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

15th
Jul 2016
Matt Pressnall
@logichammer
Jul 15 2016 00:08
Wondering if folks have been able to get openxshareable to work to pull data back over BLE to the Edison? Found @bewest's code for it:
https://github.com/openaps/openxshareble
I see folks (@TC2013 / @loudestnoise )have tried with the Edison but wasn't sure if there were successes. Thanks!
Dana Lewis
@danamlewis
Jul 15 2016 00:08
we’ve got it working somewhat
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:08
I'm only using openxshareble now
Matt Pressnall
@logichammer
Jul 15 2016 00:08
And I'm a g4 Share dude
With Edison, Jason?
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:09
yes
Matt Pressnall
@logichammer
Jul 15 2016 00:09
Nice...ok...shooting for the stars on this new rig, then!
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:09
it's not perfect, but working pretty good
Matt Pressnall
@logichammer
Jul 15 2016 00:10
Do you have anything in your cron to do cleanup / checks or can I just use things as a normal device?
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:10
I have 2 loops (different dirs) 1 for the pump and 1 for the cgm that are both always running
Matt Pressnall
@logichammer
Jul 15 2016 00:10
Ah
ok
I have some scripts https://github.com/jasoncalabrese/opendex-tools that I'm using to try getting the timing better
Matt Pressnall
@logichammer
Jul 15 2016 00:11
Perfecto! And something in your pump loop to validate you don't have stale data I am assuming
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:11
if they work for you or anyone else it probably makes sense to get them to oref0
Matt Pressnall
@logichammer
Jul 15 2016 00:12
is-fresh looks like what I was thinking
Thanks a ton...will be good to have this to go off of...don't want internet dependency if I can help it
Martin Haeberli
@mhaeberli
Jul 15 2016 00:12
@danamlewis thx I’ll do so in future. Should I re-submit yesterday’s already approved (by Scott) dev PR against master?
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:12
Matt Pressnall
@logichammer
Jul 15 2016 00:13
we go in the woods a bunch and want a small rig for her
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:13
this worked pretty good on vacation in and out of network range
sarahspins
@sarahspins
Jul 15 2016 00:13
okay, I'm starting from scratch again because I botched something up in my last openaps instance trying to get the ns uploader to work (it still wsan't, and now nothing works), and I'm stuck again on getting my G5 to work again
openaps use cgm oref0_glucose --no-raw --hours 2.0 works just fine, but when I try to make it a report and invoke it, I get an error
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:15
I fixed a couple little logging things here openaps/openxshareble#6
Matt Pressnall
@logichammer
Jul 15 2016 00:15
Awesome! I have tested killing network dependency here at home (only for an hour or so) and things worked great...but we went out into the woods this weekend with a regular pi rig and after 3 or 4 hours it just stopped working...even after battery swaps and restarting the rig (which I never normally need to do)...back in network range, it all works. I am going to have an SD card / rig version that doesn't require ANY network dependency and see how that works...excited to work on the small rig, though!
Chris Oattes
@cjo20
Jul 15 2016 00:15
@jasoncalabrese have you considered only running mmtune when the loop fails?
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:15
on the plane there were some devices that didn't have a name or had non-ascii properties
Chris Oattes
@cjo20
Jul 15 2016 00:16
(just been looking at your git repo)
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:16
yeah, I'd like to do that, and do refreshes more like what @scottleibrand is doing
Matt Pressnall
@logichammer
Jul 15 2016 00:16
Dang, glad you chimed in...you are going to save me a ton of time!
thanks, man!
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:17
think we need some more helper scripts for refreshing, trying to avoid the complicated aliases for new stuff
sarahspins
@sarahspins
Jul 15 2016 00:17
I wasn't pulling cgm data that way before because I couldn't get it to work, I was just using openaps use cgm iter_glucose 5 but I think that was causing some other issues since I couldn't figure out how to zone that data
sarahspins
@sarahspins
Jul 15 2016 00:17
anyone have any ideas?
