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

14th
Jun 2016
Thacher
@thacherT1D
Jun 14 2016 00:03
@dm61 I'm not using an attached dexcom device -- I'm on G5, so I'm pulling the data from NightScout, I am using the G5 receiver in addition to my iPhone though -- I previously tried to attach that receiver physically, should that work?
Lynne Butts
@flutelynne_twitter
Jun 14 2016 00:16
@thacherT1D I use the receiver attached (G5). Look at the latest doc: Here’s the link to that if you wanna look at it too - http://openaps.readthedocs.io/en/wip-understand-before-doing/docs/walkthrough/index.html
Brooke Pazoles
@PazolesBrooke_twitter
Jun 14 2016 00:23
@danamlewis working on it right now!
Dana Lewis
@danamlewis
Jun 14 2016 00:23
:) @PazolesBrooke_twitter
Chris Oattes
@cjo20
Jun 14 2016 00:26
@danamlewis how often does your cron job run?
Dana Lewis
@danamlewis
Jun 14 2016 00:30
@cjo20 every minute and checks to see if a process is still running.
Chris Oattes
@cjo20
Jun 14 2016 00:38
the docs suggest that when you start observing the closed loops, it operats in 'low glucose suspend mode', I would guess that this is meant to be achieved by setting the max_iob to 0u, but it doesn't say anywhere in the docs to do that. The only mention of max_iob is 'this is how you create it' and the example is 2u
Dana Lewis
@danamlewis
Jun 14 2016 00:38
it default to zero until you go to change it to something higher
Chris Oattes
@cjo20
Jun 14 2016 00:41
https://openaps.readthedocs.io/en/dev/docs/walkthrough/phase-2/Using-oref0-tools.html That says you need to create it, but doesn't say to set it to 0 initially. The only examples (there and in the oref0-mint-max-iob page) mention 2u
Paul Martin
@Spazholio
Jun 14 2016 01:11
Silly, basic question - what is the recommended way to install pip? When I try to use apt-get on the Edison, it wants to remove some stuff, but I’m kinda looking to avoid that.
Ben West
@bewest
Jun 14 2016 01:12
what's it removing?
Paul Martin
@Spazholio
Jun 14 2016 01:12
The following actions will resolve these dependencies:

      Remove the following packages:              
1)      cpp-4.7                                   
2)      debconf-i18n                              
3)      gcc-4.7                                   
4)      gcc-4.7-base                              
5)      liblocale-gettext-perl                    
6)      libpython2.7                              
7)      libtext-charwidth-perl                    
8)      libtext-iconv-perl                        
9)      libtext-wrapi18n-perl                     
10)     python-dev                                
11)     python2.7-dev                             
12)     tasksel                                   
13)     tasksel-data                              

      Leave the following dependencies unresolved:
14)     debconf recommends debconf-i18n
Ben West
@bewest
Jun 14 2016 01:12
what's the apt-get part though?
probably shouldn't need to remove those things
Paul Martin
@Spazholio
Jun 14 2016 01:14
Actually, I misspoke. aptitude wants to remove stuff. apt-get wants to upgrade and quite possibly do what I want.
That’s behavior I’ve not seen before.
Ugh. Nevermind. This aptitude is behaving slightly differently than I’m used to. Normally if it’s updating something, it doesn’t explicitly state it’s going to remove the old packages. It just does it in favor of the upgrade.
Brett
@burgertime14
Jun 14 2016 01:19
has anyone added a cellular shield to an edison rig?
Thacher
@thacherT1D
Jun 14 2016 01:20
@flutelynne_twitter I've followed through the directions for the attached G5, but I'm getting an empty array '[ ]' when I pull the data?
Garrett Webb
@garetis
Jun 14 2016 01:33
Is anyone here using medtronic CGM and android?
Garrett Webb
@garetis
Jun 14 2016 01:43
Also, what is the profile.json file in openaps/settings? When I cat it I don't get a return.
Brett
@burgertime14
Jun 14 2016 01:45
@garetis have you run openaps report invoke settings/profile.json?
Garrett Webb
@garetis
Jun 14 2016 01:47
I got this error:
```
pi@raspberrypi:~/openaps1 $ openaps report invoke settings/profile.json
get-profile://text/shell/settings/profile.json

module.js:340
    throw err;
          ^
Error: Cannot find module '/home/pi/openaps1/settings/bg_targets_raw.json'
    at Function.Module._resolveFilename (module.js:338:15)
    at Function.Module._load (module.js:280:25)
    at Module.require (module.js:364:17)
    at require (module.js:380:17)
    at Object.<anonymous> (/usr/local/lib/node_modules/oref0/bin/oref0-get-profile.js:44:26)
    at Module._compile (module.js:456:26)
    at Object.Module._extensions..js (module.js:474:10)
    at Module.load (module.js:356:32)
    at Function.Module._load (module.js:312:12)
    at Function.Module.runMain (module.js:497:10)
reporting settings/profile.json
I manually created my basal profile, bg targets, and read settings JSON files
Chris Oattes
@cjo20
Jun 14 2016 01:49
it says it can't find the bg_targets file
profile.json is created by get-profile
Garrett Webb
@garetis
Jun 14 2016 01:50
is that because it is looking for a file that says "bg_targets_raw.json, and I have named it bg_target.json?
Brett
@burgertime14
Jun 14 2016 01:50
yes :)
Chris Oattes
@cjo20
Jun 14 2016 01:50
that's certainly a possibility :P
Brett
@burgertime14
Jun 14 2016 01:50
I may have done that myself as some point
Chris Oattes
@cjo20
Jun 14 2016 01:50
I've done that about 5 times this evening
Garrett Webb
@garetis
Jun 14 2016 01:51
what's the best way to rename file names? nano?
Brett
@burgertime14
Jun 14 2016 01:51
mv <file name> <new file name>
Paul Martin
@Spazholio
Jun 14 2016 01:51
@jyaw IT WORKS! I have no idea what’s different now, but openxshareble is finally freaking working! I suppose I can’t upgrade from bluez now, but I don’t care! WOOO!
Garrett Webb
@garetis
Jun 14 2016 01:52
Ok, thanks @burgertime14
Brett
@burgertime14
Jun 14 2016 01:53
:thumbsup:
Jarred Yaw
@jyaw
Jun 14 2016 01:55
@Spazholio congrats! Was easier having someone on the journey with me. I'm glad to be finished with that as well!
Paul Martin
@Spazholio
Jun 14 2016 01:57
Now I just have to make sure that I’m connected 24/7. =)
Brett
@burgertime14
Jun 14 2016 01:58
@Spazholio that was why I was asking if anyone has added a cellular shield (cell modem) to their edison. I am starting to collect parts and don't want to rely on a mifi
Paul Martin
@Spazholio
Jun 14 2016 01:58
Are you getting raw info out of it when you run the openaps use share iter_glucose command?
Ben West
@bewest
Jun 14 2016 01:58
cell modem would be great
or GetFirmwareHeader
I like that one
Paul Martin
@Spazholio
Jun 14 2016 01:59
@burgertime14 Ah, I see. I’m always with my phone, so I just have my hotspot on.
Ben West
@bewest
Jun 14 2016 01:59
battery is fun
Paul Martin
@Spazholio
Jun 14 2016 01:59
@bewest Should I be getting raw data with the iter_glucose command?
Ben West
@bewest
Jun 14 2016 01:59
no
try oref0_glucose
Paul Martin
@Spazholio
Jun 14 2016 01:59
Oh good. =)
Interesting. What’s the field name for it?
  {
    "trend_arrow": "FLAT", 
    "system_time": "2016-06-14T04:00:32", 
    "direction": "Flat", 
    "sgv": 128, 
    "display_time": "2016-06-13T21:01:31", 
    "date": 1465866091000.0, 
    "dateString": "2016-06-13T21:01:31", 
    "unfiltered": 153664, 
    "filtered": 147008, 
    "device": "openaps://share", 
    "rssi": 199, 
    "type": "sgv", 
    "glucose": 128
  }
That’s what I’m seeing returned.
Ben West
@bewest
Jun 14 2016 02:13
looks good
zone it and send it to oref0/NS
unfiltered and filtered are the "raw" values
Paul Martin
@Spazholio
Jun 14 2016 02:14
And NS interprets those? They weren’t human-readable right off the bat, so wasn’t sure.
(oh, and the same instructions appear to not work for the Pi, but at least they work for the Edison)
Ben West
@bewest
Jun 14 2016 02:15
yeah, I got the field names from NS actually
it should probably be called ns_glucose but I thought that'd be more confusing...
anyway, this works for both out of the box :-)
Paul Martin
@Spazholio
Jun 14 2016 02:16
And I can use ns-upload-entries for this, right?
Ben West
@bewest
Jun 14 2016 02:16
just add zoning
Paul Martin
@Spazholio
Jun 14 2016 02:16
Yep.
Wait…what’s the RIGHT way to do zoning? I have one way, but there was someone else in here that I swear had a much simpler way that worked just as well. Wanna make sure I’m not over-complicating things. =)
Ben West
@bewest
Jun 14 2016 02:18
yes, with zones, this can be given directly to openaps use ns shell format-recent-type tz entries monitor/glucose.json
Paul Martin
@Spazholio
Jun 14 2016 02:19
Crap. I’ve got a split set-up. Don’t have it running on the Pi, so can’t integrate everything. Grrrr.
Have you seen openxshareble working on a Pi? Or just an Edison?
Ben West
@bewest
Jun 14 2016 02:19
actually, I've only got it working on pi
I have some built debs to solve it on pi
Paul Martin
@Spazholio
Jun 14 2016 02:20
That’s interesting. Can I give a quick (large) paste and you can see if anything obvious jumps out at you?
or pastebin.ca
or dpaste
if it's large
Paul Martin
@Spazholio
Jun 14 2016 02:21
This message was deleted
Ben West
@bewest
Jun 14 2016 02:21
on pi?
Yes, on Pi.
(sorry, had to edit out my serial)
Ben West
@bewest
Jun 14 2016 02:22
some instructions advise using /etc/rc.local which I don't do
I built a proper deb package for pi
both wheezy and jessie
it sets up dbus and perms and stuff by default via apt packaging
Paul Martin
@Spazholio
Jun 14 2016 02:23
I don’t use /etc/rc.local. I’m using systemd as well, so I had to modify the directions ever so slightly.
