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

14th
Jan 2016
Tim Howard
@tghoward
Jan 14 2016 00:08
Trying to remember ... what triggers a git commit in openaps? a 'report invoke'? or an alias? or both?
I've seen some discussion about trying to get on git commit per loop. What should I strive for to do that?
*one git commit per loop ^
git log -duh- is showing it committing on reports
Scott Leibrand
@scottleibrand
Jan 14 2016 00:12
IMO that is impossible with current code
*inadvisable, anyway.
Tim Howard
@tghoward
Jan 14 2016 00:13
huh, why inadvisable?
Scott Leibrand
@scottleibrand
Jan 14 2016 00:13
Because you need aliases to be conditional
Tim Howard
@tghoward
Jan 14 2016 00:13
I thought I saw @loudnate talking about this
Scott Leibrand
@scottleibrand
Jan 14 2016 00:13
only way to get down to one commit is to make them a dumb sequence of reports.
Tim Howard
@tghoward
Jan 14 2016 00:14
but the final alias is only one alias ... openaps loop
Nathan Racklyeft
@loudnate
Jan 14 2016 00:14
My loop is a dumb sequence of reports.
Tim Howard
@tghoward
Jan 14 2016 00:15
how do you minimize commits?
Nathan Racklyeft
@loudnate
Jan 14 2016 00:15
All logic and "non-dumb" stuff are openaps plugins,
Aliases.
Tim Howard
@tghoward
Jan 14 2016 00:15
not following ... if it is committing on reports
unless three reports in one alias only has one commit?
Nathan Racklyeft
@loudnate
Jan 14 2016 00:16
openaps update && openaps munge && openaps predict && openaps dose
Scott Leibrand
@scottleibrand
Jan 14 2016 00:16
So you hide all conditionality inside devices?
Nathan Racklyeft
@loudnate
Jan 14 2016 00:16
Generates only 4 commits for 36 repirts
Yeah, uses
Scott Leibrand
@scottleibrand
Jan 14 2016 00:16
still 4 commits though.
Not 1
Nathan Racklyeft
@loudnate
Jan 14 2016 00:16
You could make it one if you wanted
Tim Howard
@tghoward
Jan 14 2016 00:16
4 commits is better than a dozen :smile:
Nathan Racklyeft
@loudnate
Jan 14 2016 00:17
Create a new alias called loop that includes all the content of those 4
Scott Leibrand
@scottleibrand
Jan 14 2016 00:17
@tghoward if an alias calls bash it loses that ability. So Nate is suggesting replacing all our bash aliases with devices.
Which also eliminates a lot of the "build your own logic" elements.
Since all the logic is buried inside scripts that the average person implementing won't change.
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 00:19
iit is something that I don't understand... so openaps commits each time an alias is fired? but WHY? was it the fastest way of logging?
Scott Leibrand
@scottleibrand
Jan 14 2016 00:19
Tradeoffs all around.
Nathan Racklyeft
@loudnate
Jan 14 2016 00:20
[alias]
loop = report invoke read_status.json read_battery_status.json read_clock.json reservoir.json reservoir_history.json recent_glucose.json clean_glucose.json recent_history.json clean_history.json reconcile_history.json resolve_history.json normalize_history.json insulin_effect_without_future_basal.json carb_effect.json cob.json glucose_momentum_effect.json predict_glucose_without_future_basal.json recommended_dose.json set_dose.json clean_history_and_dose.json reconcile_history_and_dose.json resolve_history_and_dose.json normalize_history_and_dose.json insulin_effect_with_dose.json predict_glucose_with_dose.json iob_with_dose.json
Tim Howard
@tghoward
Jan 14 2016 00:20
@ktomy yes, very fast way of logging. And ensuring there is a 'paper trail' to follow
Nathan Racklyeft
@loudnate
Jan 14 2016 00:20
and it only stores diffs. lots of repeated content in each loop.
(git, that is)
Tim Howard
@tghoward
Jan 14 2016 00:23
@scottleibrand @loudnate yes, tradeoffs all around. Thanks for the perspectives. (that's some loop! ^^) Perhaps I can find a middle ground!
Nathan Racklyeft
@loudnate
Jan 14 2016 00:25
sorry to just jump into the convo there. I know there's some confusion around aliases and commits.
some of those reports are just for visualization purposes, not strictly just for dosing. I left out a few others that are tests/validation of different approaches that I run alongside.
Tim Howard
@tghoward
Jan 14 2016 00:27
did @scottleibrand capture it correctly that you are putting a lot of the decisionmaking into openaps devices?
say I want to compare two timestamps with the bash test command ... you'd make a device out of that?