Matt Pressnall
@logichammer
Jul 15 2016 00:19
Nice! I haven't tried to install yet so this will come in very handy! Need to figure out a good way to back up Edison to an image before I bork it...looks like I can mount a USB drive and then image it...not sure how I would do a restore but I should figure that out before I go breaking this thing...took awhile to get it set up.
sarahspins
@sarahspins
Jul 15 2016 00:19
I'm getting very frustrated with this because my G5 isn't working at all on my phone while I am running it as a hotspot (our home internet is out because of an equipment failure with our ISP) - either the multiple wifi connections or the interference caused from the carelink are causing problems, so I can't just pull data from my NS site, because my phone isn't getting data and thus it's not getting to NS
Matt Pressnall
@logichammer
Jul 15 2016 00:19
Really appreciate all the links!
sarahspins
@sarahspins
Jul 15 2016 00:20
I sort of had my loop working yesterday not really getting my cgm data the "right" way but I'd rather not work with creative solutions, I don't understand why what is listed in the docs just won't work
Matt Pressnall
@logichammer
Jul 15 2016 00:20
@sarahspins, can you temporarily connect the G5 directly to your AP?
sarahspins
@sarahspins
Jul 15 2016 00:20
it is connected via usb
I can get data from it, just not as a report
Matt Pressnall
@logichammer
Jul 15 2016 00:21
and what does your alias look like for getting CGM data as a report?
sarahspins
@sarahspins
Jul 15 2016 00:21
I haven't gotten that far yet
it errors out invoking the report
pi@openAPS:~/sarah $ openaps report add raw-cgm/glucose-raw.json JSON cgm oref0_glucose  --no-raw --hours 2.0
added cgm://JSON/oref0_glucose/raw-cgm/glucose-raw.json
pi@openAPS:~/sarah $ openaps report invoke raw-cgm/glucose-raw.json
cgm://JSON/oref0_glucose/raw-cgm/glucose-raw.json
raw-cgm/glucose-raw.json  raised  Python int too large to convert to C long
Traceback (most recent call last):
  File "/usr/local/bin/openaps-report", line 6, in <module>
    exec(compile(open(__file__).read(), __file__, 'exec'))
  File "/home/pi/src/openaps/bin/openaps-report", line 82, in <module>
    app( )
  File "/home/pi/src/openaps/openaps/cli/__init__.py", line 51, in __call__
    self.run(self.args)
  File "/home/pi/src/openaps/bin/openaps-report", line 75, in run
    output = app(args, self)
  File "/home/pi/src/openaps/openaps/cli/subcommand.py", line 52, in __call__
    return self.method.main(args, app)
  File "/home/pi/src/openaps/openaps/reports/invoke.py", line 40, in main
    output = task.method(args, app)
  File "/home/pi/src/openaps/openaps/uses/use.py", line 45, in __call__
    output = self.main(args, app)
  File "/home/pi/src/openaps/openaps/vendors/dexcom.py", line 536, in main
    for egv, raw in itertools.izip_longest(iter_glucose, iter_sensor):
  File "/usr/local/lib/python2.7/dist-packages/dexcom_reader-0.1.10-py2.7.egg/dexcom_reader/readdata.py", line 300, in iter_records
    for x in reversed(xrange(start, end)):
OverflowError: Python int too large to convert to C long
pi@openAPS:~/sarah $
as I mentioned before though, openaps use cgm oref0_glucose --no-raw --hours 2.0 works just fine on it's own... it returns json data that I can see
Matt Pressnall
@logichammer
Jul 15 2016 00:23
look in openaps.ini and find your raw-cgm/glucose-raw.json report
what does it look like?
mine looks like
[report "monitor/cgm-glucose.json"]
device = cgm
hours = 25.0
use = iter_glucose_hours
reporter = JSON
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:24
try openaps report add monitor/glucose-oref0.json JSON cgm oref0_glucose --no-raw --hours 2.0
oref0_glucose is like iter_glucose, but also pulls and joins the sensor records
sarahspins
@sarahspins
Jul 15 2016 00:25
@logichammer it looks like this, but I'm not sure why because I created it exactly like you see above
[report "raw-cgm/glucose-raw.json"]
count = 
use = oref0_glucose
date = display_time
reporter = JSON
seconds = 
hours = 2.0
report = raw-cgm/glucose-raw.json
gaps = 
microseconds = 
device = cgm
minutes = 
sensor = 
threshold = 100
no_raw = True
glucose =
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:25
with that you can get raw data
Matt Pressnall
@logichammer
Jul 15 2016 00:25
nice...that I did not know
sarahspins
@sarahspins
Jul 15 2016 00:25
@jasoncalabrese let me remove the old one and try that
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:25
the path I used is different
Matt Pressnall
@logichammer
Jul 15 2016 00:25
make sure the correct directories are created
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:26
you can add some other path to try it
sarahspins
@sarahspins
Jul 15 2016 00:26
I did make sure the directory was there
@jasoncalabrese same exact error with yours
Matt Pressnall
@logichammer
Jul 15 2016 00:28
I was going to say, those commands look exactly the same
sarahspins
@sarahspins
Jul 15 2016 00:28
then no surprise that it didn't work :)
Matt Pressnall
@logichammer
Jul 15 2016 00:29
yeah...doh!