Ben West
@bewest
Jun 14 2016 02:23
hmm, you were sure to ./configure --with-experimental?
or --enable-experimental
forget
Paul Martin
@Spazholio
Jun 14 2016 02:23
I did.
And that version is running.
With that flag.
Ben West
@bewest
Jun 14 2016 02:24
hmmmm
what OS/bluez?
Paul Martin
@Spazholio
Jun 14 2016 02:24
(pi) [pancreas: ~/openaps] $ ps -ef | grep blue
root       706     1  0 22:10 ?        00:00:00 /usr/local/bin/bluetoothd --experimental
Ben West
@bewest
Jun 14 2016 02:24
and kernel
hmmm
Paul Martin
@Spazholio
Jun 14 2016 02:24
Jessie, fully updated, 5.37, and...
Ben West
@bewest
Jun 14 2016 02:25
uname -a or uname -r
Paul Martin
@Spazholio
Jun 14 2016 02:25
Linux pancreas 4.4.13-v7+ #893 SMP Wed Jun 8 14:39:18 BST 2016 armv7l GNU/Linux
Ben West
@bewest
Jun 14 2016 02:25
oh
should work
you have a ble usb dongle?
Paul Martin
@Spazholio
Jun 14 2016 02:25
I agree. =)
Ben West
@bewest
Jun 14 2016 02:25
or it's pi3?
Paul Martin
@Spazholio
Jun 14 2016 02:25
I’m not using one. Yes, it’s a Pi3.
I have one I could try though.
Ben West
@bewest
Jun 14 2016 02:25
ah, pi3
hmmm
I havent' tried it on a pi3
Paul Martin
@Spazholio
Jun 14 2016 02:26
I mean, you can see in the output that it sees the DEXCOMRX, so I think the onboard Pi3 is ok for BLE stuffs.
Ben West
@bewest
Jun 14 2016 02:26
it should be the same
ah ok
yeah... it should be the same
this is a brand new device...
you have a good kernel
Paul Martin
@Spazholio
Jun 14 2016 02:26
If I want to switch to the dongle, do I just run bluetoothctl and switch the device? It’ll persist when I leave bluetoothctl?
Ben West
@bewest
Jun 14 2016 02:27
no
hmm
if you want to try the dongle, hmmm
Paul Martin
@Spazholio
Jun 14 2016 02:27
I thought about specifying the MAC in share.ini so it won’t browse, and it’ll just look for that MAC...
I think it’s finding the right device though.
Ben West
@bewest
Jun 14 2016 02:28
I think it will create two adapters in hciconfig
you'll have an hci0 (the default, builtin) and then an hci1 (the usb dongle)
think easiest way to try the alternate ble stack is to turn the 0th adapter off
so only the usb one is "on"
sudo hciconfig hci0 down
Paul Martin
@Spazholio
Jun 14 2016 02:29
Doing so now.
Well, hell.
That did it.
Ben West
@bewest
Jun 14 2016 02:30
fascinating!!!!
really?
Paul Martin
@Spazholio
Jun 14 2016 02:31
I swear. =)
I am slightly displeased that it requires a dongle, but freaking ecstatic that it’s working.
I’d LOVE to know why that is though.
Ben West
@bewest
Jun 14 2016 02:32
yeah, me too
Paul Martin
@Spazholio
Jun 14 2016 02:32
Do you have a Pi3 to tinker on?
Ben West
@bewest
Jun 14 2016 02:32
no
Paul Martin
@Spazholio
Jun 14 2016 02:32
Want one? =)
Ben West
@bewest
Jun 14 2016 02:32
hmm sort of
Paul Martin
@Spazholio
Jun 14 2016 02:32
Seriously, I have an extra. I’d be willing to send it out to you to hack on. You know, in all your free time. =)
Ben West
@bewest
Jun 14 2016 02:33
hehe, that's the thing
David Cintron
@loudestnoise
Jun 14 2016 02:33
Give BEN ALL TEH THINGS
Paul Martin
@Spazholio
Jun 14 2016 02:33
Look, if there’s anything else out there that’s more important than my wants and needs, I want it found, captured, and shot.
David Cintron
@loudestnoise
Jun 14 2016 02:33
LOL
@bewest is your rig a RPi?
I just assumed you had one of these fancy tiny rigs
apabari
@apabari
Jun 14 2016 02:35
Does anyone have any updated documentation on how to get OpenAPS running on an Edison? I've followed the guidelines to try and get Debian working, and then Yocto... the OS seems to load, and OpenAPS seems to install, but when I try launching commands from OpenAPS I get errors. I'm just reflashing with Debian now, but thought I'd ask for help too!
Ben West
@bewest
Jun 14 2016 02:41
@loudestnoise I try to maintain one of each, across matrix of OS's
that way I can ensure best quality/compatibility across sytems
makes the whole thing more effective as a platform, more amenable to community
David Cintron
@loudestnoise
Jun 14 2016 02:41
gotcha
Ben West
@bewest
Jun 14 2016 02:41
(hardly any of them actually work at any time, I'm usually debugging something)
I lost my beautiful jessie rig at qsph :-(
down a pi, also broke a ti cc1111 by accident
Dan Caron
@dancaron
Jun 14 2016 02:43
Hey @bewest! Just joining here on Pete's recommendation. You said you spoke to Insulet recently? Care to share anything of interest?
Ben West
@bewest
Jun 14 2016 02:43
yes
oh
hmmm
I met them at a conference by diabetes technology society a few weeks ago called DTSEC
and that was actually a really good interaction
Dan Caron
@dancaron
Jun 14 2016 02:44
That's great
Ben West
@bewest
Jun 14 2016 02:44
the focus of the conference was on security
and the takeaway was "you (vendor) need to work with people who want to collaborate with you on data/security, regardless of your contract with them"
people will use your data in ways can't expect
Dan Caron
@dancaron
Jun 14 2016 02:46
can't argue that
Ben West
@bewest
Jun 14 2016 02:46
I mean, I didn't have to make that point :-)
chris hannemann hammered it home though
we exchanged contacts and I received email from them
so they are thinking about how to work with us, or at least converse with us
at ADA, I spent some time in their booth early Sunday
got their demo and talked with some people about openaps
gave them my demo :-)
went back again with Scott and we ran into different group of people
we wound up talking for probably 45 minutes or maybe hour
the hall closed and the lights were off
we were last ones out
Paul Martin
@Spazholio
Jun 14 2016 02:49
It’s heartening to hear that all these different vendors are actually open to discussion. Whether or not something actually comes from it is another story, but…it’s a start.
Ben West
@bewest
Jun 14 2016 02:49
they are kind of "where" dexcom was when this all started
they are very intrigued but also a little nervous
they don't feel they are in a position to help us directly
but they do want an opportunity to learn and perhaps build in better ways to work together in the future
Paul Martin
@Spazholio
Jun 14 2016 02:50
Well, re-implementing temporary boluses remotely on new firmwares would be a nice start. That’s just re-enabling old functionality.
Ben West
@bewest
Jun 14 2016 02:50
one thing they said I thought was really fasicinating
they were willing to say they wanted to focus on hardware and not on software
I thought that was one of the most interesting things I heard
yeah, the way the system depends on PDM makes it a bit awkward for them to help us "today"
but in future, if we could have a wireless "chat" with the PDM, that could allow us to augment their system very easily
pure software
Paul Martin
@Spazholio
Jun 14 2016 02:52
On another note, I tried to write the BLE Dexcom info to a file via > and it worked, but I got all the Bluetooth negotiation text as well. How to best omit that?
Dan Caron
@dancaron
Jun 14 2016 02:54
that's very cool @bewest thanks for sharing!
Ben West
@bewest
Jun 14 2016 02:54
hmmm
hmm
mine doesn't...
oh
update your Adafruit thingy
Paul Martin
@Spazholio
Jun 14 2016 02:54
I did.
git pull says I’m up to date.
I’m on master. Switch to dev?
Ben West
@bewest
Jun 14 2016 02:55
bewest@bewest-MacBookPro:~/src/Adafruit_Python_BluefruitLE$ git rev-parse HEAD
ohhhh
hehe
git checkout wip/bewest/custom-gatt-profile
sans typos
Paul Martin
@Spazholio
Jun 14 2016 02:56
In the Adafruit dir?
Ben West
@bewest
Jun 14 2016 02:56
yeah
I'm surprised it worked at all
Paul Martin
@Spazholio
Jun 14 2016 02:56
Oops.
I was working on the openxshareble dir.
However….
Already on 'wip/bewest/custom-gatt-profile’
Ben West
@bewest
Jun 14 2016 02:57
hmmm openxshareble should probably be on master I think
Paul Martin
@Spazholio
Jun 14 2016 02:57
Yeah, that one is.
Ben West
@bewest
Jun 14 2016 02:57
bewest@bewest-MacBookPro:~/src/Adafruit_Python_BluefruitLE$ git branch 
  master
* wip/bewest/custom-gatt-profile
  wip/pydbus
  wip/studying-glib-mainloop
bewest@bewest-MacBookPro:~/src/Adafruit_Python_BluefruitLE$ git rev-parse HEAD
3b15cad0245497d7d043e2b2ba7c447f177bd877
bewest@bewest-MacBookPro:~/src/Adafruit_Python_BluefruitLE$
Paul Martin
@Spazholio
Jun 14 2016 02:58
(pi) [pancreas: ~/src/Adafruit_Python_BluefruitLE] $ git rev-parse HEAD
c0f2ad44b5170145a75a9f026d5c1a00c183a24f
(pi) [pancreas: ~/src/Adafruit_Python_BluefruitLE] $
Well, there’s the problem. Unsure how to rectify though. =)
Ben West
@bewest
Jun 14 2016 02:59
check which branch your on
but probably just git pull
Paul Martin
@Spazholio
Jun 14 2016 02:59
(pi) [pancreas: ~/src/Adafruit_Python_BluefruitLE] $ git branch
* wip/bewest/custom-gatt-profile
(pi) [pancreas: ~/src/Adafruit_Python_BluefruitLE] $
Ben West
@bewest
Jun 14 2016 02:59
yeah, git pull
Paul Martin
@Spazholio
Jun 14 2016 02:59
git pull says I’m up to date.