Nathan Racklyeft
@loudnate
Jan 14 2016 00:36
Yeah, everything is registered as a use command in one of four device plugins
Every use has to output something, so in that way every decision is logged
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 00:39
@scottleibrand @bewest Could you please check openaps/oref0#54 ?
Tim Howard
@tghoward
Jan 14 2016 00:41
@loudnate Ok, thanks.
Verterax
@Verterax
Jan 14 2016 01:00
Hello everyone. Happy be here, thank you in advance Scott Leibrand and Dana M. Lewis for your great project, and to everyone of the community that's helped move it forward. I'm on the lookout for a 2.4A firmware or lower MiniMed, but one day, I hope to count our instance of the project amongst the ranks. :)
Dana Lewis
@danamlewis
Jan 14 2016 01:02
:) glad you popped over @Verterax . Hard to track Facebook posts, so you'll tend to get a response here for sure
Scott Leibrand
@scottleibrand
Jan 14 2016 01:28
@TC2013 @bewest I'm getting the same lack of pairing with the other Dexcom share receiver using openxshareble. has anyone done this successfully with a Pi yet?
Scott Leibrand
@scottleibrand
Jan 14 2016 01:36
I would've thought so, based on https://github.com/adafruit/Adafruit_Python_BluefruitLE#linux--raspberry-pi-requirements, but I thought @bewest said he only had it working on Linux...
Thomas Emge
@ThomasEmge
Jan 14 2016 02:56
I have now set the PATH for cron as well as using /usr/local/bin/mm-stick warmup and it is still exiting with 1....arrrggghh
Scott Leibrand
@scottleibrand
Jan 14 2016 02:57
What user are you running command line as? And what user's crontab?
Thomas Emge
@ThomasEmge
Jan 14 2016 02:58
however...if I don't really care that is 1 and keep going it actually succeeds with the rest of the script
I am running as user pi
Scott Leibrand
@scottleibrand
Jan 14 2016 03:00
Is the rest of it working and logging now?
Thomas Emge
@ThomasEmge
Jan 14 2016 03:01
very strange...I don't understand this linux thing ;) yep, logging just fine now
Toby Canning
@TC2013
Jan 14 2016 03:15
@scottleibrand I just tried installing openxshareable on the Edison. It can find the Dexcom on BT and I set the SN to SM52xxxxxxx However, when I go to pull records, I get: AttributeError: 'NoneType' object has no attribute 'iter_records'
Is that where you are at too?
Scott Leibrand
@scottleibrand
Jan 14 2016 03:16
Yep, same.
Does BT icon on receiver ever stop blinking?
Toby Canning
@TC2013
Jan 14 2016 03:16
I don't know, I'll check
Scott Leibrand
@scottleibrand
Jan 14 2016 03:16
mine blinks for 20m and then errors out with a "pairing failed"
Jason Calabrese
@jasoncalabrese
Jan 14 2016 03:19
is there a way to have multiple reports point to the same file? besides something ugly like symlinks
Toby Canning
@TC2013
Jan 14 2016 03:20
I just turned share off and on. Got it blinking, ran list_dexcom and that stopped the blinking, but it doesn't show paired
Error:
Scott Leibrand
@scottleibrand
Jan 14 2016 03:23
@jasoncalabrese don't think so. I do a conditional mv or rsync -tu afterwards.
Toby Canning
@TC2013
Jan 14 2016 03:23
paired? 0
pairing...
paired
finding service
SERVICE None
Traceback (most recent call last):
  File "/home/edison/Adafruit_Python_BluefruitLE/Adafruit_BluefruitLE/bluez_dbus/provider.py", line 121, in _user_thread_main
    self._return_code = target()
  File "/home/edison/openxshareble/openxshareble/__init__.py", line 85, in run
    output = self.main(args, app)
  File "/home/edison/src/openaps/openaps/vendors/dexcom.py", line 91, in main
    for item in self.dexcom.iter_records('EGV_DATA'):
AttributeError: 'NoneType' object has no attribute 'iter_records'
I took out some junk in that
Scott Leibrand
@scottleibrand
Jan 14 2016 03:24
"Paired" is meaningless.
but blinking stopping is a good sign.
Toby Canning
@TC2013
Jan 14 2016 03:25
Did you install python-dbus?