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:29
oh, I didn't notice that error before
Matt Pressnall
@logichammer
Jul 15 2016 00:29
Underlying python issue?
Or weird prereqs issue?
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:29
maybe
Matt Pressnall
@logichammer
Jul 15 2016 00:29
Sarah, maybe try rebooting (old school fix!)
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:29
or some strange data
sarahspins
@sarahspins
Jul 15 2016 00:30
I've rebooted a bunch of times already, but I can try :)
Matt Pressnall
@logichammer
Jul 15 2016 00:30
dah...that won't help then
yeah, it might be malformed data but I don't know how to troubleshoot that
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:30
I don't know enough about the python code
also might be that oref0_glucose doesn't work on a g5
or the format is a little different
sarahspins
@sarahspins
Jul 15 2016 00:31
but it does work outside of a report?
Matt Pressnall
@logichammer
Jul 15 2016 00:32
If the format of the JSON that comes back is different, it might not know how to parse it correctly
I'll post what a G4 cgm.json file looks like
sarahspins
@sarahspins
Jul 15 2016 00:32
it doesn't look different to me than what is returned by iter_glucose?
I mean there are more fields of data but the format is the same
Matt Pressnall
@logichammer
Jul 15 2016 00:33
[  {
    "_id": "57870aefac3c45f54dd3aadf",
    "device": "xDrip-DexcomShare",
    "date": 1468467875277,
    "dateString": "2016-07-13T20:44:35.277-0700",
    "sgv": 112,
    "direction": "Flat",
    "type": "sgv",
    "filtered": 121000,
    "unfiltered": 129000,
    "rssi": 100,
    "noise": 1,
    "glucose": 112
  },
  {
    "_id": "57870af2ac3c45f54dd3aaea",
    "device": "xDrip-DexcomShare",
    "date": 1468467574907,
    "dateString": "2016-07-13T20:39:34.907-0700",
    "sgv": 107,
    "direction": "Flat",
    "type": "sgv",
    "filtered": 119000,
    "unfiltered": 124000,
    "rssi": 100,
    "noise": 1,
    "glucose": 107
  }
]
does it look like this, though?
sarahspins
@sarahspins
Jul 15 2016 00:34
yes
sort of
Matt Pressnall
@logichammer
Jul 15 2016 00:34
Let's see!
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:34
if iter_glucose works I'd stay with that for now
you only need glucose_oref0 if you're going to use raw data
and you can't do that with the g5 anyway
Matt Pressnall
@logichammer
Jul 15 2016 00:35
gotta boogie...good luck, Sarah and thanks, Jason!
sarahspins
@sarahspins
Jul 15 2016 00:36
meh, okay, but how do I zone that data? I'm assuming that not using zoned data was partially what was causing my not updating NS issue from earlier
I mean I kinda-sorta had my loop working but without any visualition, which sucked
then I broke it, so I figured I'd start over and follow the docs exactly, but that doesn't work at all :(
iananders
@iananders
Jul 15 2016 00:37
@sarahspins you have added the tz device, right?
sarahspins
@sarahspins
Jul 15 2016 00:37
yes
iananders
@iananders
Jul 15 2016 00:40
and are you using the tz device to "rezone" your glucose data when you fetch fresh cgm data? (i had a similar issue)
sarahspins
@sarahspins
Jul 15 2016 00:40
I had some issue creating one of the reports with the iter_glucose data, so no, I wasn't
that is however, what I want to do
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:40
@logichammer 1 more thing I forgot to add, you need to set the mac address for the receiver or it will try connecting to any ble device in range
sarahspins
@sarahspins
Jul 15 2016 00:41
miracle upon miracle though now it seems to be working just fine when I try it again... grr
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:41
if you get it working we should work on some docs for it
sarahspins
@sarahspins
Jul 15 2016 00:42
so now that I can create that zoned data, can someone confirm I really don't need to worry about oref0_glucose?