Ben West
@bewest
Jun 14 2016 02:59
oh
...
oh dear...
no... it's fine
oh
Paul Martin
@Spazholio
Jun 14 2016 03:00
That’s what says when something just caught on fire, but they don’t want to panic others...
Ben West
@bewest
Jun 14 2016 03:00
hmm
well someone updated the readme
I did a git pull
c0f2ad44b5170145a75a9f026d5c1a00c183a24f
now I get same as you
ok
sudo python setup.py develop
should just re-install it... mine goes to syslog
I mean all that extra stuff goes to syslog automatically now
Paul Martin
@Spazholio
Jun 14 2016 03:01
Looks like it just reinstalled:
(pi) [pancreas: ~/src/Adafruit_Python_BluefruitLE] $ sudo python setup.py develop
running develop
running egg_info
writing Adafruit_BluefruitLE.egg-info/PKG-INFO
writing top-level names to Adafruit_BluefruitLE.egg-info/top_level.txt
writing dependency_links to Adafruit_BluefruitLE.egg-info/dependency_links.txt
reading manifest file 'Adafruit_BluefruitLE.egg-info/SOURCES.txt'
writing manifest file 'Adafruit_BluefruitLE.egg-info/SOURCES.txt'
running build_ext
Creating /usr/local/lib/python2.7/dist-packages/Adafruit-BluefruitLE.egg-link (link to .)
Adafruit-BluefruitLE 0.9.0 is already the active version in easy-install.pth

Installed /home/pi/src/Adafruit_Python_BluefruitLE
Processing dependencies for Adafruit-BluefruitLE==0.9.0
Finished processing dependencies for Adafruit-BluefruitLE==0.9.0
(pi) [pancreas: ~/src/Adafruit_Python_BluefruitLE] $
Ben West
@bewest
Jun 14 2016 03:03
bewest@bewest-MacBookPro:~/Documents/foobar$ openaps use cgm GetFirmwareHeader
{ 
  "BLEVersion": "1.0.0.24",
  "DexBootVersion": "3",
  "BLESoftDeviceVersion": "1.0.52.1",
  "RFVersion": "1.0.0.28",
  "ApiVersion": "2.4.0.0",
  "ProductName": "Dexcom G4 Receiver with Share",
  "SoftwareNumber": "SW10586",
  "TestApiVersion": "2.9.0.0",
  "SchemaVersion": "1",
  "PortVersion": "4.6.4.59",
  "FirmwareVersion": "3.0.1.026",
  "ProductId": "G4ShareReceiver"
}bewest@bewest-MacBookPro:~/Documents/foobar$
on mine, it sends all that extra stuff to syslog
only the data gets printed now
hmmm
Paul Martin
@Spazholio
Jun 14 2016 03:04
How did you facilitate that? As long as stuff still can go to my /var/log/openaps/loop.log, I’m fine with THIS stuff spitting to syslog...
Without all my BT negotiation cruft, my Firmware info is identical to yours.
Paul Martin
@Spazholio
Jun 14 2016 03:11
Also, this is the way I’m zoning that data: openaps use tz rezone --date system_time --date display_time test.json. Does that look accurate? It reports accurate data, but is it the “right” way?
Paul Martin
@Spazholio
Jun 14 2016 03:17
And I ran openaps use ns shell format-recent-type tz entries test-zoned.json and while it seems to go (no errors) then nothing shows up in NS.
Paul Martin
@Spazholio
Jun 14 2016 03:28
Alright, I figured out how I’m a bonehead (that last command was just to format, not upload), but when I finally decided to ACTUALLY upload something, I get the following:
Pete Schwamb
@ps2
Jun 14 2016 03:29
@bewest On the HW vs SW front and device makers, I think that approach makes a lot of sense. Users expect a smarter, more humane interface to their systems nowadays, and that’s a high bar for a device manufacturer to keep up with, when phone manufactures are spending billions on it. For example, Riley is really frustrated with the UI on her pump; if she’s really high for some reason, and has to scroll all the way up to do a correction. She’s jealous of her friend who has a tslim, and just punches in the numbers.
Paul Martin
@Spazholio
Jun 14 2016 03:29
(also, there’s no raw data showing in NS - should there be?)
Pete Schwamb
@ps2
Jun 14 2016 03:31
Of course I immediately thought of how we could put a complete pump UI on the phone after we crack omnipod.
Scott Leibrand
@scottleibrand
Jun 14 2016 03:35
:-)
I'm beginning to have a hard time keeping up with all y'all. When that happens it'll be hopeless. ;-)
Ben West
@bewest
Jun 14 2016 03:36
@ps2 but riley will appreciate the scrolling UI when she's driving or riding a bike
@ps2 tandem demands you pull over and look at it
I did meet a like minded 9 year old or so at maker faire
makes sense though to have touch when it makes sense
and scrolling when needed
shouldn't be either or
@Spazholio that's a little different from what I suggested
Paul Martin
@Spazholio
Jun 14 2016 03:40
Which part? =)
Ben West
@bewest
Jun 14 2016 03:40
the interface to ns shell upload-non-empty-type is different from ns shell upload
upload-non-empty-type is specially prepared to handle timeseries data like entries/treatments
it actually calls upload along with sanitizing things underneath
oh
hmm
sorry
hmm
Paul Martin
@Spazholio
Jun 14 2016 03:41
Ok, so openaps use ns shell upload entries test-final.json is incorrect?
Ben West
@bewest
Jun 14 2016 03:41
lemme check that
Paul Martin
@Spazholio
Jun 14 2016 03:41
k
Ben West
@bewest
Jun 14 2016 03:41
openaps report show nightscout/recent-missing-entries.json --cli
openaps use ns shell format-recent-type tz entries monitor/glucose.json
so that's the prep step
upload
Paul Martin
@Spazholio
Jun 14 2016 03:42
Wait…does that assume that the glucose.json is zoned or no?
Ben West
@bewest
Jun 14 2016 03:42
yes
it's already zoned
Paul Martin
@Spazholio
Jun 14 2016 03:42
Ok, cool.
Ben West
@bewest
Jun 14 2016 03:42
$ openaps report show nightscout/uploaded-entries.json  --cli
openaps use ns shell upload-non-empty-type entries.json nightscout/recent-missing-entries.json
entries.json is the trick there
John Males
@johnmales
Jun 14 2016 03:42
Hi All, great to see there is some movement or at least discussion with the hardware makers. Can I ask a question about oref0 meal? Specifically what does mealCOB mean? I keep getting a result where meal COB is in the 300's whilst the carbs and boluses are giving results that I would expect. I realize I don't understand what mealCOB actually means.
Ben West
@bewest
Jun 14 2016 03:42
entries == txt
entries.json == json
Paul Martin
@Spazholio
Jun 14 2016 03:43
Here’s the dumb question - how do I get nightscout/recent-missing-entries.json?
Ben West
@bewest
Jun 14 2016 03:44
see previous

```
openaps report show nightscout/recent-missing-entries.json --cli
openaps use ns shell format-recent-type tz entries monitor/glucose.json

```

Paul Martin
@Spazholio
Jun 14 2016 03:44
Oh damn it. I missed that. Just looked at the second line.
Ben West
@bewest
Jun 14 2016 03:44
yeah, that's why I showed both :-)
when I showed the second one, it wouldn't make sense :-)
you just leveled up, I believe
that's actually good question
David Cintron
@loudestnoise
Jun 14 2016 03:45
:100:
Ben West
@bewest
Jun 14 2016 03:46
this is the kind of thing I'm trying to get in the walk through/docs
then a walk through of what they mean/how to use them
we need some dry explanations of how each tool works
so "how do I get that file like that?"
some tutorial needs to go kind of back and forth iteratively :)
not quite a walk through
more of a tut
Paul Martin
@Spazholio
Jun 14 2016 03:47
I’m so close…do you know how I can get rid of the BT connection output?
Ben West
@bewest
Jun 14 2016 03:47
ah
well you can use repor
report just saves the json
instead of >
ah ok well your adafruit is up to date
suppose we should check openxshareble
Paul Martin
@Spazholio
Jun 14 2016 03:48
Let me see if adding as a report instead of a use helps out with that.
Ben West
@bewest
Jun 14 2016 03:49
oh...
ok
hmm
please hold :-)
well
you have openxshareble installed via git?
if you do git checkout dev and git pull origin it should work
maybe with another sudo python setup.py develop
I'll release that presently
Paul Martin
@Spazholio
Jun 14 2016 03:52
Done. And now it should be good just as a use instead of only as a report?
Paul Martin
@Spazholio
Jun 14 2016 03:58
Indeed it does. Neat.
Also, sometimes it takes a LONG time to run. I’m only pulling one hour via oref0_glucose —hours 1 and it took a little longer than 30s. I imagine that’s just one of those “hey, sometimes it takes a while to run” things?
Oh, wait, just so I’m clear:
openaps report add nightscout/recent-missing-entries.json JSON ns shell format-recent-type tz entries monitor/glucose.json
is incorrect, and shoud be:
openaps report add nightscout/recent-missing-entries.json text ns shell format-recent-type tz entries monitor/glucose.json
Did I understand you correctly earlier?
Paul Martin
@Spazholio
Jun 14 2016 04:05
Because openaps use ns shell format-recent-type tz entries <filename> spits out JSON, so I may be misunderstanding.
Scott Leibrand
@scottleibrand
Jun 14 2016 04:18
@johnmales which oref0 branch are you using?
John Males
@johnmales
Jun 14 2016 04:18
Dev
I haven't tried AMA as yet
Scott Leibrand
@scottleibrand
Jun 14 2016 04:20
The mealCOB in dev is informational only. Supposed to be based on a fixed 30g/hr carb decay.
John Males
@johnmales
Jun 14 2016 04:20
So it should be equal to or smaller than carbs?
Scott Leibrand
@scottleibrand
Jun 14 2016 04:20
What does your full meal.json look like? And what were the recent meal carbs and boluses?
correct.
John Males
@johnmales
Jun 14 2016 04:21
I'll log in remotely to check but something like carbs 40, boluses 3.0, mealCOB 310 is typical
Scott Leibrand
@scottleibrand
Jun 14 2016 04:22
Yeah, that sounds buggy. If you want to we can debug.
