munge://JSON/resolve/01.prepare/mm-resolved.json
reporting 01.prepare/mm-resolved.json
munge://JSON/normalize/01.prepare/mm-normalized.json
reporting 01.prepare/mm-normalized.json
get-profile://text/shell/02.diyps/profile.json
reporting 02.diyps/profile.json
calculate-iob://text/shell/02.diyps/iob.json
reporting 02.diyps/iob.json
determine-basal://text/shell/02.diyps/diyps.json
{"delta":-2,"glucose":193,"avgdelta":-1.6666666666666665}
{"duration":16,"rate":2,"temp":"absolute"}
{"iob":1.5046812853697713,"activity":0.008714916363636364,"bolusiob":0.6748951403772839}
{"carbs_hr":28,"max_iob":0,"dia":4,"type":"current","current_basal":0.9,"max_daily_basal":1.1,"max_basal":2,"max_bg":125,"min_bg":106,"carbratio":15,"sens":45}
IOB: 1.50, Bolus IOB: 0.67
Avg. Delta: -1.7, BGI: -2.0
15m deviation: 1
BG: 193-2 -> 126-172 (Unadjusted: 125-171)
reporting 02.diyps/diyps.json
I'm going to need more pis
Scott Leibrand
@scottleibrand
Sep 25 2015 04:25
:-)
Ben West
@bewest
Sep 25 2015 08:58
ok
I've got a bed-side rig going I think
we'll see
I've had lot of indian food, rich curry, samosas
definitely only going up tonight
Chris Oattes
@cjo20
Sep 25 2015 08:59
morning
Ben West
@bewest
Sep 25 2015 08:59
no way
morning
but...
was just like 3 hours ago we spoke
I'm going to bed my friend
Chris Oattes
@cjo20
Sep 25 2015 08:59
heh. Closer to 5
time for work for me :P
sleep well
Ben West
@bewest
Sep 25 2015 08:59
lol
first over night!
Chris Oattes
@cjo20
Sep 25 2015 09:06
Good luck
Scott Leibrand
@scottleibrand
Sep 25 2015 12:03
@bewest awesome! Did you use it at all overnight?
Dana Lewis
@danamlewis
Sep 25 2015 12:04
I don't think he'd be awake yet :)
Btw if anyone ever does a red eye, the long cables to the Carelink are excellent so you can leave the Pi and battery under your feet on just run the Carelink up into your pocket near the pump.
Scott Leibrand
@scottleibrand
Sep 25 2015 12:05
Hope he's not still up, but you never know. ;-)
Nice
Things you learn when you fry a Pi. ;-)
Dana Lewis
@danamlewis
Sep 25 2015 12:06
I guess that applies to normal long flights, too. I've just kept the whole thing on my lap for most <5 hour flight but will probably travel with this longer cable now for long flights
yea yea learn something every time ;)
Chris Oattes
@cjo20
Sep 25 2015 12:17
danamlewis: so just a long usb cable?
Scott Leibrand
@scottleibrand
Sep 25 2015 12:31
Yeah, USB male to female extension.
Scott Leibrand
@scottleibrand
Sep 25 2015 13:32
@bewest you really should replace your sensor after 2 weeks when it starts to ??? if you're going to run OpenAPS.
Ali Mazaheri
@amazaheri
Sep 25 2015 14:41
he is still sleeping
lets call him and wake him up :)
Had a bad battery issue at 3am
luckily I was up but still had to recover a 55 ugh! :)
gotta love push notification to keep u on the edge
Chris Oattes
@cjo20
Sep 25 2015 14:45
his NS hasn't updated in a while
Dana Lewis
@danamlewis
Sep 25 2015 14:45
Oh good yall are watching him?
Ali Mazaheri
@amazaheri
Sep 25 2015 14:46
guardians of galaxy

openAPSunicorns

Chris Oattes
@cjo20
Sep 25 2015 14:46
thought I'd take a look. It's been over an hour since last CGM data though
closer to 3 hours
Ali Mazaheri
@amazaheri
Sep 25 2015 15:04
:+1:
Ali Mazaheri
@amazaheri
Sep 25 2015 15:28
Capture.PNG
Batt failure at 1:30am note to self! Check battery everynight!