Toby Canning
@TC2013
Jan 14 2016 03:31
It has got to be close: Name = DEXCOMRX
Paired = 0
Adapter = /org/bluez/hci0
LegacyPairing = 0
Alias = DEXCOMRX
Connected = 1
UUIDs = dbus.Array([dbus.String(u'f0aca0b1-ebfa-f96f-28da-076c35a521db')], signature=dbus.Signature('s'), variant_level=1)
Address = CA:A7:DD:E9:CE:F9
Trusted = 0
Blocked = 0
None
paired? 0
pairing...
paired
Scott Leibrand
@scottleibrand
Jan 14 2016 03:32
Yeah, had to do python-dbus to get it to even try.
Toby Canning
@TC2013
Jan 14 2016 03:33
Okay, I'll let you enjoy the evening. I'll let you know if I figure anything out
Ben West
@bewest
Jan 14 2016 03:55
devices can have many uses
the reason we're adding ad-hoc devices is because we're still discovering, collectively, what the thing needs to do
that's a signal that either new devices with new uses are needed, or additional uses could be added to some existing devices
we've already done this quite a bit with the nightscout and oref0 wrapper tools
Pete Schwamb
@ps2
Jan 14 2016 03:58
‘devices’ in this context don’t necessarily map to physical devices, right?
Ben West
@bewest
Jan 14 2016 04:14
in particular, a device is a bit of config which names which a vendor implementation with it's extra config
so the oref0- are all "process" aka shell/bash processes all process vendor types have a shell use
I've tried to demonstrate how to accept an $1 to allow single process to take more than one kind of use
but it's a lot easier in python
device is a name, a registered vendor type, and any extra optional config
Thomas Emge
@ThomasEmge
Jan 14 2016 04:32
@scottleibrand I see you are tinkering with your own challenge...however tapping into the vast pool of knowledge.... even though my cron script is 'working' what else could be reasons for mm-stick warmup to produce a 1 while the openaps reports are working?
schulzedj
@schulzedj
Jan 14 2016 04:50
I got a fun error tonight when running use cgm iter_glucose 5

``` raise GitCommandError(command, status, stderr_value)
git.exc.GitCommandError: 'git commit -avm openaps-use cgm iter_glucose 5

  TODO: better change descriptions
  /usr/local/bin/openaps-use cgm iter_glucose 5
  ' returned with exit code 128

stderr: 'error: file write error (No space left on device)
fatal: unable to write sha1 file'```

Does this mean it's not rewriting over prior reports?
Scott Leibrand
@scottleibrand
Jan 14 2016 04:58
Your disk is full.
Nothing will work until you free up some disk space.
If you didn't do it when you first installed Raspbian, you can try sudo raspi-config and do the expand file system thing.
If you've already done that you'll need to use tools like du to find out what is using up all the space so you can clean up.
schulzedj
@schulzedj
Jan 14 2016 05:23
So I left it alone for 5 minutes and was then able to run the report...weird.
Am running loop and am getting the following output: ValueError: No Closing Quotation No CGM data
Jason Calabrese
@jasoncalabrese
Jan 14 2016 05:26
What does df -h show?
schulzedj
@schulzedj
Jan 14 2016 05:27
pi@raspberrypi:~/daneopenaps $ df -h Filesystem Size Used Avail Use% Mounted on /dev/root 3.5G 3.3G 1.4M 100% / devtmpfs 459M 0 459M 0% /dev tmpfs 463M 0 463M 0% /dev/shm tmpfs 463M 6.4M 457M 2% /run tmpfs 5.0M 4.0K 5.0M 1% /run/lock tmpfs 463M 0 463M 0% /sys/fs/cgroup /dev/mmcblk0p1 60M 20M 41M 34% /boot tmpfs 93M 0 93M 0% /run/user/1000
Is there a better way to copy and paste my output so it's recognized in this window?
Scott Leibrand
@scottleibrand
Jan 14 2016 05:29
Put ``` on its own line before and afterward.
Dana Lewis
@danamlewis
Jan 14 2016 05:29
Do three of the ticks, go to a new line and paste, and then three ticks on another line below that
Jason Calabrese
@jasoncalabrese
Jan 14 2016 05:29
Your root partition is full
Scott Leibrand
@scottleibrand
Jan 14 2016 05:29
Those are backticks, not single quotes.
Jason Calabrese
@jasoncalabrese
Jan 14 2016 05:29
Only 1.4M left
Need to delete something before it get corrupted
schulzedj
@schulzedj
Jan 14 2016 05:32
So I have zero idea why it's full.
how does one go about addressing this full root partition
Jason Calabrese
@jasoncalabrese
Jan 14 2016 05:33
Try du -sh /var/log
schulzedj
@schulzedj
Jan 14 2016 05:34
and I guess trying to figure out what's dumping into this
I get the permission denied
Jason Calabrese
@jasoncalabrese
Jan 14 2016 05:36
Use sudo
schulzedj
@schulzedj
Jan 14 2016 05:37
Output: 4.5M /var/log
Jason Calabrese
@jasoncalabrese
Jan 14 2016 05:38
You can try some other dirs /home /usr
Since could take a lot of time
schulzedj
@schulzedj
Jan 14 2016 05:39
what exactly is -sh doing?
