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

14th
Oct 2015
Scott Leibrand
@scottleibrand
Oct 14 2015 00:00
I'm gonna declare lazy consensus on this one and merge it if no one else does so (or objects) by EOD. :)
Actually, @tghoward if you could give it a +1 on the PR if you want it merged, that's also good enough
Dana Lewis
@danamlewis
Oct 14 2015 00:01
:)
Tim Howard
@tghoward
Oct 14 2015 00:01
Ha! Ok, I'll check it out.
Tim Howard
@tghoward
Oct 14 2015 00:11
Made a 'looks good' comment - if that's what you meant. What's slightly bothersome to me about these directions is that the iOS directions have you editing WPA-supplicant and the windows/linux directions have you editing the interfaces file. I think either will work and perhaps you need to edit both in some cases. If the very simple edits to WPA_supplicant.conf work, then perhaps it should be written that way for all. But afraid to suggest such a radical change. Will explore how it works.
Scott Leibrand
@scottleibrand
Oct 14 2015 00:12
yeah, that bugged me too
the challenge is that if you're editing the file offline (on the SD card from your computer) you can't run commands, just edit files
so we should probably write it so that the direct interfaces edits are just a bootstrap, and then you do the add-both-networks thing once you get it online regardless of how you bootstrapped
Tim Howard
@tghoward
Oct 14 2015 00:13
Yes, true (and I have been generally connecting via cable and cheating!)
but both supplicant.conf and interfaces are files so could be edited just on the SD card, right?
Scott Leibrand
@scottleibrand
Oct 14 2015 00:14
yeah, if you're willing to put plain-text passwords in you could do it that way
(that's what we did originally)
Tim Howard
@tghoward
Oct 14 2015 00:16
hmm, right. And that's how it is now written -- using plain text passwords.
Scott Leibrand
@scottleibrand
Oct 14 2015 00:16
which is fine for most wireless networks
usually your password is pretty freely given out to anyone you let into your space
Tim Howard
@tghoward
Oct 14 2015 00:18
by not plain text, do you mean something like the debian psk setup here? https://wiki.debian.org/WiFi/HowToUse#WPA-PSK_and_WPA2-PSK
Eric
@ecc1
Oct 14 2015 00:20
@oskarpearson I had better luck with the SRF stick, but I pretty much started from scratch for firmware (so I could build with sdcc under linux). I can send and receive to my MM pumps from about 6 feet away. I haven't put anything up on github yet but would be happy to share what I have.
Scott Leibrand
@scottleibrand
Oct 14 2015 00:26
@tghoward those commands in the osx section of the docs generate a hashed version of the wifi password, along with a commented plain-text version
Tim Howard
@tghoward
Oct 14 2015 00:28
ah - ok, thanks for the clarification.
Scott Leibrand
@scottleibrand
Oct 14 2015 00:30
I've been confused at least once by changing the comment and forgetting to regenerate the hash. it might be easier to just recommend people use plain-text unless they have a super-secure wifi password
Tim Howard
@tghoward
Oct 14 2015 00:40
Would it be cleanest to simply write the instructions around having an ethernet cable to connect the pi to the computer? Then you could use the wpa_passphrase command in both cases and no plain text anywhere? Seems like a pretty small addition to the hardware requirements - an ethernet cable.
Eric
@ecc1
Oct 14 2015 00:41
@oskarpearson Here's what I have so far: https://github.com/ecc1/cc1111
Scott Leibrand
@scottleibrand
Oct 14 2015 01:05
ok, I just merged all outstanding pull requests on the docs. I probably broke something, so @tghoward or others, if you could find what else needs tweaked now, feel free to do so and send in your PRs. :)
Peter Miller
@ochenmiller
Oct 14 2015 01:08
@ecc1, wirelessthings.net website has some pretty interesting documentation on adding and modifying antennas. Was your 6ft range achieved using the built in chip antenna?