Dana Lewis
@danamlewis
Sep 25 2015 15:43
Yep. Ps @channemann Not seeing a variance in battery with Duracell vs energizer triple a
diabeticgonewild
@diabeticgonewild
Sep 25 2015 15:54
I can't merge my old github with the updated OpenAPS docs repo...I have been trying for at least 2 hours and it's mind numbing.
This message was deleted
Dana Lewis
@danamlewis
Sep 25 2015 15:55
Which way are you trying to merge? Are you trying to fork the updated version, or pr in some changes?
diabeticgonewild
@diabeticgonewild
Sep 25 2015 15:55
Fork the updated version
But it's "afraid" of me removing some of my (irrelevant) edits
Chris Oattes
@cjo20
Sep 25 2015 15:56
do you need to stash/ commit / revert changes you've made locally?
diabeticgonewild
@diabeticgonewild
Sep 25 2015 15:56
Basically yeah cause local repo is openAPS/docs
but I can't because it's "afraid" of me losing edits on a couple of files
on the remote repo
Chris Oattes
@cjo20
Sep 25 2015 15:57
paste the error message
diabeticgonewild
@diabeticgonewild
Sep 25 2015 15:58
there's all sorts. I am winging it.
Chris Oattes
@cjo20
Sep 25 2015 15:58
try and do what you originally set out to do, and then paste the error message
diabeticgonewild
@diabeticgonewild
Sep 25 2015 15:59
K
Can't copy and paste in git bash....the "recommended" alternative command does something crazy.
screenshot.jpg
git commit -a puts me through some sort of text editor that I can't get out of or scroll through for the life of me.
Chris Oattes
@cjo20
Sep 25 2015 16:03
do you want to keep the changes you have made to hardware.md?
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:03
on the local repo, yes, not on the remote
Chris Oattes
@cjo20
Sep 25 2015 16:05
put the contents of hardware.md on gist or similar
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:06
Thanks. I also have to clone the repo again under a different name cause the directory name is the same as the repo name. N00b mistake.
Stupid question. I deleted both hardware.md in my local and remote repos cause I got desperate. It belongs in the local one right?
Chris Oattes
@cjo20
Sep 25 2015 16:08
what do you mean?
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:09
That file doesn't exist in either the local or the remote repos cause I was told to delete them on the command line.
or delete it
not them
Chris Oattes
@cjo20
Sep 25 2015 16:11
so you want to keep the changes you've made, even though you've already deleted the file from your computer?
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:11
Yeah I guess, so I need to reclone the repo.
I mean, it's not clear what is going on. It just tells me to delete crap off of a repo not specifiying remote or local and I end up guessing and playing around forever.
Chris Oattes
@cjo20
Sep 25 2015 16:15
So you're trying to update diabeticgonewild/docs to have all of the changes from openaps/docs, as well as your own changes.
  1. clone diabeticgonewild/docs to your hard drive
  2. git fetch upstream
  3. git checkout master
  4. git merge upstream/master
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:15
Screenshot2.jpg
Is upstream syntax for something?
Sorry I'm so clueless
It sounds lke it to me.
Chris Oattes
@cjo20
Sep 25 2015 16:17
upstream should refer to where you forked from
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:17
should that be https:// or can it be local?
I guess https://
Chris Oattes
@cjo20
Sep 25 2015 16:18
you type in what I typed
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:18
So you type in upstream?
Chris Oattes
@cjo20
Sep 25 2015 16:18
yes
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:18
This sounds so wrong to me... but I guess its not
``` git checkout master
docs/Hardware/hardware.md: needs merge
error: you need to resolve your current index first.
Chris Oattes
@cjo20
Sep 25 2015 16:20
ok, copy and paste docs/Hardware/hardware.md in to gist
Chris Oattes
@cjo20
Sep 25 2015 16:22
what is the output of git status
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:23
no changes added to commit (use git add and/or git commit -a)
Chris Oattes
@cjo20
Sep 25 2015 16:24
is that all of the output?