Jason Calabrese
@jasoncalabrese
Jan 14 2016 05:40
Summary humanized
It includes all messed dirs and files
Nested
schulzedj
@schulzedj
Jan 14 2016 05:40
So you're saying try something like `sudo du -sh /home?
Jason Calabrese
@jasoncalabrese
Jan 14 2016 05:40
Yeah
schulzedj
@schulzedj
Jan 14 2016 05:41
Going to have to Google that for greater context
Jason Calabrese
@jasoncalabrese
Jan 14 2016 05:41
Something will be big
Scott Leibrand
@scottleibrand
Jan 14 2016 05:41
sudo du -xd 3 -h | grep [0-9]G
or sudo du -xd 3 -h / | grep [0-9]G
Jason Calabrese
@jasoncalabrese
Jan 14 2016 05:42
My take a while on a full files system
Might
Scott Leibrand
@scottleibrand
Jan 14 2016 05:43
if it only returns something like 2.1G / then try a variation like sudo du -xd 3 -h / | grep [0-9][0-9][0-9]M
that will give you all the big directories and subdirectories down to 3 levels deep
the [0-9]G one says to look for directories of at least one GB
the other is >100M
takes almost no time to run those for me. SSDs are pretty fast for reads
schulzedj
@schulzedj
Jan 14 2016 05:45
Just as a point, whenever I use the "```" I am not able to actually submit. It's almost like I'm holding down the shift+enter
Scott Leibrand
@scottleibrand
Jan 14 2016 05:46
you should be able to after doing the second one
schulzedj
@schulzedj
Jan 14 2016 05:46
Output is `1.3 G /usr/lib 2.2G /usr 3.3G /
Scott Leibrand
@scottleibrand
Jan 14 2016 05:46
it wants to let you paste everything (including the newlines in the paste) before it treats a newline as an "I'm done"
schulzedj
@schulzedj
Jan 14 2016 05:47
I know I should...and I know it's user error
Scott Leibrand
@scottleibrand
Jan 14 2016 05:47
do you use the graphical interface on your Pi?
per the 2nd answer, you can run something like dpkg-query -Wf '${Installed-Size}\t${Package}\n' | sort -n | tail to see which packages are using the most space
schulzedj
@schulzedj
Jan 14 2016 05:49
Are you referring to the raspi-config?
for GUI?
Scott Leibrand
@scottleibrand
Jan 14 2016 05:49
no, the point and click interface with a mouse and windows
or do you do everything via the text command line?
(most of us don't use the GUI, but I know @esteward does)
if you never plug it into a monitor mouse and keyboard, then you can't/don't use the GUI
if you do use the mouse, you do
schulzedj
@schulzedj
Jan 14 2016 05:51
I have in the past to get the IP address but no longer have a separate monitor so would have to do it at work
Scott Leibrand
@scottleibrand
Jan 14 2016 05:51
reason I ask is, that GUI takes up a lot of space, so if it shows up in that dpkg-query output, you could remove it if you don't use it
re: raspi-config, have you tried the expand filesystem thing yet? If your SD card is actually bigger than 4GB, that would let it use the extra space
schulzedj
@schulzedj
Jan 14 2016 05:56
pi@raspberrypi:~/daneopenaps $ dpkg-query -Wf '${Installed-Size}\t${Package}\n' |sort -n | tail
43834   nuscratch
45673   scratch
47401   libpython3.4-dev
55920   pypy-upstream
75760   libreoffice-common
94821   libreoffice-core
106515  sonic-pi
110561  raspberrypi-bootloader
181992  oracle-java8-jdk
448821  wolfram-engine
Ah...need to press ctrl+enter
Scott Leibrand
@scottleibrand
Jan 14 2016 05:56
ok, then do sudo apt-get remove wolfram-engine to start with
then do the same for the oracle, sonic, libreoffice, and scratch stuff
all of that is GUI stuff you won't use even if you want to log in and use the mouse
wolfram is a mathematica thing, libreoffice is a free clone of MS office, sonic is audio stuff...
java is used for a lot of things, but nothing we care about
schulzedj
@schulzedj
Jan 14 2016 05:58
Ok- then is it reasonable to think my reports are not rewriting but just stacking up files?