John Males
@johnmales
Jun 14 2016 04:22
Sure
Scott Leibrand
@scottleibrand
Jun 14 2016 04:22
Or you can just skip to AMA and see if it does the same thing. :-)
Paul Martin
@Spazholio
Jun 14 2016 04:23
@bewest Can I see the report show —cli for your monitor/glucose.json?
John Males
@johnmales
Jun 14 2016 04:23
Ok :) will try.
John Males
@johnmales
Jun 14 2016 04:31
the [carbhistory] entry in oref0 meal - if carbs are being entered into the pump is this necessary? If so which report would give you this data?
Having now installed oref0 advanced-meal-assist branch, meal.json is giving me this output: {"carbs":7,"boluses":0.8} with no mealCOB, but this is on stale data.
Ben West
@bewest
Jun 14 2016 04:44
@Spazholio
bewest@bewest-MacBookPro:~/Documents/foobar$ openaps report show  --cli monitor/glucose.json 
openaps use tz rezone --date timestamp --date dateString --date start_at --date end_at --date created_at --date system_time --date display_time --astimezone raw-cgm/glucose-raw.json
bewest@bewest-MacBookPro:~/Documents/foobar$ openaps report show  --cli raw-cgm/glucose-raw.json
openaps use cgm oref0_glucose --threshold "100" -n "20" --hours "2.0"
I think the main thing there is the --astimezone switch
esp if you are fetching from ns
you need --date dateString for sure
--date display_time --date system_time
Paul Martin
@Spazholio
Jun 14 2016 04:45
Just trying to rejigger my entire loop since I’m now using BLE. What’s the -n flag?
Also - am I supposed to be seeing extra info in NS now that I’m uploading stuff from openxshareble?
And should my reporter for the nightscout/recent-missing-entries.json report be text or JSON?
Paul Martin
@Spazholio
Jun 14 2016 04:51
Because my loop keeps hiccuping on generating that report. =/
Scott Leibrand
@scottleibrand
Jun 14 2016 04:52
@johnmales you don't need carbhistory if you enter carbs via bolus wizard into pumphistory.
John Males
@johnmales
Jun 14 2016 04:52
ok thanks @scottleibrand
John Males
@johnmales
Jun 14 2016 05:07
Comment about mealCOB made in docs & PR'ed
Paul Martin
@Spazholio
Jun 14 2016 05:14
@scottleibrand So auto-sens needs the last 24-28 hours to make a decision, right? So if I’m moving to openxshareble, if I understand it correctly, I’d need to use NS data. Because there is NO WAY I can pull that much data from the G4 via BLE in a reasonable amount of time….right?
Brian Bonner
@itsn0tm3
Jun 14 2016 05:16
Anyone ever have shits and fits getting the pumphistory out of their pump for no reason?
Paul Martin
@Spazholio
Jun 14 2016 05:17
Yep.
Brian Bonner
@itsn0tm3
Jun 14 2016 05:17
I'm in a hotel in SLC and I've been having problems with it for the last 2-3 days and I can't figure out WHY
Paul Martin
@Spazholio
Jun 14 2016 05:17
The 24h one is a LOT worse for me personally.
Brian Bonner
@itsn0tm3
Jun 14 2016 05:17
I actually changed it to 14h
and still having problems with it
at home in ROC I ><NEVER< have problems
Paul Martin
@Spazholio
Jun 14 2016 05:18
Maybe there’s lots of interference in the hotel?
Brian Bonner
@itsn0tm3
Jun 14 2016 05:18
nah I have the same problem outside too
ugh
I hooked up a rig to both sides of the bed
Paul Martin
@Spazholio
Jun 14 2016 05:19
Grab your pump and your rig, then wrap yourself in outdoor screening. Ghetto Faraday cage, yo. =)
Brian Bonner
@itsn0tm3
Jun 14 2016 05:19
and the other one can't even get the model out of the pump before giving up
Paul Martin
@Spazholio
Jun 14 2016 05:19
What are you using? TI stick? Carelink?
Brian Bonner
@itsn0tm3
Jun 14 2016 05:19
LOL! I'm close to considering that unfortunaetly
ti stick
Paul Martin
@Spazholio
Jun 14 2016 05:19
Huh.
That thing’s a beast.
Brian Bonner
@itsn0tm3
Jun 14 2016 05:19
0.8 on both sticks and the latest mmeowlink
the error kind of looks like this:
Jun 14 01:18:48 micheyl Invalid Packet Received - '' - retrying: 1 of 3
Jun 14 01:18:48 micheyl Invalid Packet Received - '' - retrying: 2 of 3
Jun 14 01:18:49 micheyl Invalid Packet Received - '' - retrying: 3 of 3
Jun 14 01:18:51 micheyl openaps: }settings/settings.json  raised  'Pump' object has no attribute 'model'
but it usually shits out on the pumphistory
but it'll randomly do it in different places
Paul Martin
@Spazholio
Jun 14 2016 05:20
That…is nothing like any error I’ve ever seen.
Kinda wouldn’t mind getting that kind of info in my logs, to be honest. =)
Brian Bonner
@itsn0tm3
Jun 14 2016 05:21
heh that's from the master mmeowlink code
I've got PLENTY of different error codes
Paul Martin
@Spazholio
Jun 14 2016 05:21
That’s weird. I’m pretty sure that’s what I’m running. I’ll have to look into that.
Brian Bonner
@itsn0tm3
Jun 14 2016 05:21
ones I've never seen before at home
other one looks like this:
Jun 14 01:15:54 livingroomaps Timed out or other comms error - Received an error response Timeout - retrying: 1 of 3
Jun 14 01:15:54 livingroomaps Timed out or other comms error - Received an error response Timeout - retrying: 2 of 3
Jun 14 01:15:56 livingroomaps Timed out or other comms error - Received an error response Timeout - retrying: 3 of 3
Jun 14 01:15:58 livingroomaps openaps: monitor/pumphistory.json  raised  need more than 0 values to unpack
Paul Martin
@Spazholio
Jun 14 2016 05:22
Just because you’re outside, doesn’t mean you’re not encountering interference. If you’re connected enough, I’d say take the rig outside, and go….I dunno, get in the car and get away from the hotel for a bit. Like a mile or something. See if you see the same errors then?
Brian Bonner
@itsn0tm3
Jun 14 2016 05:23
saw the same thing in a conference room in an entirely different hotel for a presentation I was at
but that was an intel edison with ti rig
Paul Martin
@Spazholio
Jun 14 2016 05:23
Man, that’s just peculiar.
Brian Bonner
@itsn0tm3
Jun 14 2016 05:24
I haven't used the wait for silence code yet, I'm thinking about trying that out tmw after work and seeing if that helps at all
heres another weird error I'm getting
Jun 14 01:24:13 micheyl Invalid Packet Received - '' - retrying: 1 of 3
Jun 14 01:24:13 micheyl Invalid Packet Received - 'Error decoding FourBySix packet' - retrying: 2 of 3
Jun 14 01:24:15 micheyl Timed out or other comms error - Received an error response Timeout - retrying: 3 of 3
Paul Martin
@Spazholio
Jun 14 2016 05:26
Assuming you’re running a preflight? mmtune?
Brian Bonner
@itsn0tm3
Jun 14 2016 05:26
yup
Paul Martin
@Spazholio
Jun 14 2016 05:26
What does it say? Everything looks good, etc?
Brian Bonner
@itsn0tm3
Jun 14 2016 05:27
yep left rig is getting -57 signal
left rig is getting -94
err right rig
and it's only like 6 feet away from me
the left rig is closer to me at the moment tho
I do see a door for the air conditioning/heater unit
wonder if that's putting off interference or something
Paul Martin
@Spazholio
Jun 14 2016 05:28
Are they interfering with each other? What happens if you turn one off? I mean, I can’t imagine that’s it since Scott and Dana run 6 in one apartment, but...
Brian Bonner
@itsn0tm3
Jun 14 2016 05:29
nah I was only running 1 rig before
and it was failing like no tomorrow
it makes it every once in a while
so I figured I'd set up the rig and the timing is just off enough to make sure they don't interfere with each other
if it is interference, I can't imagine what would cause THIS much interference
amusingly this is making me want to make an interference tester out of a pi zero
and use it to test the hotel room for interference before I take posession of it
"Oh I'm just looking for electronic bugs before I take posession, those damn things are loud!"
Paul Martin
@Spazholio
Jun 14 2016 05:33
Yeah, no kidding. =)
I swear, if I get one more “Your CGM data is too old” error message, there’s gonna be property damage.
Unfortunately, it’ll likely be MY property, but those are just details.
Brian Bonner
@itsn0tm3
Jun 14 2016 05:40
still working on trying to get the BT to work?
Paul Martin
@Spazholio
Jun 14 2016 05:42
No, actually. With some help from @jyaw and @bewest I managed to get it working on the Edison, and then the Pi3. Turns out, I needed to use a BT dongle for the Pi3 - the onboard BT just isn’t working for some reason. It can see the Dexcom, meaning the whole BLE thing shouldn’t be the issue, but it won’t talk to it for some reason.
Brian Bonner
@itsn0tm3
Jun 14 2016 05:43
oh great!
sounds like the time zone might be off on the rig then if you're getting the error that much?
(but then i'm sure you've already checked that)
Paul Martin
@Spazholio
Jun 14 2016 05:44
No, that’s the funny thng - the date on the rig is the one that’s always right. Plus, I invoke ntp at regular intervals to make sure that’s not the issue. No, it seems that when I poke around at my glucose source (change it from NS to wired CGM to openxshareble) it will throw that error for an hour or so before settling down. It “feels” like it’s working out bad data, but I have no idea how or if that’s even the case.
Brian Bonner
@itsn0tm3
Jun 14 2016 05:45
thats weird
sorry to talk and run but I need to hit the sack, got an early start in the morning
Paul Martin
@Spazholio
Jun 14 2016 05:46
Same here. Good luck with the interference issue. Report back if you discover what it is? I’m definitely curious now.
Brian Bonner
@itsn0tm3
Jun 14 2016 05:47
will do, I'm hoping the wait for silence will help put some light on it since it works about half of the time on the left rig so has to be SOMETHING
nigth!
err night!