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:24
oh and a ton of random files in red in that random directory
screenshot3.jpg
It says unmerged paths
and unmerged files
Screenshot4.jpg
I need to take a break from this. I have been in front of the computer for at least 4 hours now. Sorry.
Chris Oattes
@cjo20
Sep 25 2015 16:28
try running git rebase --abort
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:29
then try running the previous commands?
Chris Oattes
@cjo20
Sep 25 2015 16:29
yes
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:29
like #2
Your branch is up-to-date with origin/master
for git checkout master
Chris Oattes
@cjo20
Sep 25 2015 16:30
now run the 3rd command
err, 4th
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:31
should I be putting in master? (that is the 3rd command)
Chris Oattes
@cjo20
Sep 25 2015 16:31
run git merge upstream/master
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:32
Automatic merge failed; fix conflicts and then commit the result
Chris Oattes
@cjo20
Sep 25 2015 16:32
ok, now put the contents of hardware.md in to gist
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:33
I did.
Chris Oattes
@cjo20
Sep 25 2015 16:33
I mean again
a new copy
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:33
This doesn't make a ton of sense to me with respect to syntax and the machine knowing what it is/what it isn't
K
add file or create gist?
I just noticed the add file...so I screwed up before
Chris Oattes
@cjo20
Sep 25 2015 16:35
create gist
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:35
alright
done
Chris Oattes
@cjo20
Sep 25 2015 16:35
link?
Chris Oattes
@cjo20
Sep 25 2015 16:37
what has happened so far is you've downloaded the information about openaps/docs, made sure your hard drive has the latest diabeticgonewild/docs, and then told it to try and merge openaps/docs with what you have on your computer. It tried it's best, but it wasn't able to work out how to reconcile some of the differences between the openaps version and the diabeticgonewild version, so it is asking you to fix those yourself
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:37
K Thanks
Chris Oattes
@cjo20
Sep 25 2015 16:38
I would have expected the file to contain some sections that look like
some text here
<<<<<<< HEAD
a
=======
b
>>>>>>>
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:38
It's in *.md format
I copied and pasted it

is HEAD

### is head
## is Head too
Chris Oattes
@cjo20
Sep 25 2015 16:39
what did you open it in to do the copy and paste?
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:39
Sublime Text 2
Chris Oattes
@cjo20
Sep 25 2015 16:40
can you show me the output of git status again?
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:42
ss1.jpg
ss2.jpg
Chris Oattes
@cjo20
Sep 25 2015 16:43
what is the full filename of the file you uploaded to gist? including the path
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:43
hardware.md
probably made a mistake
Chris Oattes
@cjo20
Sep 25 2015 16:43
that isn't including the path
the path is which directory it is in
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:44
didn't include the path in the filename
on the gist
Chris Oattes
@cjo20
Sep 25 2015 16:44
I know. But it is in a directory on your computer. What is that directory?
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:44
oh
repo = docs_final/docs/Hardware/hardware.md
not logical but you get the point
repo = docs_final
Chris Oattes
@cjo20
Sep 25 2015 16:48
what happens if you run git mergetool
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:52
sss3.jpg
That's the output of git commit -a
I don't know what to do with that
Chris Oattes
@cjo20
Sep 25 2015 16:52
you cant commit because you still have conflicts
you need to resolve the conflicts before you can commit
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:53
So when I ran git mergetool it basically ran tortisesvn inside of git bash and it wouldn't merge on its own
tortoisesvn
Chris Oattes
@cjo20
Sep 25 2015 16:54
what is the output of git diff docs/Hardware/hardware.md
diabeticgonewild
@diabeticgonewild
Sep 25 2015 16:55
sss4.jpg
Look I really, really appreciate your help but I didn't sleep at all last night. I really need to rest a little bit. My body can't handle the lack of sleep like it used to. Too many autoimmune diseases. But thank you so much. I think I am going to call it quits for awhile but you made it more bearable.
Thanks
Chris Oattes
@cjo20
Sep 25 2015 16:56
seeya
Chris Hannemann
@channemann
Sep 25 2015 17:06
@danamlewis re: AAA batteries good to know. How long are you getting right now on yours? I switched back to an alkaline this go-around after the last lithium died (rapidly) overnight. Still keeping my results here.