Scott Leibrand
@scottleibrand
Jan 14 2016 05:58
scratch is some graphical learn-to-code environment for kids.
schulzedj
@schulzedj
Jan 14 2016 05:59
well maybe I should keep scratch ;-)
Scott Leibrand
@scottleibrand
Jan 14 2016 05:59
heh. run it on your computer instead. :)
or get another Pi for that kind of stuff if you want
I think Pi's come with minecraft too
generally your loop will fail in some way if the disk is full, and be unable to enact any temps
schulzedj
@schulzedj
Jan 14 2016 06:02
I did get a 8 G card as a back up... maybe use that once I get everything running on this one and copied over
Scott Leibrand
@scottleibrand
Jan 14 2016 06:02
yeah. you'll want to make a backup of the smaller one, since you can easily write a small image to a large card, but not the other way around
schulzedj
@schulzedj
Jan 14 2016 06:03
That was my plan
so my /dev/root is now down to 75%
definitely better
Scott Leibrand
@scottleibrand
Jan 14 2016 06:11
:+1:
schulzedj
@schulzedj
Jan 14 2016 06:12
Doing the raspi-config thing now
And that brought me down to 37%...Go team!
Scott Leibrand
@scottleibrand
Jan 14 2016 06:14
:-)
schulzedj
@schulzedj
Jan 14 2016 06:20
Looks like I have an error with my glucose.json
pi@raspberrypi:~/daneopenaps $ openaps loop
cgm://JSON/iter_glucose/glucose.json
reporting glucose.json
Traceback (most recent call last):
  File "/usr/local/bin/openaps", line 4, in <module>
    __import__('pkg_resources').run_script('openaps==0.0.7', 'openaps')
  File "build/bdist.linux-armv7l/egg/pkg_resources/__init__.py", line 745, in run_script
  File "build/bdist.linux-armv7l/egg/pkg_resources/__init__.py", line 1670, in run_script
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.7-py2.7.egg/EGG-INFO/scripts/openaps", line 167, in <module>
    app( )
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.7-py2.7.egg/openaps/cli/__init__.py", line 44, in __call__
    self.run(self.args)
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.7-py2.7.egg/EGG-INFO/scripts/openaps", line 153, in run
    builtins.dispatch(args, self)
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.7-py2.7.egg/openaps/builtins.py", line 48, in dispatch
    get_alias(args.command, app)(args)
  File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.7-py2.7.egg/openaps/builtins.py", line 27, in __call__
    spec_command = shlex.split(self.spec.fields.get('command'))
  File "/usr/lib/python2.7/shlex.py", line 279, in split
    return list(lex)
  File "/usr/lib/python2.7/shlex.py", line 269, in next
    token = self.get_token()
  File "/usr/lib/python2.7/shlex.py", line 96, in get_token
    raw = self.read_token()
  File "/usr/lib/python2.7/shlex.py", line 172, in read_token
    raise ValueError, "No closing quotation"
ValueError: No closing quotation
No CGM data.
Which is weird because there is output when I cat glucose.json
Scott Leibrand
@scottleibrand
Jan 14 2016 06:22
"No closing quotation", huh? might not be glucose.json that is the problem
schulzedj
@schulzedj
Jan 14 2016 06:22
Yeah...I am wondering if it's in my alias for the loop
Scott Leibrand
@scottleibrand
Jan 14 2016 06:22
try running individual reports (the components of the openaps loop alias)
narrow it down until you have just one report that fails
schulzedj
@schulzedj
Jan 14 2016 06:28
I did not use /monitor in my setup. I am wondering how the instructions for openaps alias add gather '! bash -c "rm monitor/*; openaps monitor-cgm && openaps monitor-pump && openaps get-settings"' would look.
I'm trying rm*;
What is the rm in reference to here?
Scott Leibrand
@scottleibrand
Jan 14 2016 06:30
don't do that.
rm means remove
reason we put all the monitor reports in that dir was to be able to remove them all each loop without affecting other stuff.
schulzedj
@schulzedj
Jan 14 2016 06:30
so I should go back and add everything with for monitor?
Scott Leibrand
@scottleibrand
Jan 14 2016 06:31
That's one approach. Another would be to list all the individual files you want to rm
schulzedj
@schulzedj
Jan 14 2016 06:32
for example I do openaps pump vs openaps monitor-pump
Scott Leibrand
@scottleibrand
Jan 14 2016 06:32
That doesn't matter.