Scott Leibrand
@scottleibrand
Jun 14 2016 06:01
@Spazholio we disable autosens and AMA when we go completely offline and can't get BG data from NS. I suppose you could also use slightly (but not too) old NS data for autosens...
Paul Martin
@Spazholio
Jun 14 2016 06:01
Ok, I think I figured out the “Your CGM data” error issue out. It seems that it’s looking at my NS data, and calling that my “CGM” data. But if the loop doesn’t run, or the BLE doesn’t pull data (which is seems to work 2/3 of the time) and since that’s how I’m getting data into NS, there’s going to be gaps until the next iteration, when the BLE data uploads properly and it re-pulls from NS, so there isn’t a gap anymore. There has to be a way to mitigate this, and I’m sure I’m just overlooking it since it’s late and I’m tired.
I gotta be honest, I’d have to look through my openaps.ini just to figure out how the heck to disable auto-sens. It’s been so long since I set it up initially. =)
Scott Leibrand
@scottleibrand
Jun 14 2016 06:02
We actually have two loops on our openxshareble rigs: one just refreshes CGM data, and the other does pump stuff.
We disable it by passing a shorter glucose.json. ;-)
@jasoncalabrese is doing something similar with a CGM-only loop.
Paul Martin
@Spazholio
Jun 14 2016 06:02
I was JUST thinking about doing that - setting up 2 loops. How often do you pull from the CGM, and how many hours do you pull?
Scott Leibrand
@scottleibrand
Jun 14 2016 06:03
There is a cgm-loop branch in my openaps-sh.
Paul Martin
@Spazholio
Jun 14 2016 06:03
k
Scott Leibrand
@scottleibrand
Jun 14 2016 06:03
Youd have to do more to upload too.
Paul Martin
@Spazholio
Jun 14 2016 06:03
How do you mean?
Scott Leibrand
@scottleibrand
Jun 14 2016 06:03
We don't upload. You'd need to add that.
(Upload CGM data)
Paul Martin
@Spazholio
Jun 14 2016 06:05
Oh, gotcha. The reports are already there for that, I’d just have to call them differently, outside of the regular loop.
Oh, you pull CGM data every minute. Nice.
Do you find that that affects the battery of the Dexcom much, if at all?
Scott Leibrand
@scottleibrand
Jun 14 2016 06:08
We don't use it enough to notice, but I doubt it would.
We plug in the Dex every night anyway to the bedside Pi.
Paul Martin
@Spazholio
Jun 14 2016 06:09
And last question before I pass out - how many hours do you pull from the CGM? I was thinking 2-3 seemed like a decent tradeoff for the time it takes.
Scott Leibrand
@scottleibrand
Jun 14 2016 06:11
Yeah, check the script, but I think it's something like that. Maybe less.
Ben West
@bewest
Jun 14 2016 07:57
@Spazholio I updated openxshareble master so that the logging goes to syslog
any kind of git pull you do in openxshareble should update :-)
Matt Pazoles
@Pazoles
Jun 14 2016 12:06
Hey all - in oref0 determine basal, I noticed that it won't issue a new temp basal if the current temp basal is within +/-0.1 of the recommendation. This seems like it's too large of a step for @PazolesBrooke_twitter, but before I start tinkering I just wanted to check in to see what the reason behind that is so I can make sure I update it safely
Chris Oattes
@cjo20
Jun 14 2016 13:04
This is weird O.o suddenly iter_glucose isn't returning sensor values
It stopped at 1319 and hasn't read one since. It finds the calibration I did at 1350 fine
Andrew Baugh
@baughaw
Jun 14 2016 13:06
@cjo20 Are you using Dex or Enlite?
Chris Oattes
@cjo20
Jun 14 2016 13:06
enlite
Andrew Baugh
@baughaw
Jun 14 2016 13:06
I have that problem sometimes with Enlite. I would check the battery level and also recalibrate again
Chris Oattes
@cjo20
Jun 14 2016 13:07
it isn't reporting as low battery
Andrew Baugh
@baughaw
Jun 14 2016 13:08
just curious which pump are you using? I had a lot of issues with a 522 doing this same thing. I switched to a 523 and doesnt seem to happen as much
Chris Oattes
@cjo20
Jun 14 2016 13:08
554
Andrew Baugh
@baughaw
Jun 14 2016 13:09
i would try to recalibrate again...
Chris Oattes
@cjo20
Jun 14 2016 13:09
It's showing values ok on the pump screen though
Andrew Baugh
@baughaw
Jun 14 2016 13:13
yeah it will, this is just a known issue. Calibrating is usually the only way to fix it
Chris Oattes
@cjo20
Jun 14 2016 13:14
hmm, it stopped roughly an hour before I was due to calibrate anyway. I wonder if they're just calculating the time wrong internally
Paul Martin
@Spazholio
Jun 14 2016 16:27
@bewest So I’m having a sorta problem with openxshareble. Hoping you can help. When I went to bed last night, I integrated it into my loop in the way that @scottleibrand recommended - outside the main loop inside a loop of its own. Every minute, it will attempt to pull glucose data from the Dexcom, but the BLE connection is flaky as hell. It would regularly hang for 10+ minutes, never pulling any data, thereby crashing my loop. So much so that I had to build some kill functionality into the loop, but I suspect it’s not working, since it hasn’t once pulled data in the last 8 hours. Help?
Ben West
@bewest
Jun 14 2016 16:29
hmmm
Paul Martin
@Spazholio
Jun 14 2016 16:31
I noticed that before integrating - sometimes it would start to run, but the BT icon would never actually show. And the process would just hang, so I’d have to kill it and start again.
Same thing right now - I have the loop running, there were no other BLE processes running, it SAYS it’s attempting to pull data, but I’m getting no BT icon, and no data. =/
Ben West
@bewest
Jun 14 2016 16:33
mine is a little more simple
I have do-everything alias
do-everything ! bash -c "(openaps preflight && openaps gather-clean-data && openaps do-oref0 && openaps enact-oref0); openaps report-nightscout"
Paul Martin
@Spazholio
Jun 14 2016 16:34
Well, I had it integrated into the main do-everything but it broke the loop whenever there wasn’t data to work from. Which was a lot. All it had to do was miss one or two BLE pulls, and the entire loop would crash for the next iteration or so.
Ben West
@bewest
Jun 14 2016 16:36
when doing oref0_glucose you mean?
Paul Martin
@Spazholio
Jun 14 2016 16:36
Correct.
And I’m only pulling 2 hours.
Ben West
@bewest
Jun 14 2016 16:37
ah, that was my next question
Paul Martin
@Spazholio
Jun 14 2016 16:39
I tried 24 hours once. That was funny.
Also - and I hate to harp on this - but I’m getting no extra data in NS? No unfiltered/raw/whatever. And that was kinda a large part of the reason I was doing this.
Ok….so I restarted BT and that fixed the problem of me not pulling data in the last 8 hours. I think I may need to do that periodically. =/
Matt Pazoles
@Pazoles
Jun 14 2016 16:46
Ran into an interesting issue switching from an iphone hotspot to home wifi (may have been the same issue @PazolesBrooke_twitter ran into last week) - APS connects to home wifi (I can ssh into it), but can't connect to the pump. RPI was also behaving weirdly - not very responsive. Rebooted and it was fine. Anyone else see this before?
Paul Martin
@Spazholio
Jun 14 2016 16:47
I haven’t. I know I have some issues switching WiFi networks sometimes, but I’ve never seen it affect pump comms.
When you say you can’t connect to the pump, what are you seeing in the logs?
Garrett Webb
@garetis
Jun 14 2016 16:53
So, I'm looking to switch from Medtronic CGM to Dexcom. Is there full support for the G5 at this point? Any reason I should get the G4 over the G5? I'm an Android user...
Sarah Hodson Grady
@shgrady
Jun 14 2016 16:58
@garetis I heard they said no Android until at least 2017 on the G5 officially, but I read some people were beta testing the G5 with Xdrip
The G4 has a slimmer profile, longer battery, and as a female, what is one more device(the receiver) in my purse, and stays paired well with my S7
Garrett Webb
@garetis
Jun 14 2016 17:01
Thanks @shgrady , I think I am not going to try to use my phone for uploading to NS, and will be using my rig instead. Would that make a difference?
Sarah Hodson Grady
@shgrady
Jun 14 2016 17:04
the rest in here are an expert, I am actually going to hook an older G4(without Share) to my rig via wire, and use my other G4 with Share full time. My understanding is that the G4 with Share is radio to transmitter, and bluetooth to phone, G5 is bluetooth both ways or skip receiver. Since X5 still in Beta with xDrip, I am ordering another G4 when my warranty is up because I am impatient, you can always upgrade the receiver firmware and switch to G5 transmitters (6 mos later when your transmitter battery dies)
they actually sent me two G5 transmitters, but I did not upgrade my receiver's firmware because they are twice the height and I have Android.
David Cintron
@loudestnoise
Jun 14 2016 17:09
@garetis I use a G5 transmitter with my G4 with Share receiver (that I upgraded to the G5 software) connected to my RPi rig. I also have the G5 paired to my iPhone and at the moment that’s how I pull in values into my Nightscout. I’m not sure if anyone has added the documentation for using G5 on the docs, but I got mine working by reading the conversations here on gitter
Garrett Webb
@garetis
Jun 14 2016 17:15
So, the "share receiver" is the device that you read your glucose off of? So @loudestnoise , your G5 transmitter is being sent to both your receiver and your iPhone, and you are uploading your NS data through your phone?
Chris Oattes
@cjo20
Jun 14 2016 17:16
@scottleibrand Why is basal truncated to 2 d.p. here: https://github.com/cjo20/oref0/blob/dev/lib/determine-basal/determine-basal.js#L26
Garrett Webb
@garetis
Jun 14 2016 17:17
@loudestnoise And the receiver is connected to the rig so it has the glucose data to act upon, but you aren't syncing data from the rig->NS or NS->rig ?