Chris Hannemann
@channemann
Sep 25 2015 17:57
Does anyone happen to know what the cutoff voltage is for Medtronic pumps (i.e. when does it go into 'Low Battery' state)?
diabeticgonewild
@diabeticgonewild
Sep 25 2015 18:00
You could just pull out a multimeter next time the battery goes low on your pump. But I don't like putting 2 metal leads onto two terminals of a battery, even if they complete a circuit (with the internal resistance of the multimeter that exists) cause batteries put out a lot of current.
It it got shorted it's dangerous and cause it puts out a lot of current you can end up with like burns on your face.
I prefer battery testers
Chris Oattes
@cjo20
Sep 25 2015 18:02
from a AAA battery?
Internal resistance of a multimeter is probably at least 10k, if not 100k. A dead battery is probably somewhere from, what, 1-1.2v? 1.2v/10k = 0.12mA
diabeticgonewild
@diabeticgonewild
Sep 25 2015 18:03
That's true but it makes me uncomfortable
Chris Hannemann
@channemann
Sep 25 2015 18:04
Not sure if pump is reporting Voc or V at some load
But I'll test my next dead one
diabeticgonewild
@diabeticgonewild
Sep 25 2015 18:04
There are stories of people putting screwdriver tips to the terminals and getting horrific burns on their faces from the burns from basic batteries
From sparks
Chris Oattes
@cjo20
Sep 25 2015 18:05
seeing as it is being used in the pump at the time, it'll be under the load of running processing / comms / perhaps the motor running
diabeticgonewild
@diabeticgonewild
Sep 25 2015 18:05
Cause of the current. They can deliver a lot of current. That's the danger of batteries.
Chris Oattes
@cjo20
Sep 25 2015 18:05
from a AAA battery?
maybe from a car battery
Scott Leibrand
@scottleibrand
Sep 25 2015 18:12
If you put about a half dozen 9V's in series and touch the wires, you might be in danger (it is certainly shocking, as I discovered as a teenager). But a single battery you can carry with two fingers is not capable of damaging you through your skin. And even a larger one is not going to damage a multimeter if it's in proper voltage-measuring mode.
Chris Oattes
@cjo20
Sep 25 2015 18:14
Pretty sure i've licked a 9v battery as a kid :P
scottleibrand @scottleibrand waits eagerly for @bewest to wake up and let us know if he managed to run openaps overnight, and if so how it went.
Scott Leibrand
@scottleibrand
Sep 25 2015 18:14
That's the proper way to test if they're dead. ;-)
Chris Hannemann
@channemann
Sep 25 2015 18:14
@cjo20 As a kid? I'm licking one right now ...
Chris Oattes
@cjo20
Sep 25 2015 18:14
@channemann I've moved on to eclairs... :P
Chris Hannemann
@channemann
Sep 25 2015 18:14
:tongue: :battery:
Scott Leibrand
@scottleibrand
Sep 25 2015 18:14
:+1:
diabeticgonewild
@diabeticgonewild
Sep 25 2015 18:18
Yes it is. My professors told me.
Scott Leibrand
@scottleibrand
Sep 25 2015 18:18
Not everything you learn in school is relevant in the real world.
diabeticgonewild
@diabeticgonewild
Sep 25 2015 18:19
If you short it by putting a screwdriver shaft on each terminal you can burn your face from the Sparks due to current.
Chris Oattes
@cjo20
Sep 25 2015 18:19
how do you think you are gonig to get sparks from a screwdriver?
Scott Leibrand
@scottleibrand
Sep 25 2015 18:19
In theory, there is no difference between theory and practice. In practice... ;-)
diabeticgonewild
@diabeticgonewild
Sep 25 2015 18:19
Current
from the short
True. Even though engineering is applied science...