Dana Lewis
@danamlewis
Oct 14 2015 01:09
cool. thanks for deleting my old patches, too @scottleibrand . @tghoward, I will look for yours as next round before I go do in and do a master review on org & other high level stuff.
Scott Leibrand
@scottleibrand
Oct 14 2015 01:11
I just did a PR to pull all that stuff into your dev branch too.
Dana Lewis
@danamlewis
Oct 14 2015 01:12
thx
Dana Lewis
@danamlewis
Oct 14 2015 01:21
All: if anyone wants to hop on the phone tonight for an impromptu call and talk mobile project related stuff for OpenAPS, let me know. Right now just looks like @amazaheri, @audiefile, @scottleibrand, & I. (Or if there's a ton of interest and tonight at 8:30 Pacific doesn't work, can also schedule another call with better time zone accomodation/more notice.)
Matthias Granberry
@mgranberry
Oct 14 2015 01:37
I'm interested in hearing. Is there a bridge number?
Dana Lewis
@danamlewis
Oct 14 2015 01:38
I had set up as a Google Hangout, but if people aren't at their computers for that, could do a call-in only instead
Matthias Granberry
@mgranberry
Oct 14 2015 01:43
I must not be on a mailing list or something. Can you send me info at matthias.granberry@gmail.com?
Dana Lewis
@danamlewis
Oct 14 2015 01:44
Yea, it's an ad-hoc last minute talk, I hadn't sent it out to the list as an official call. Will shoot info. If anyone else wants it, ping me here or privately.
Sent, @mgranberry.
Dana Lewis
@danamlewis
Oct 14 2015 02:19
Based on interest, putting out a Doodle to get people's availability for the mobile call (cancelling tonight's). If you're interested, use this Doodle to share availability over next two weeks for OpenAPS mobile discussion call: http://bit.ly/1MxUhLe . I also created a Gist to start storing some commentary related to expressing interest/input related to this for now until we get some work moving & can point discussion to an active repository/project: http://bit.ly/1k3CPTT
And if there's anyone abroad interested for which those time zones just royally stink, let me know that, too?
Eric
@ecc1
Oct 14 2015 02:26
@ochenmiller I'm using unmodified SRF sticks and TI dongles. Not sure I'd trust my hardware skills beyond soldering on headers and serial cables :-)
eszcloud
@eszcloud
Oct 14 2015 03:16
@johnmales: @channemann has looked into range and directionality regarding the carelink stick. The transmitter in the carelink stick is on the bottom gray side. Have you tried pointing that side toward the pump?
John Males
@johnmales
Oct 14 2015 10:12
@eszcloud @channemann Yes I think pointing the stick in the appropriate direction helps to a degree, will be trying to experiment a bit more with this. @oskarpearson I note the TI CC1111EMK868-915 device that I'm using with an android phone for night scout overnight is far more forgiving range wise - basically anywhere in the same room seems fine to get reliable data from the medtronic cgm and pump.
John Males
@johnmales
Oct 14 2015 10:24
@danamlewis thanks for the multiple rPi comments. Would be keen to be involved in the google hangout, if only to say hello! My timeslots are limited due to timezones putting the hangouts right in the middle of my work day, but understand the difficulties of this sort of thing.
Ed Raskin
@msrcgm
Oct 14 2015 12:26
Hi Gang, I'm still stuck because my iob calculation is still not including recent bolus information. Ben helped me tag the problem here: https://github.com/msrcgm/myopenaps/releases/tag/why-is-iob-zero
This message was deleted
Ed Raskin
@msrcgm
Oct 14 2015 12:32
Tagging @amazaheri in response to your Oct. 5 question to me. Was traveling for a week, so getting back to hopefully finishing an implementation. Seems like I am close, but feels far.
Ali Mazaheri
@amazaheri
Oct 14 2015 15:36
Hey @msrcgm welcome back. Go ahead and upload some new jsons to your repo so we can look into your iob calc issues.
eszcloud
@eszcloud
Oct 14 2015 17:47
Anyone have favorite schemes or toolboxes for de/encoding JSON files in matlab? @diabeticgonewild Sounds like you're an avid matlab user and might have a rec?