David Cintron
@loudestnoise
Jun 14 2016 17:17
Yeah, sorry it’s confusing because dexcom released multiple versions of the G4 receiver. There’s regular G4 receiver and there’s G4 with Share. The share part is what let you send your readings to your iPhone. Then G5 came out and you could either get a new G5 Receiver or upgrade your G4 with Share receiver as it already had the bluetooth bits in it, just needed new softweare
I’m using Nightscout Bridge, which pulls the values directly from dexcom’s servers
I haven’t got it working to sync up readings from my Rig > Nightscout and correct I’m not using Nightscout > rig
The particular reason I use the receiver is because I can rely on having CGM data in the absence of an internet connection
I do want to eventually get the rig sending “missing” entries to Nightscout because the iPhone sometimes will have gaps, and for whatever reason the receiver seems to not miss them
Garrett Webb
@garetis
Jun 14 2016 17:20
Ok. Two things: I've seen small things here and there about the bridge, but haven't seen it in the docs (yet?). Could you throw me a link of where I could find it? Second thing, which you may have just answered, but the reason that the G4 receiver is connected to the rig is when you are absent of wifi data, it is basically a backup?
David Cintron
@loudestnoise
Jun 14 2016 17:21
Bridge is part of the current version of Nightscout. I can’t recall when it was added, but you just have to enable it like you do other plugins
and then you include your BRIDGE_USERNAME and BRIDGE_PASSWORD
Here’s some documentation on the Nightscout site - http://www.nightscout.info/wiki/welcome/nightscout-with-ios-and-dexcom-share
@garetis are you doing Nightscout on Heroku or Azure?
Garrett Webb
@garetis
Jun 14 2016 17:23
Ok, I guess I didn't look too hard at the NS info with dexcom because I wasn't set on using it yet.
I'm using Azure
set all that up yesterday
But I was going to use my old 712 and there is no way to get my medtronic cgm to NS (or the pump), so moving to dexcom
But am also in the process of getting a 722 or usable 723
David Cintron
@loudestnoise
Jun 14 2016 17:25
Ok, cool yeah you would add the bridge stuff in your Azure application settings
but that would require you have an iPhone, which sounds like you don't
Best of luck on obtaining another pump. There were some leads posted here a few days ago
Also talk to @baughaw I think he has some to sell
Garrett Webb
@garetis
Jun 14 2016 17:29
I do have some 6s sitting around that I have, but I am an Android user. So the bridge is specifically used with the iPhone? Is the -curl thing for using the rig to update to NS?
David Cintron
@loudestnoise
Jun 14 2016 17:29
Yeah, unfortunately all of that requires iOS
You either need the Receiver with Share to iOS running dexcom share 2 app or G5 with G5 Mobile app
Garrett Webb
@garetis
Jun 14 2016 17:30
I actually bought a 723 with firmware 2.5A, because I saw that in the documentation, but failed to realize that it was mentioning that was the version it couldn't be, not the last one it could be. So, now I'm trying to contact sellers to see if they are able to swap -- essentially an upgrade for them
Dana Lewis
@danamlewis
Jun 14 2016 17:31
:(
@garetis let us know if we can improve the docs to make that more clear
Garrett Webb
@garetis
Jun 14 2016 17:31
My intention is to connect the dexcom receiver directly to the rig and use it to upload to NS, is that something that people do?
Dana Lewis
@danamlewis
Jun 14 2016 17:31
or submit a PR to fix directly
Garrett Webb
@garetis
Jun 14 2016 17:31
@danamlewis haha, it's all good =)
David Cintron
@loudestnoise
Jun 14 2016 17:31
@garetis Yes, people do that for sure
and I think that would be a great setup
I guess the only thing it would require is your rig have internet connection to upload to Nightscout
Garrett Webb
@garetis
Jun 14 2016 17:32
@danamlewis Yeah, was thinking about it as soon as I make some headway
David Cintron
@loudestnoise
Jun 14 2016 17:32
but with the receiver connected to rig you can still loop
even if you don’t have Internet
Garrett Webb
@garetis
Jun 14 2016 17:32
I have unlimited data with t-mobile, so I don't think it'll be too much of an issue
David Cintron
@loudestnoise
Jun 14 2016 17:32
:+1:
I have plenty of data too, but hotspot is er spotty sometimes for me and of course there’s time when cellular signal ain’t great
Garrett Webb
@garetis
Jun 14 2016 17:35
So, just to clarify, but the -curl device is the one that uploads rig data to NS? I remember seeing that, but when I have gone back looking a couple times I haven't seen it
Also, this may be a silly question, but there's no way to connect the Bluetooth from the rig (I'm using a Pi3) directly to the G5?
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 14 2016 17:37
Hello everyone, I just got started (found a 722 and ordered my first Rig parts). I am so encouraged by all the great work you all are doing.
Paul Martin
@Spazholio
Jun 14 2016 17:37
Congrats @jlucasvt. Welcome! =)
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 14 2016 17:37
@garetis (are you in GA?) you've got the tech bee...
Thanks @Spazholio .
David Cintron
@loudestnoise
Jun 14 2016 17:38
@garetis I think you want to use the ns device that can help you upload your BG values
and there are others ways to get the BG off the dexcom transmitters like xdrip and some magical thing called openxshareble but you can ask @Spazholio how easy that is to setup
I’ve not done it, but I know others have
Paul Martin
@Spazholio
Jun 14 2016 17:41
@bewest What are your thoughts on keeping a persistent BT connection open from the rig to the Dexcom? I think it’s the negotiation/connection that gets things junked up every once in a while. Rather than opening/closing (resulting in me having to restart BT every once in a while) maybe we could just make the connection persistent, like the Share app does?
@garetis It wasn’t all that easy. But if you’re using a Pi, you can just plug the Dexcom into the rig. That works like a champ.
David Cintron
@loudestnoise
Jun 14 2016 17:43
+1 for receiver to rig
Dana Lewis
@danamlewis
Jun 14 2016 17:44
@garetis I’d encourage you to PR small things like that as soon as you encounter them, so it doesn’t happen to the next person :)
Garrett Webb
@garetis
Jun 14 2016 17:45
Easy sounds like a good way to start... I'll wait for my feet to get more wet before I get off the beaten path -- but at least now I am starting to understand the conversations more. I'm thinking I'll probably end up wearing both metronic and dexcom cgm for a little while
@danamlewis yeah, I thought that earlier when I said it, and internally committed to doing it later today
Dana Lewis
@danamlewis
Jun 14 2016 17:45
:+1:
:) :clap:
Paul Martin
@Spazholio
Jun 14 2016 17:48
"but at least now I am starting to understand the conversations more.” <— this part is sorta key. When I was starting out, I had no clue what people were talking about, but at the stage I’m at now, I remember those conversations from the beginning and thinking, “Man, I wish I knew what they were talking about then, because I think I’m running into the same issue now…"
David Cintron
@loudestnoise
Jun 14 2016 17:48
mmmmhm
it’s a secret language
but I’m glad if I can help people figure that out, and let others help with the more technical questions
Lynne Butts
@flutelynne_twitter
Jun 14 2016 17:57
I have another noob question: When setting up bitbucket, and it shows me the ' $ ......' lines, do I ssh from my computer to the pi and put those command lines in? I really don't want to mess anything up.
Also, I contacted Tendure battery co., and asked if it were possible to charge the tendure A3 with another battery pack. They said it should be ok, but I am wondering if a surge could happen and corrupt the sd card...
David Cintron
@loudestnoise
Jun 14 2016 18:03
I’ve done that before @flutelynne_twitter. Charge one battery with another. Didn’t have any issues.
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 14 2016 18:10
Google.com
oops sorry.
Lynne Butts
@flutelynne_twitter
Jun 14 2016 18:11
@loudestnoise Thanks! Gonna have a 20hr+ day, and would like to use the rig.
Garrett Webb
@garetis
Jun 14 2016 18:11
@flutelynne_twitter What hardware/system/rig are you using, and how long does your A3 last?
Lynne Butts
@flutelynne_twitter
Jun 14 2016 18:13
The A3, when at home, seems to last at least 10 hrs. I end up swapping out with a 6000 battery while the A3 is charging for me to use overnight. I am waiting for the delivery of a second pi2 so I could just plug it in at night.
I have the RPi2, CareLink stick, G5 receiver. I use a larger Lowepro bag, not the 20, because my cords are long, and I use the regular Dexcom charging cord. I end up putting the Carelink on the outside of the bag for better room and reception. Seems to work ok, although I do have to unplug and replug the CL stick a few times a day.
David Cintron
@loudestnoise
Jun 14 2016 18:16
@flutelynne_twitter does the battery you use not let you charge the battery while also powering your Pi?
Lynne Butts
@flutelynne_twitter
Jun 14 2016 18:17
Well, I tried charging the Zendure while using the rig at night, and I wasn't having good connections to the CL stick.
David Cintron
@loudestnoise
Jun 14 2016 18:18
Hmm, I got this one called the Moku - https://amzn.com/B00FGN1OBU
Lynne Butts
@flutelynne_twitter
Jun 14 2016 18:19
Then I though because I was also charging my phone, there may be a problem with that (some people on some pages used to talk about the dexcom receiver placement and not getting readings when other devices were charging).
How many hours do you get from that battery?
David Cintron
@loudestnoise
Jun 14 2016 18:19
it’s 9000mAh and lasts most of the day, but if I’m going to be out later than normal I charge it at work to keep me going until later
I’m thinking I easily get 12-14 hours charge on it
Lynne Butts
@flutelynne_twitter
Jun 14 2016 18:20
Hmm. The Zendure is 10000, and I guess that I am chicken to let the battery get down to 2 lights (25%-?).
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 14 2016 18:21
I am putting together my parts for a Pi2 Rig with Carelink as a starting point (looking at Eidson/TI when I get the reference build done) Is the carelink range really that bad? Does it need to be right next to pump or will it work well overnight on bed stand?
Lynne Butts
@flutelynne_twitter
Jun 14 2016 18:21
Have you piggybacked that 9000 battery with a lower battery?
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 14 2016 18:23
Lynne Butts
@flutelynne_twitter
Jun 14 2016 18:25
The carelink seems to work better for me when I hang the lowepro bag on my bedpost, with the Carelink Stick (gray side facing out). I still get times when that drops connectivity. But for walking and doing stuff, I have it on my back in a small backpack, and it's great. On a band job, it worked the whole time being 2-3 ft. away from me. Only lost connection when I had to go up front for a feature.
David Cintron
@loudestnoise
Jun 14 2016 18:27
I pretty much leave mine in the Dashpoint 20 all the time
Lynne Butts
@flutelynne_twitter
Jun 14 2016 18:28
Wow! How does it all fit in there?