Scott Leibrand
@scottleibrand
Sep 25 2015 18:23
Just learned that was a quote from the late Yogi Berra
Chris Oattes
@cjo20
Sep 25 2015 18:25
I'm doubtful. Internal resistance of a battery is probably, what, 300mohm, which would limit the current to 5 amps if it was a 0 ohm short. A 5mm copper conductor (I know, screwdrviers aren't copper) can carry about 100amps (4 AWG wire is rated for that much)
so I think even in theory, getting sparks from a AAA battery that would cause damage to your face is unlikely
Ali Mazaheri
@amazaheri
Sep 25 2015 18:34
paging @bewest
Ben West
@bewest
Sep 25 2015 18:37
my f** sensor failed like 6am
Ali Mazaheri
@amazaheri
Sep 25 2015 18:37
k
Ben West
@bewest
Sep 25 2015 18:37
trying to figure it out myself
Ali Mazaheri
@amazaheri
Sep 25 2015 18:37
I was worried
Ben West
@bewest
Sep 25 2015 18:37
looks like it was doing stuff till then
Ali Mazaheri
@amazaheri
Sep 25 2015 18:37
yeah :) u were being watched by couple of us
Ben West
@bewest
Sep 25 2015 18:38
got it set up around 2 finally
went to bed with it
looks I promptly rolled on sensor
Ali Mazaheri
@amazaheri
Sep 25 2015 18:39
:fire:
Ben West
@bewest
Sep 25 2015 18:39
actually it failed twice
because I tried restarting sensor around 8am
Scott Leibrand
@scottleibrand
Sep 25 2015 18:39
You need a new sensor. They get flaky with age.
Ben West
@bewest
Sep 25 2015 18:39
well, openaps needs a new sensor
I'm fine with this one
Ali Mazaheri
@amazaheri
Sep 25 2015 18:39
lol
Ben West
@bewest
Sep 25 2015 18:40
lol
Ali Mazaheri
@amazaheri
Sep 25 2015 18:40
Sleeze Ben Wee
Scott Leibrand
@scottleibrand
Sep 25 2015 18:40
And they're jumpy when new, so it's ideal if you have two transmitters and receivers so you can overlap them.
Ben West
@bewest
Sep 25 2015 18:40
yeah, might be pre-install one for a soak
Ali Mazaheri
@amazaheri
Sep 25 2015 18:40
or u be banned from this room :)
Ben West
@bewest
Sep 25 2015 18:40
let it soak for awhile
Scott Leibrand
@scottleibrand
Sep 25 2015 18:40
No, you need a new sensor too. You could do way better with better data.
Taking off: laters.
Ali Mazaheri
@amazaheri
Sep 25 2015 18:40
safe travels
Ben West
@bewest
Sep 25 2015 19:04
also, looks like it was getting glucose only
stick wasn't working
heh
Scott Leibrand
@scottleibrand
Sep 25 2015 19:32
Bummer
Ben West
@bewest
Sep 25 2015 19:44
also, I didn't have node installed
also I didn't have mmhistory upto date
also openaps-js not installed
Scott Leibrand
@scottleibrand
Sep 25 2015 19:45
Details details...
Ben West
@bewest
Sep 25 2015 19:46
ironically placing the stick is not an issue
I have this "boom" architect lamp at bedside, where the lamp shade is mounted on a flexible, extended arm
taped it to cool part of lamp shade, and it's pointed 45 degrees down at my hips naturally, about 3 feet away
going to need another sensor install
this one is too flaky
won't even survive warmup now
Ben West
@bewest
Sep 25 2015 21:15
lots of issues
make sure pi is in right timezone ;-)
sudo dpkg-reconfigure tzdata
Scott Leibrand
@scottleibrand
Sep 25 2015 21:16
is that in the docs?