What matters is monitor/something.json reports.
schulzedj
@schulzedj
Jan 14 2016 06:33
And I don't have any of those
Scott Leibrand
@scottleibrand
Jan 14 2016 06:33
Those go into the monitor/ directory if you do it that way.
schulzedj
@schulzedj
Jan 14 2016 06:33
I never created a monitor directory...
though it sounds like I should
that is the general approach my loop also follows, and that we used in the docs.
schulzedj
@schulzedj
Jan 14 2016 06:39
ok- I think I understand what you're showing there. It will be a good exercise to go back and remove and add the goods back in.
@scottleibrand Thanks!
Scott Leibrand
@scottleibrand
Jan 14 2016 06:42
You might also want to start over in a new directory so you have a "clean" setup.
And so you can keep using your current one until the new one is done.
schulzedj
@schulzedj
Jan 14 2016 06:44
so where I'm in daneopenaps do something like 2openaps
Scott Leibrand
@scottleibrand
Jan 14 2016 06:46
Yep
Thomas Emge
@ThomasEmge
Jan 14 2016 07:16
I don't think I quite understand what monitor/clock_zoned.json is used for? Would it turn 2016-01-13T23:10:38 coming from the pump into something like 2016-01-13T23:10:38-0800?
Scott Leibrand
@scottleibrand
Jan 14 2016 07:18
Exactly.
For NS upload mostly. It needs absolute timestamps.
Thomas Emge
@ThomasEmge
Jan 14 2016 07:22
hmm, I see openaps device add tz timezones in your script. How does this work? my openaps is complaining that timezones is not in the list of expected arguments
Scott Leibrand
@scottleibrand
Jan 14 2016 07:25
Look up higher for where we added the vendor.
Thomas Emge
@ThomasEmge
Jan 14 2016 07:32
:worried: I am overlooking code again...I think tomorrow is another wonderful day to tackle the NS uploads. Thanks for your help today.
Scott Leibrand
@scottleibrand
Jan 14 2016 07:34
Heh. Have a good night.
eyim
@eyim
Jan 14 2016 07:43

@oskarpearson Hey Oskar - I am attempting Option 1 of https://github.com/oskarpearson/mmeowlink/wiki
and am getting this error when trying to install jessie on the edison according to these instructions: https://gist.github.com/oskarpearson/b5b24a725a04e4932c33

Any idea why it can't find the files? I was able to get ubilinux installed

on the command apt-get update I get the following error

root@alanedison:~# apt-get update
Err http://http.debian.net jessie Release.gpg
  Could not resolve 'http.debian.net'
Err http://security.debian.org jessie/updates Release.gpg
  Could not resolve 'security.debian.org'
Err http://http.debian.net jessie-updates Release.gpg
  Could not resolve 'http.debian.net'
Err http://http.debian.net jessie-backports Release.gpg
  Could not resolve 'http.debian.net'
Reading package lists... Done
W: Failed to fetch http://http.debian.net/debian/dists/jessie/Release.gpg  Could not resolve 'http.debian.net'

W: Failed to fetch http://http.debian.net/debian/dists/jessie-updates/Release.gpg  Could not resolve 'http.debian.net'

W: Failed to fetch http://security.debian.org/dists/jessie/updates/Release.gpg  Could not resolve 'security.debian.org'

W: Failed to fetch http://http.debian.net/debian/dists/jessie-backports/Release.gpg  Could not resolve 'http.debian.net'

W: Some index files failed to download. They have been ignored, or old ones used instead.
Ben West
@bewest
Jan 14 2016 07:46
that looks like wifi config
and the you are failing to download apt info because wifi isn't working?
eyim
@eyim
Jan 14 2016 07:47
I have it hooked up via a serial connection. Okay let me try to get wifi up again
Thanks Ben - I messed up the wifi setup. Downloading working now
Ben West
@bewest
Jan 14 2016 07:53
nice :-)
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 21:53
Good evening/morning everyone!
how to determine firmware version of my pump?
(preferably reotely)
Jason Calabrese
@jasoncalabrese
Jan 14 2016 21:57
on the pump just press esc and scroll down
Dana Lewis
@danamlewis
Jan 14 2016 22:04
@ktomy Just saw your entry :) let me know and I'll update the spreadsheet!
Doesn't have to be now if pump isn't nearby
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:06
my pump is in the bed and is budy looping now :)
So thank you so much everyone for this, my wife if happy, she can sleep without intreruptions (me too :) , but I can bbolus/temp without really waking up :) )
So my next tasks would be:
1) to see why treatment records are duplicated while uploading
2) to put my scripts (which are replacing aliases with bash commands) somewhere, I think they are more comprehensive
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:12
3) to study oib and determine-basal calculation and to understand why loop is not working well when I have active carbs
4) to help with documenting and development
Scott Leibrand
@scottleibrand
Jan 14 2016 22:15
what symptoms are you seeing on #3?