Chris Oattes
@cjo20
Jun 14 2016 18:28
Carelink sometimes seems ok from 1m+ away, other times it doesn't work when the pump is 30cm away
David Cintron
@loudestnoise
Jun 14 2016 18:29
Yeah, makes me want to switch to a TI stick
Here’s what my rig looks like, complete with hair tie holding it together
RPi_rig.jpg
Paul Martin
@Spazholio
Jun 14 2016 18:30
TI sticks are amazing. I’d highly recommend getting one.
@jlucasvt Can that battery be charged while in use? I don’t see anything on the page that indicates that it can.
Lynne Butts
@flutelynne_twitter
Jun 14 2016 18:34
I have a TI stick coming, but am not sure about setting that up. Would this page be ok to come to for that, or should I go to the oskar page? @loudestnoise amazing! I bought hair ties, but I still have the receiver in a cut-out dexcom case, original cord, long carelink cord, too.
I guess I was afraid of heat building up too much and something malfunctioning.
David Cintron
@loudestnoise
Jun 14 2016 18:35
Mine gets warm, but nothing crazy.
Lynne Butts
@flutelynne_twitter
Jun 14 2016 18:36
I will have to experiment.
Paul Martin
@Spazholio
Jun 14 2016 18:36
I haven’t had any heating issues, and I keep mine in a closed camera case. =)
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 14 2016 18:36
@Spazholio not sure on charge and deplete at the same time. I have one on order and will report back.
Brooke Pazoles
@PazolesBrooke_twitter
Jun 14 2016 18:36
I've been carrying my rig in a fanny pack. It's been getting warm as well.
Chris Oattes
@cjo20
Jun 14 2016 18:36
my pi gets hot against my leg when I put it in my pocket
Dana Lewis
@danamlewis
Jun 14 2016 18:37
@flutelynne_twitter for mmeowlink related stuff for TI you’ll want to ask in the mmeowlink channel for the best help so it doesn’t get lost in everything else here
Garrett Webb
@garetis
Jun 14 2016 18:37
I bought a TI stick and got it last week knowing I wanted to switch eventually, but am still working out all my other issues for the first rig... is there anyone I could send it to that would be willing to flash the software to it? Or is anyone near Dallas that could help out?
Ha, that works for me, too, haha
Lynne Butts
@flutelynne_twitter
Jun 14 2016 18:37
Mine still gets warm in the camera case, too. When sitting down for a little while, I open up the case to let it breathe/cool down. @danamlewis Thank you. That's what I thought.
Dana Lewis
@danamlewis
Jun 14 2016 18:37
:+1:
David Cintron
@loudestnoise
Jun 14 2016 18:38
http://imgur.com/a/HUCHa - album of my rig as it is now
@garetis I’m in Houston, but haven’t bought a TI Stick or debugger yet. I plan on eventually
Chris Oattes
@cjo20
Jun 14 2016 18:39
https://pbs.twimg.com/media/Ck6_Z9vWYAAcGQF.jpg:large my pi2, as 'packaged' as it gets. It tends to live on my desk or in my cargo trouser pockets
David Cintron
@loudestnoise
Jun 14 2016 18:40
nice ^ although I’d be afraid of breaking off the USB to that CareLink
Lynne Butts
@flutelynne_twitter
Jun 14 2016 18:40
@loudestnoise That was the dashpoint 20? Esp. like the pulp fiction pic that popped up, too. ; )
David Cintron
@loudestnoise
Jun 14 2016 18:40
Yeah, that’s the 20. It’s a real snug fit
Chris Oattes
@cjo20
Jun 14 2016 18:40
@loudestnoise I use a small usb cable when I'm going further than to the office and back
David Cintron
@loudestnoise
Jun 14 2016 18:41
Yeah, I’ve got a collection of 90 degree cables from Amazon
I just got a pair of 90 degrees microUSB to USB for battery to Pi
I was about to break the cable on the normal cables having it shoved in the camera bag
Chris Oattes
@cjo20
Jun 14 2016 18:42
After seeing how small Dana's rig is, I don't think I'm going to be able to stick to a pi. Someone just needs to design an edison shield with a c1111
Paul Martin
@Spazholio
Jun 14 2016 18:43
Shield?
I know @eluhrs has designed a nice little 3D printable case for an Edison/TI rig: https://www.thingiverse.com/thing:1525618
Chris Oattes
@cjo20
Jun 14 2016 18:45
I think they're more commonly referred to as blocks for edison maybe. The thing the compute module plugs in to, so you can plug in power / USB / etc.
Paul Martin
@Spazholio
Jun 14 2016 18:47
Oh! Ok, now I see what you mean.
I generally see that referred to as the “breakout board”. I’d kill for one with a C1111 integrated though. =)
Chris Oattes
@cjo20
Jun 14 2016 18:47
It should be possible to design a circuit that replaces the 4 rows of holes with a C1111 circuit (possibly even just copy over the reference design?)
Scott Leibrand
@scottleibrand
Jun 14 2016 19:13
@cjo20 had to round it to something, and 2 digits seemed reasonable, given stuff like https://github.com/cjo20/oref0/blob/dev/lib/determine-basal/determine-basal.js#L198
I would also be receptive to an argument for 3 digits there if that's what you're thinking...
Paul Martin
@Spazholio
Jun 14 2016 19:16
Has anyone developed a method to upload to share when not using the Dexcom Share app? Just so I can give my endo a login to Dexcom so they can see numbers, rather than having to teach them NS. =)
Chris Oattes
@cjo20
Jun 14 2016 19:17
@scottleibrand yeah, given that basal rates (on my pump at least) can go up in increments of 0.025. It might actually be better to round it (everywhere?) in that file to a multiple of the scroll rate
Scott Leibrand
@scottleibrand
Jun 14 2016 19:21
@cjo20 that's what @loudnate did in his code. Want to figure out the multiplier to use and PR a fix?
Chris Oattes
@cjo20
Jun 14 2016 19:22
I'll take a look
Scott Leibrand
@scottleibrand
Jun 14 2016 19:22
@Spazholio @cjo20 Sparkfun and Intel are on board with producing a cc1110 sparkfun board. board design is almost done already :)
Chris Oattes
@cjo20
Jun 14 2016 19:22
sparkfun are designing it?
that's gooooooooooooooooood
Dana Lewis
@danamlewis
Jun 14 2016 19:22
Independent designer, but theyll be producing it :D
Chris Oattes
@cjo20
Jun 14 2016 19:23
who is designing it?
Andrew Baugh
@baughaw
Jun 14 2016 19:23
sweet!
Paul Martin
@Spazholio
Jun 14 2016 19:23
Holy cow. That’s awesome. Here’s the big question though - how long until it can be purchased? =)
Scott Leibrand
@scottleibrand
Jun 14 2016 19:23
no, Intel found a third party independent designer here in Seattle (who I happened to know already, and who has a family diabetes connection)
Dana Lewis
@danamlewis
Jun 14 2016 19:23
@Spazholio a few short months probably ;)
Scott Leibrand
@scottleibrand
Jun 14 2016 19:23
dunno how long production will take, but design has taken less than a week so far :)
David Cintron
@loudestnoise
Jun 14 2016 19:24
WANT
Chris Oattes
@cjo20
Jun 14 2016 19:24
so it's quite recent then
Dragan Maksimovic
@dm61
Jun 14 2016 19:24
That's awesome!
Dana Lewis
@danamlewis
Jun 14 2016 19:24
very recent
Paul Martin
@Spazholio
Jun 14 2016 19:24
Well, sign me up for a preorder then. =)
Dana Lewis
@danamlewis
Jun 14 2016 19:24
we had a phone call, we met the same evening, and a week later, board design round 1!
:)
serendipity is awesome
Chris Oattes
@cjo20
Jun 14 2016 19:24
I'd love to get some insight in to the process
Dana Lewis
@danamlewis
Jun 14 2016 19:25
we’ll definitely keep you all posted as soon as we hear anything more about timeline, @ps2 has been lending his briliance & hundreds (thousands?) of hours of expertise to this, too.
JaysonEwer
@JaysonEwer
Jun 14 2016 19:25
💪👍🎉
Dragan Maksimovic
@dm61
Jun 14 2016 19:25
One board = sparkfun base + TI?
Scott Leibrand
@scottleibrand
Jun 14 2016 19:26
no, but the new board should replace the battery block and spacer block
Paul Martin
@Spazholio
Jun 14 2016 19:26
So Battery + Spacer + TI, all integrated?
Scott Leibrand
@scottleibrand
Jun 14 2016 19:26
so we will just need edison, base board, and this
Dragan Maksimovic
@dm61
Jun 14 2016 19:27
that's fantastic
Scott Leibrand
@scottleibrand
Jun 14 2016 19:27
the battery will be a separate thing. possibly with a JST connector if we're willing to make it non-stackable
Dragan Maksimovic
@dm61
Jun 14 2016 19:27
has anyone tried the sparkfun battery block?
Scott Leibrand
@scottleibrand
Jun 14 2016 19:27
still figuring out that part and whether it needs any extra height or width for a good antenna
we use the battery block with a spacer block
Dragan Maksimovic
@dm61
Jun 14 2016 19:27
how long does it last?
Scott Leibrand
@scottleibrand
Jun 14 2016 19:28
like 2 hours
Dragan Maksimovic
@dm61
Jun 14 2016 19:28
ok, not too bad
Dana Lewis
@danamlewis
Jun 14 2016 19:28
…not anywhere near what we need though!
Scott Leibrand
@scottleibrand
Jun 14 2016 19:28
that rig is usually not mobile: I use another battery if I want it to be
Chris Oattes
@cjo20
Jun 14 2016 19:36
blob
Scott Leibrand
@scottleibrand
Jun 14 2016 19:37
so... 40?