Ben West
@bewest
Sep 25 2015 21:16
don't think so
also there's an issue wit logging should get fixed
Scott Leibrand
@scottleibrand
Sep 25 2015 21:17
yeah, I have a really aggressive logrotate in openaps-js to work around that
Ben West
@bewest
Sep 25 2015 21:17
it's duplicated the logging instance for each report running in the same process
so if you run 4 reports in same proc
decocare logs 4x duplicate output in last report...
eesh
makes rsysog unhappy
but decocare can also provide way to filter out logs better
doesn't need to log so much stuff for this....
logging should be reset between runs I guess
but eliminating the duplicate logging should go long way
great, finally ran correctly
need to make a reset_session use for medtronic
to easily wipe out the expires line
it matters a lot if you switch timezones
maybe also introduce timezone data to the expires line
Scott Leibrand
@scottleibrand
Sep 25 2015 21:19
heh ya
Ben West
@bewest
Sep 25 2015 21:20
I was wondering why it wasn't working
finally noticed no power sent
think maybe minutes should be set to 2 or something
10 minutes puts it on edge of second loop :-(
that's bad
need option to try less hard, also
early on, it was important to keep trying all kinds of different things
but it appears now like if it doesn't work it never will...
so why bother keep trying
yeah, second/third loop just failed
because session expires right after it starts
maybe run every 6 minutes instead so the loops line up better
Ali Mazaheri
@amazaheri
Sep 25 2015 21:23
Ben
U changed that in Dev for me
in Florida
Ben West
@bewest
Sep 25 2015 21:23
yeah, I have that
Ali Mazaheri
@amazaheri
Sep 25 2015 21:23
related?
Ben West
@bewest
Sep 25 2015 21:23
sort of
going to try minutes = 28 for session length
will hopefully overlap the loops in the middle instead of at the edges
Ben West
@bewest
Sep 25 2015 21:29
that patch from florida is in master now :-)
Ali Mazaheri
@amazaheri
Sep 25 2015 21:29
cool
18 or newer version
19?
Ben West
@bewest
Sep 25 2015 21:30
in openaps 0.0.6 I think
Ali Mazaheri
@amazaheri
Sep 25 2015 21:30
cool
Ali Mazaheri
@amazaheri
Sep 25 2015 21:35
u published it?
people get it with -ZU?
Chris Hannemann
@channemann
Sep 25 2015 21:41
@bewest regarding timezone, I remember we had to straighten that out on my system. In the docs right now, there is a section that says:
"Run sudo raspi-config to expand filesystem, change user password and set timezone (in internationalization options). This will take effect on the next reboot, so it is a good idea to go ahead and run sudo reboot now."
Some elaboration would be very helpful :-)
Not sure which is preferable, that method or sudo dpkg-reconfigure tzdata
John Males
@johnmales
Sep 25 2015 22:01
Hi all, gradually working through setting up reports and aliases with the generous help of @eszcloud - thanks :) In the meantime I was wondering is anyone here using the carelink - Pi combination to upload to Nightscout? I have a working Nightscout setup with an Android phone but would like to use the Carelink setup to send data to Nightscout.
Chris Oattes
@cjo20
Sep 25 2015 22:02
I've got a pi and a carelink, I'm not sure which tool I'm meant to use to get it to NS though
I imagine @bewest has a magic openaps incantation to do it
John Males
@johnmales
Sep 25 2015 22:06
Yep looking to see if someone has created an uploader to send data from the Pi - carelink to the MongoDB instance that I have setup for NS.
Ben West
@bewest
Sep 25 2015 22:08
the easiest way is to use the NS rest api
avoid interacting with mongo directly
John Males
@johnmales
Sep 25 2015 22:11
Is there a document about that somewhere by any chance?
docs are self-hosted/hot/live
to post, you need the API-SECRET header set to the hashed value of the plain text secret
post to /api/v1/entries.json
set content-type: application/json
Scott Leibrand
@scottleibrand
Sep 25 2015 22:13
there is a script to do that in openaps-js
Ben West
@bewest
Sep 25 2015 22:14
right, should be able to add that script as openaps device/use/report
Ben West
@bewest
Sep 25 2015 22:15
set API_SECRET="<hashed-secret>" in cron job
Scott Leibrand
@scottleibrand
Sep 25 2015 22:15
and NIGHTSCOUT_HOST too
Ben West
@bewest
Sep 25 2015 22:15
ahhhhh much nicer with cleaner logs
openaps/openaps#52
decocare only shows the nasty BAD AILING and other stuff BAD empty CRC?