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:21
Effect is that bg raises and there is no enough insulin because low-temps are enacted
but not very sure yet, I need more data
I am not sure, but I think that loop dowsn't know that I have carbs on board
...last statement is not true as I can see bolussnooze != 0
:)
anyway, I'm not sure yet, 4 nights is definitly not enough to understand if the system needs fine-tuning
Scott Leibrand
@scottleibrand
Jan 14 2016 22:25
when do you see it enacting the low-temps? right after you bolus, or an hour or two later?
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:25
imediately after
Scott Leibrand
@scottleibrand
Jan 14 2016 22:25
I would like to look at the output with you and help figure out what it's doing and why
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:27
ok, thank you, in this night I don't have abefore-sleep-eating situation and I have deleted old data after git definite corruption.
after next reproduction I'll contact you and we'll see what happens
is it ok?
Scott Leibrand
@scottleibrand
Jan 14 2016 22:28
k. if you're uploading to nightscout, that often has good retrospective data
(independent of git)
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:31
yes, I know, but I could see this effect before I changed pump settings dramaticly (isf, basals) and one of the possible causes was a wrong ISF, so I'd prefer to wait untill next ocurance.
also a strange think i've seen that when loop low-temp-ed the pump to 0 it doesn't say "nothing to do" but is setting temp basal each time to 0
and I think it is a simple bug somewhere
Scott Leibrand
@scottleibrand
Jan 14 2016 22:34
if the adjustment required takes longer than 30m at 0 to achieve, it will extend the temp to 30m
no, that's how it's supposed to work
if the temp required is between 0 and normal basal, it won't extend the temp
but if it's less than zero (i.e. requires more than 30m at zero) it extends
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:35
blob
blob
Scott Leibrand
@scottleibrand
Jan 14 2016 22:35
we could have it set a temp to zero for 60m, but usually something changes by then, so you actually don't end up needing a full 60m temp to zero
do you have the openaps pill for that timeframe?
would like to see the tooltip from it at the time it initially temped to zero
(or we can look next time)
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:36
just a second
blob
Scott Leibrand
@scottleibrand
Jan 14 2016 22:41
deviation -170? that's crazy
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:41
blob
Scott Leibrand
@scottleibrand
Jan 14 2016 22:42
what do you have his ratios and basal rates set to?
maybe just paste your profile.json
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:42
just a second
blob
ah, sorry
just a second, profile.json is reloading
{"max_iob":4,"type":"current","dia":2,"current_basal":0.5,"max_daily_basal":0.9,"max_basal":1.5,"min_bg":105,"max_bg":110,"sens":120}
I can give you all monitor/ and settings/ files if you you need more info
Scott Leibrand
@scottleibrand
Jan 14 2016 22:46
hmm, sens of 120 and basals in the 0.5-0.9 range? how did you determine sens?
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:47
:) kind of try-and-fail
Scott Leibrand
@scottleibrand
Jan 14 2016 22:47
I would expect a sens closer to 80 for that level of basal
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:48
sens should be bigger but in last few days I have lowered it because loop didn't lower bg enough
Scott Leibrand
@scottleibrand
Jan 14 2016 22:48
have you ever done a test where BG is high and flat, you give a single correction bolus, and see how much he drops from it?
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:49
yes, some time ago (about a year ago tried that) ang get something like 0.1u for 1.5mmol , meaning about 270
Scott Leibrand
@scottleibrand
Jan 14 2016 22:49
what is his weight?
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 22:50
but since tht time my chilg has grown and I think sensitivity decreased dramaticaly
16kg
tdd is about 11-13u
carb ratios are between 20 and 30
Scott Leibrand
@scottleibrand
Jan 14 2016 22:52
I guess that ISF is consistent with that TDD... if carb ratio to ISF ratio were constant, that'd give an ISF of 80-120 as well
something else is going on here besides just ISF
I need to see the full determine-basal output (stderr + stdout) to see how it gets a -170 deviation
what were the BGs 15m prior to the temp to zero?
(I'm trying to calculate the 15m avgdelta)
drspike
@drspike
Jan 14 2016 22:57
where is the best place to see how to get OpenAPS to pull data from nightscout? I have created my curl device but how do I alter my monitor/glucose.json report? I am only on Phase 1 and have only messed with a few commands and made a few different reports
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:10
Sorry, internet connection failed, pump is down (i did mot configured offline mode) :( cannot give you logs right now as i am on my phone now
Sorry, pump is ok, loop is down :)
Scott Leibrand
@scottleibrand
Jan 14 2016 23:14
ok. let's revisit when it happens again
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:16
my internet is back, so, I think that the key of that 170 (which I wasn't aware of, i just seen 0 temp) is that I had no readings for some time and only 3 readings (or if it acted before 3rd reading - 2 readings before that 0-temp, just a secont' I'll show you.