Chris Oattes
@cjo20
Jun 14 2016 19:38
As far as I know, medtronic pumps should be ok with 40
Not sure what the minimum scroll rate is on the older models though
Scott Leibrand
@scottleibrand
Jun 14 2016 19:39
could do a multiple of 40 (80 or 120) if we also want to handle newer pumps with more resolution
Chris Oattes
@cjo20
Jun 14 2016 19:39
well, ideally it would match the scroll rate that is set in the pump
Scott Leibrand
@scottleibrand
Jun 14 2016 19:39
if we only want to support older openaps-compatible MDT, then I think that'd be 40
Chris Oattes
@cjo20
Jun 14 2016 19:39
the multiplier is just 1/scroll_rate
Scott Leibrand
@scottleibrand
Jun 14 2016 19:40
120 would have the benefit of always rounding to something that the pump would round to the same value as if we rounded to 40
80 would cause it to always "round up"
so I'd say either 40 or 120
Chris Oattes
@cjo20
Jun 14 2016 19:41
160, not 120. But 40 is probably better.
Scott Leibrand
@scottleibrand
Jun 14 2016 19:41
no, 160 has the same problem as 80
if something is close to but slightly below the midpoint between two different 40ths, it will get rounded to the midpoint, which the pump will then always round up
Chris Oattes
@cjo20
Jun 14 2016 19:42
120 doesn't give a nice number
Scott Leibrand
@scottleibrand
Jun 14 2016 19:43
ok, let's just do 40 then :)
Chris Oattes
@cjo20
Jun 14 2016 19:43
blob
Paul Martin
@Spazholio
Jun 14 2016 19:55
@scottleibrand Do you ever find that openxshareble just hangs and prevents more readings from happening?
Scott Leibrand
@scottleibrand
Jun 14 2016 19:57
not sure, but I wouldn't be surprised if I've seen that. We have a killall -g --older-than 7m openaps in cron that might be helping with that, not sure
we usually use NS, and only pull from openxshareble when we lose connectivity, which of course means I'm usually not logged in watching it when it's using BLE
Matthias
@ceben80
Jun 14 2016 19:58
0
Paul Martin
@Spazholio
Jun 14 2016 19:58
Well, I’ve shortened that line to 4m, but last night it never seemed to actually work. I had processes running from 10 mins prior.
Scott Leibrand
@scottleibrand
Jun 14 2016 19:59
only exception is when I leave my hotspot running when the plane takes off, or when I do something to emulate that on purpose :)
Matthias
@ceben80
Jun 14 2016 19:59
Ups sorry
Scott Leibrand
@scottleibrand
Jun 14 2016 19:59
if you catch it in the act, check if you need to add a -9 to the killall
Paul Martin
@Spazholio
Jun 14 2016 19:59
The loop keeps breaking since it keeps thinking that my CGM time is off due to the reasons I was outlining last night. Hmm.
Scott Leibrand
@scottleibrand
Jun 14 2016 20:01
you're pushing the envelope with trying to get openxshareble reliable enough to use as a primary upload method
Paul Martin
@Spazholio
Jun 14 2016 20:01
I’ve actually added a pkill statement to the end of the CGM loop that I got from your repo. And it only runs every 2 mins, since it could take over a minute for a legit execution.
Scott Leibrand
@scottleibrand
Jun 14 2016 20:01
Ben is the only other person I know of who uses openxshareble as a primary (vs. backup) method of getting BGs, and his dies all the time any time he's in a noisy 2.4GHz environment
Paul Martin
@Spazholio
Jun 14 2016 20:01
Ugh, it appears so. And the whole reason I did it was to get the raw data into NS, and that is just flat-out not working. No idea why.
Chris Oattes
@cjo20
Jun 14 2016 20:03
@scottleibrand How do you run the tests on oref0?
Sarah Hodson Grady
@shgrady
Jun 14 2016 20:17
@Spazholio I found a pump on Craigslist! On its way...
Paul Martin
@Spazholio
Jun 14 2016 20:17
Congrats!
Chris Oattes
@cjo20
Jun 14 2016 20:24
Thanks
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 14 2016 20:26
I'm sure this has been asked in other places but is there any comms with the Tandem line of pumps? I really love my massive tFlex resevoir
Dragan Maksimovic
@dm61
Jun 14 2016 20:42
Does configuring BLE tethering on Edison/ubilinux follow the same approach as described in the doc for RPI2+BLE stick?
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jun 14 2016 20:42
@scottleibrand Thanks, You and @danamlewis have done something magical. Looks like I will be going back to Medtronic for my AP for now, but I do have a call into tandem to be yet another voice in the masses asking for them to enable the bluetooth on the pump.
Brooke Pazoles
@PazolesBrooke_twitter
Jun 14 2016 20:44
@jlucasvt I left Tandem to go (back to) Medtronic for APS. No regrets, but I feel your pain.
Dana Lewis
@danamlewis
Jun 14 2016 20:44
:sparkles: :sparkling_heart: :)
Garrett Webb
@garetis
Jun 14 2016 20:49
This loss of functionality that happened with Medtronic after firmware 2.4... is that something that will ever be able to be worked around? Would a home user ever be able to get a copy of older firmware and install it on a pump themselves, or is that pretty much impossible?
Dana Lewis
@danamlewis
Jun 14 2016 20:49
No one has yet successfully done that
@jaylagorio has asked for old broken pumps so he can investigate that
but right now it’s not known to be possible
Paul Martin
@Spazholio
Jun 14 2016 20:52
Getting remedial for a moment, why does meal.json need pumphistory-24h.json?
Scott Leibrand
@scottleibrand
Jun 14 2016 20:56
In AMA, we calculate COB based on actual BG deviations relative to insulin activity, so we need pumphistory back to DIA hours before the first carb entry
that ends up being more like 6 hours than 24, but most people have a 24h pumphistory for autosens, so we use that
@dm61 No one has successfully set up BT hotspot tethering on an Edison with ubilinux. If you can figure that out please document. :)
Paul Martin
@Spazholio
Jun 14 2016 21:00
Ah, ok. That’s just where my loop ususally dies if everything else is ok and I couldn’t quite remember why we needed it. Is it possible that that information could be pulled from NS instead of the pump, assuming that everything else has been working and uploading to NS for the last 24+DIA hours?
Thacher
@thacherT1D
Jun 14 2016 21:01
@loudestnoise looking at how to get the G5 data into openAPS -- I can get values when I use openaps use ns shell get entries.json 'count=10' but then when I try to create a report I get an empty array []
David Cintron
@loudestnoise
Jun 14 2016 21:06
so the openaps use ns shell get entries.json 'count=10’ pulls BGs from Nightscout
when you say G5 data, do you mean G5 data that is now in Nightscout?
or are you also using a receiver that’s connected to your RPi
Thacher
@thacherT1D
Jun 14 2016 21:14
I am not currently using a receiver
with the openAPS
I have a G5 receiver that is active along with my iPhone, I was looking at connecting that to my openAPS rig if it was better/easier/necessary
currently my G5 data is in nightscout, but not pulling through to the dex report in openAPS
David Cintron
@loudestnoise
Jun 14 2016 21:20
I originally thought I would want to do the same thing, but because it’s just so much easier I went the receiver connected to RPi route
Dragan Maksimovic
@dm61
Jun 14 2016 21:21
@scottleibrand not exactly the answer I was looking for :worried:
Paul Martin
@Spazholio
Jun 14 2016 21:22
Ok, apropos of nothing, but Robomongo is proving to be an excellent tool for connecting to a MongoDB instance.
David Cintron
@loudestnoise
Jun 14 2016 21:22
@thacherT1D so that to say I don’t probably know how to help you with that, but I’m sure someone does.
Scott Leibrand
@scottleibrand
Jun 14 2016 21:26
@dm61 it does seem to work with Yocto, though. not sure what the difference is.
Thacher
@thacherT1D
Jun 14 2016 21:28
okay -- so if I go the connected G5 receiver route -- should I uninstall the G5bridge stuff to night scout? or can I leave the night scout G5 stuff intact and just use the G5 receiver for the openAPS rig?
This message was deleted
Dragan Maksimovic
@dm61
Jun 14 2016 21:28
I've just went through the trouble of loading ubilinux onto Edison, maybe I should have kept yocto
@thacherT1D you do not need to uninstall the bridge - just leave it as is
Thacher
@thacherT1D
Jun 14 2016 21:45
okay, cool -- I'll just go through the steps on the openAPS to setup the G5 receiver as a connected receiver. Thanks!
PieterGit
@PieterGit
Jun 14 2016 22:04
Can't wait till you this wait with openaps... GlucoPred non invasive glucose measuring (trial in Norway), see https://vimeo.com/162799829 and https://clinicaltrials.gov/ct2/show/study/NCT02595580#contacts
Paul Martin
@Spazholio
Jun 14 2016 22:27
So…installing connman seems to have soft-blocked all of my interfaces. Is that normal behavior? Not configuring it, just installing it.
Matt Walker
@walker0
Jun 14 2016 22:58
@dm61 I have a rig with the battery block and have found on yocto it gets around 3.5 hours
Paul Martin
@Spazholio
Jun 14 2016 23:05
Is this the right 10-pin connector for use with the debugger? https://smile.amazon.com/2-54mm-Pitch-2x5Pin-Ribbon-12-inch/dp/B00NQ8BFF2/
Matt Walker
@walker0
Jun 14 2016 23:07
Looks the same as the one that comes in the box
@scottleibrand with the cc1110 block design that is being done will it have a usb port for charging so that it could be used without the base block for during normal use?
Paul Martin
@Spazholio
Jun 14 2016 23:12
Do Weaved or MeshCentral allow you to SSH into the devices, or do you have to use a web-client?
Scott Leibrand
@scottleibrand
Jun 14 2016 23:13
@walker0 still working out details, but there is talk of One Block to Rule Them All. ;-)
Matt Walker
@walker0
Jun 14 2016 23:18
great to here, I actually have a layout in eagle for the same idea based off the spark fun battery block, but one that actually produced would be even better because it saves extremely fiddly soldering of the connectors and chips. I also think for antennas a chip antenna with reasonable range for minimal size and then possibly an additional option to add range by adding a wire whip antenna or similar would be what I would prefer
kstdennis
@kstdennis
Jun 14 2016 23:19
Brandon Faloona
@bfaloona
Jun 14 2016 23:23
Greetings everyone! I'm getting started on an APS for my mom. I have the hardware I need (I think) and I'm working through the FANTASTIC docs. Bravo to everyone who has helped build these tools.
Paul Martin
@Spazholio
Jun 14 2016 23:57
Ok, all kidding aside, Weaved might be a godsend.