John Males
@johnmales
Sep 25 2015 22:16
That script is only in the dev branch of open apps-js? Maybe that is why I didn't see it
Ben West
@bewest
Sep 25 2015 22:17
sudo npm install -g 'git://github.com/openaps/openaps-js.git#dev'
Scott Leibrand
@scottleibrand
Sep 25 2015 22:19
yeah, I need to get some stuff merged back into master
next up is new-loop I think
Ben West
@bewest
Sep 25 2015 22:19
would encourage removing loop/breaking it up into actual standalone jobs
John Males
@johnmales
Sep 25 2015 22:21
Brilliant thanks that gives me something to play with and figure out
Scott Leibrand
@scottleibrand
Sep 25 2015 22:21
yeah, probably better to do that than merge it
Ben West
@bewest
Sep 25 2015 22:22
you're going to love this reduced logging
Scott Leibrand
@scottleibrand
Sep 25 2015 22:22
yes. :)
Ben West
@bewest
Sep 25 2015 22:23
whole loop just ran, only logged to bad zero CRC? lines
everything else ran silent/beautiful
this is set to a 3 minute session, too
not sure if that's better for battery
or worse
Scott Leibrand
@scottleibrand
Sep 25 2015 22:24
we run continually, so doesn't seem like it'd really help
Ben West
@bewest
Sep 25 2015 22:24
would be interesting to try a 28 minute session, I don't have patience to try it right now
it goes a long way to schedule the loops such that the session lifecycle on the pump does not exactly match the cron job
Chris Oattes
@cjo20
Sep 25 2015 22:25
what is session length for?
Ben West
@bewest
Sep 25 2015 22:25
the pump needs "permission" sort of to use the extra battery/RF
so, we tell it, "I'm going to talk to you for $x minutes, so stay ON"
and there's probably an effect on battery using this command more vs less often
so coasting by for an hour on same "session" might well use lot less battery than asking for new session every 5 minutes
Chris Hannemann
@channemann
Sep 25 2015 22:27
Interesting, so the session request is what takes battery, not the session duration?
Ben West
@bewest
Sep 25 2015 22:28
it's unclear what the relationship is
Scott Leibrand
@scottleibrand
Sep 25 2015 22:28
yeah, that's worth a try if someone has the patience. :)
Chris Hannemann
@channemann
Sep 25 2015 22:28
Sounds like something I would do ....
Ben West
@bewest
Sep 25 2015 22:28
but I'm willing to hypothesize that there is an interesting relationship
Chris Oattes
@cjo20
Sep 25 2015 22:28
I'd be surprised if the request took up more battery
Chris Hannemann
@channemann
Sep 25 2015 22:28
Oh, but it would take quite some time to determine the drain
Ben West
@bewest
Sep 25 2015 22:28
I think what happens is there is a big capacitor
it's a big bag inside the pump
the rf on command takes longer than other commands
so you can imagine they might be using that time to charge the capapictor and let it drain slowly
such that asking it to charge itself often would be more expensive to battery
Chris Oattes
@cjo20
Sep 25 2015 22:29
using the cap for what?
Ben West
@bewest
Sep 25 2015 22:30
no one knows for sure, again unclear
but would make sense to use that to power the RF
sense the pump motor should ideally be unaffected by whatever else is going on...
John Males
@johnmales
Sep 25 2015 22:30
Do you mean the session timeout set in openaps.ini?
Ben West
@bewest
Sep 25 2015 22:30
you wouldn't want to split power like that
yes, it's an option minutes
there's no command switch for it, but it'll use whatever is in openaps.ini
Chris Hannemann
@channemann
Sep 25 2015 22:30
Isn't that a question that a current/former Medtronic employee snooping in this gitter channel should be able to answer? When is that individual going to come out of the woodwork?
Ben West
@bewest
Sep 25 2015 22:32
IMO, this should be in the user manual....
not sure why we are subject to surprise/guessing for a life-critical machine
here's this machine, btw the RF stuff may or may not change it's behavior, no you can't know how, good luck
ok, quite happy with how the pi is performing now
btw, scott, for your multi-pi setup, what will help is after init'ing your openaps instance, make them use a branch not master
like rpi-home instead of master
so they're on different branches
might well be easier to just git checkout rpi-mobile -- openaps.ini for single file no matter what
and they're all different trees, really
diabeticgonewild
@diabeticgonewild
Sep 25 2015 22:39
IDK if what I said on the openaps/docs PR is me being anal or if it's actually decent recommendations...