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:17
blob
Scott Leibrand
@scottleibrand
Jan 14 2016 23:17
@drspike are you asking how to accomplish the second half of that? one approach is simply to edit the .ini with a text editor.
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:18
and I think I need to copy .git directory to my desktop to identify all the values in all the reports
Scott Leibrand
@scottleibrand
Jan 14 2016 23:18
@ktomy what was the last BG reading before that calibration?
was it hours before and/or super high?
drspike
@drspike
Jan 14 2016 23:18
using the nano text editor? Ill try that. Thank you
Scott Leibrand
@scottleibrand
Jan 14 2016 23:19
yeah
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:19
blob
Scott Leibrand
@scottleibrand
Jan 14 2016 23:19
aha!
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:19
exactly
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:21
so I think in determine-basal we should check also time of last 3 gbs, not only their values
Scott Leibrand
@scottleibrand
Jan 14 2016 23:21
yes. that has been a TODO since I first wrote that code. want to tackle it? :)
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:21
yes, saw that when modified dateString handling
yes, let me try
Scott Leibrand
@scottleibrand
Jan 14 2016 23:22
sweet
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:24
can you please tell me what ide are you using (if you're running windows)? I really don't want to write in mcedit again :)
Scott Leibrand
@scottleibrand
Jan 14 2016 23:24
vim ;-)
(not an IDE, just an editor with syntax highlighting and decent indent support)
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:25
yea, ok :) I will never understand it
Scott Leibrand
@scottleibrand
Jan 14 2016 23:25
sublime text 2 is good on Mac. dunno for Windows.
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:26
so I will try to run it in my vs.net
Scott Leibrand
@scottleibrand
Jan 14 2016 23:26
actually yeah, I think that's what @amazaheri uses
Ali Mazaheri
@amazaheri
Jan 14 2016 23:27
correct, VS Code or VS Studio has built in python, nodejs support allowing to debug and test too :)
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:29
yes, and I really can't understand people programming without step-by-step debugger
Ali Mazaheri
@amazaheri
Jan 14 2016 23:30
:)
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:30
the alert('I was here') style is 20 years old :)
Ali Mazaheri
@amazaheri
Jan 14 2016 23:30
well they can run Code on MAC or Linux and its opensource :)
Visual Studio Code that is :)
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:31
I am sure that there are native linux IDEs, but people is proud of using vim and emacs :)
Ali Mazaheri
@amazaheri
Jan 14 2016 23:32
joking aside VS makes debugging and coding nodejs and python really easy, that's how I ported OpenAPS to Windows (was planning to use Intel Stick couple of months ago) but @bewest and @oskarpearson got me back to Linux and Edison hehe
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:34
why edisson? i mean, are there really reasons to use x86?
Ali Mazaheri
@amazaheri
Jan 14 2016 23:35
Intel Compute Stick runs both Windows and 10 and idea was built in wifi, BLE and size. Edison gives u all that and super compact will less power usage, I am sold :)
But Windows is my main dev platform as I am doing extra related stuff on Azure and VS covers all that
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:36
yes, embedded ble and wifi are good reasons to make a switch
Ali Mazaheri
@amazaheri
Jan 14 2016 23:38
join https://gitter.im/oskarpearson/mmeowlink if interested a lot of good stuff happing there
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:38
yes, i work as a .net dev last 15 years
Ali Mazaheri
@amazaheri
Jan 14 2016 23:38
I don't feel alone anymore :)
time for another meeting! talk to you all later
Alin Artiom Kenibasov
@ktomy
Jan 14 2016 23:39
k
Scott Leibrand
@scottleibrand
Jan 14 2016 23:51
:-) FWIW, I use vim because it allows me to edit anywhere on any server. I find it much easier to do things in place rather than try to edit on my laptop and then push the code wherever it needs to be
but that's also a sign of me being more of a sysadmin type than a proper developer
proper developers tend to find it worthwhile to use IDEs and have proper dev and test environments
Matthias Granberry
@mgranberry
Jan 14 2016 23:57
I've always seen vim/IDE stuff as a left-hand right-hand thing. Sure, one is easier to use, but it isn't always convenient.
and vim isn't exactly hard to use after 15 years