Ben West
@bewest
Oct 14 2015 17:51
apparently, matlab will decode web requests as json automatically
eszcloud
@eszcloud
Oct 14 2015 17:52
that's interesting and useful
Ben West
@bewest
Oct 14 2015 17:52
so you might simply run simple web server, python -m SimpleHTTPServer
then use the default webrequest against it
that serves anything in current dir on port 8000
eszcloud
@eszcloud
Oct 14 2015 17:54
how do I access port 8000?
it's part of the url, it's be like http://localhost:8000/clock.json in webread Import
eszcloud
@eszcloud
Oct 14 2015 18:00
thanks!
Oskar Pearson
@oskarpearson
Oct 14 2015 20:10
@johnmales Which software are you using with the TI CC1111 and nightscout?
Ben West
@bewest
Oct 14 2015 20:36
 if [ -z $(openaps fetch-model ) ] ; then ( mm-stick warmup fail "NO PUMP MODEL RESPONDED!" || mm-stick fail "NO MEDTRONIC CARELINK STICK AVAILBLE") ;  else  echo "OK" ; fi
 File "/usr/lib/python2.7/dist-packages/serial/serialposix.py", line 466, in write
    if not self._isOpen: raise portNotOpenError
ValueError: Attempting to use a port that is not open
/usr/local/bin/mm-stick FAIL
fail NO MEDTRONIC CARELINK STICK AVAILBLE
bewest@hither:~/Documents/foo$
openaps alias add fetch-model "use pump model"
John Males
@johnmales
Oct 14 2015 20:37
Hi @oskarpearson I'm using https://github.com/arbox0/MedtronicUploader - currently it is one release behind on nightscout 0.7 if I recall, but it otherwise works pretty well. Also the interesting thing for me as well is it gets data from the medtronic CGM directly, so I'm hoping there will be a way to do the equivalent of dexcom raw data.
Oskar Pearson
@oskarpearson
Oct 14 2015 20:43
@johnmales And you flashed the CC1111 with mmcommander I assume?
BTW: Hi - nice to meet you ;)
John Males
@johnmales
Oct 14 2015 20:47
Hi as well Oskar :) yes I followed the guide in that github repo. I will probably try flashing the cc1111 firmware version that can transmit next and have a play - with the pump disconnected of course
Eric
@ecc1
Oct 14 2015 21:36
@oskarpearson If you tried MMCommander on an SRF-Stick, then I see why it didn't work. The GPIOs are connected differently, and of course there's no pushbutton, so the firmware needs to be modified. But it doesn't build under Linux so I didn't go that route.
Oskar Pearson
@oskarpearson
Oct 14 2015 21:47
@ecc1 It's not so much that mmcommander didn't work - it's more that any firmware I could find wouldn't work at all. It seems like it might be because I've got a new version of the srf stick. The blogs all appear to refer to older releases. rfcat: wouldn't work (even the hex file tweeted by the rfcat author previously). Installing the CC bootloader doesn't work (the device immediately no longer even shows up as a USB device). The closest description I could find is https://e2e.ti.com/support/wireless_connectivity/f/155/t/240883 (see the description and the reply on "Jun 6, 2013 10:34 AM" for why I think it might be a different chipset or similar)
All the lights would also immediately come on after flashing - and never go off again
Ben West
@bewest
Oct 14 2015 23:22
just put in preflight check to my loop
preflight = ! bash -c "rm -f model.json && openaps report invoke model.json && test -n $(json -f model.json) && echo \"PREFLIGHT OK\" || ( mm-stick warmup fail 'NO PUMP MODEL RESPONDED' || mm-stick fail 'NO MEDTRONIC CARELINK STICK AVAILABLE')"
Ben West
@bewest
Oct 14 2015 23:56
I put a few more mm-stick uses in openaps/openaps-js#39
warmup and fail