I am going to try to pull a copy and push it to a new repo cause I give up on my old repo
I'm going to try to clean it up some...
GitHub can be such a pain to use though, but I am getting to work on it...
Scott Leibrand
@scottleibrand
Sep 25 2015 22:43
is anyone still using openaps-js new-loop instead of master or dev? thinking about just deleting it.
Ben West
@bewest
Sep 25 2015 22:46
I was only looking at dev
what's in new-loop that's not in dev?
dev looks newer
John Males
@johnmales
Sep 25 2015 22:46
No
Ben West
@bewest
Sep 25 2015 22:50
@diabeticgonewild not sure what you mean
if you're working on docs, the best suggestion is to fork the existing docs repo, and propose changes to that
Chris Oattes
@cjo20
Sep 25 2015 22:52
@bewest she has a repo that is a couple of months behind openaps/docs, and she wants to merge the upstream/master repo with hers to get the new stuff
Ben West
@bewest
Sep 25 2015 22:52
ah, that's possible too
Chris Oattes
@cjo20
Sep 25 2015 22:53
but was strugglngi with conflicts
diabeticgonewild
@diabeticgonewild
Sep 25 2015 22:54
Yeah I'm just making a new repo now. I am going to clean it up so this stuff follows a convention, along with added stuff like passwords are not displayed in Linux AKA you won't see the terminal screen move.
I'm still having trouble making it into a new repo but I will figure it out.
Chris Oattes
@cjo20
Sep 25 2015 22:58
https://github.com/openaps/docs go to that page, click "fork" then follow the instructions on that page
if you want to get a completely new copy of the repo with none of your changes in it etc.
diabeticgonewild
@diabeticgonewild
Sep 25 2015 23:05
Got it now.
Once I ditched Git Bash for Windows Command Line things got a lot simpler....crazy
diabeticgonewild
@diabeticgonewild
Sep 25 2015 23:10
At least I can finally work on it now....
Chris Oattes
@cjo20
Sep 25 2015 23:11
it might be a good idea to not leave it so long before syncing with openaps/docs next time
Scott Leibrand
@scottleibrand
Sep 25 2015 23:12
new-loop is behind dev, hence the desire to just delete it.
diabeticgonewild
@diabeticgonewild
Sep 25 2015 23:14
The problem is I am sick a lot and I just don't get around to doing things sometimes, unfortunately.
Chris Oattes
@cjo20
Sep 25 2015 23:23
then all the more reason to make changes in small chunks and then commiting them and making a pull reuqest
that way you don't end up spending a large chunk of what time you have dealing with merging two files that are vastly different
diabeticgonewild
@diabeticgonewild
Sep 25 2015 23:28
Yeah true. It's just hard. So much to do, so little time. Honestly right now I am figuring out how to clean this up by putting all of the text through a screenreader Kurzweil 3000 so I can hear it all in pure spoken English before I make hardcore edits to it. If I listen to it a couple of times I can process it a lot better. It's a very wordy document right now.
Kurzweil 2.jpg
It reads straight from the browser and the Markdown editor I am using is fortunately very accessible. Screenreaders are nice because they don't "express" English well, like a "voice actor" would, so you can cut down on excess and just weird language usage.
diabeticgonewild
@diabeticgonewild
Sep 25 2015 23:38
You guys are actually lucky I don't read much.
Or at least in a traditional way...I rely on accessible text a lot.
diabeticgonewild
@diabeticgonewild
Sep 25 2015 23:44
Also I am removing all italics. You are supposed to replace italics with bold text to improve readability. Already made lots of changes.
Dana Lewis
@danamlewis
Sep 25 2015 23:45
Cool. Looking forward to seeing a PR
diabeticgonewild
@diabeticgonewild
Sep 25 2015 23:46
Thanks!