iananders
@iananders
Jul 15 2016 00:43
you don't if you're using a G5
Hannah Crabtree
@lollydaggle
Jul 15 2016 00:43
I'm also confused if the loop needs zoned pumphistory
sarahspins
@sarahspins
Jul 15 2016 00:43
Hannah I suspect it does, I had my loop working without any zoned data, but NS wasn't updating with anything... even though it was getting the data
Hannah Crabtree
@lollydaggle
Jul 15 2016 00:43
And zoned glucose and clock
Or if it's just nightscout that needs the zoned ones
sarahspins
@sarahspins
Jul 15 2016 00:44
I don't think the loop cares
mine was working until I set about trying to fix the nightscout problem, and broke everything, and started over (again)
Hannah Crabtree
@lollydaggle
Jul 15 2016 00:44
Cool -- that's what it seems like
sarahspins
@sarahspins
Jul 15 2016 00:44
I really want the visualization on NS though
Hannah Crabtree
@lollydaggle
Jul 15 2016 00:45
I am trying your approach of looping than figuring out nightscout but if you're having trouble, I may rethink
sarahspins
@sarahspins
Jul 15 2016 00:45
I broke my sandbox NS site earlier trying to switch my deployment from master to dev and had to make a new one.. so yeah, I've not had a good day :P
on the bright side though, it's good it was already a "spare" site, but still quite annoying
Scott Leibrand
@scottleibrand
Jul 15 2016 00:46
@mhaeberli merged dev <-> master
sarahspins
@sarahspins
Jul 15 2016 00:47
I had a loop last night that was actually updating NS for about three hours, but I am not sure what I did to break it and I haven't been able to get there again :(
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:47
you want zoned data for everything
Paul Martin
@Spazholio
Jul 15 2016 00:48
Is it possible to have 2 NS installs on the same self-hosted machine? I don’t recall specifying a port to run on, so I wasn’t sure if that was easily customizable...
sarahspins
@sarahspins
Jul 15 2016 00:49
I don't know, I've had two sites set up on azure for a long time (since last year) using the same mongo db
Scott Leibrand
@scottleibrand
Jul 15 2016 00:49
@Spazholio I run our NS on a custom port.
Paul Martin
@Spazholio
Jul 15 2016 00:50
@lollydaggle If you don’t find anyone in DC, you’re welcome to mail it to me and I’ll flash it for you. I’ve done it for several people here already. =)
Jason Calabrese
@jasoncalabrese
Jul 15 2016 00:51
memory usage gets to be a problem with multiple NS instances on a single host
Martin Haeberli
@mhaeberli
Jul 15 2016 00:53
@scottleibrand thx - and as advised I’ll do further work on master
sarahspins
@sarahspins
Jul 15 2016 00:56
progress: got some missing cgm data to upload to NS :)
Jason Calabrese
@jasoncalabrese
Jul 15 2016 01:04
:smile:
Hannah Crabtree
@lollydaggle
Jul 15 2016 01:06
@Spazholio thanks! May hit you up on that.
If everything should be zoned, that's another walk through tweak. Phase 2 definitely works off of the raw pump stuff
And I'm still not 100% sure how the tz device works but I'll play around with it when I'm near my pi again
sarahspins
@sarahspins
Jul 15 2016 01:08
@lollydaggle agreed - the walkthrough seems to take some shortcuts when it doesn't need to, and make some giant leaps out of order
Hannah Crabtree
@lollydaggle
Jul 15 2016 01:10
What's helped me is making a giant word doc with tables listing folders, reports, code to add reports and report dependencies
Paul Martin
@Spazholio
Jul 15 2016 01:13
@jasoncalabrese It might not even be a big deal - just didn’t want to mix dev with production when it comes to data that’s going to be regulating my insulin. =)
Jason Calabrese
@jasoncalabrese
Jul 15 2016 01:14
it's a hard system to test
Scott Leibrand
@scottleibrand
Jul 15 2016 01:18
@sarahspins and @lollydaggle: if it's apparent what needs fixed, please do so ASAP and submit a PR so the next newbie doesn't trip on the same thing.
sarahspins
@sarahspins
Jul 15 2016 01:21
@scottleibrand I'm tripping on so many things at the moment that I don't know if anything I could contribute would be helpful or just make things worse :) I cay say though that walking through the nightscout setup as outlined towards the of phase 1, I needed to reference a couple of things in phase 2.... and having been through this a few (or I mean, like 5) times now, it's gotten easier, but I was legit stuck a bunch of times the first two times I tried going through this stuff ;)
Hannah Crabtree
@lollydaggle
Jul 15 2016 01:29
@scottleibrand once I get my setup going, I can take a stab at updating the walkthrough from a newbie perspective. My documentation is really thorough.
sarahspins
@sarahspins
Jul 15 2016 01:39
hmm... so I have temp basal data uploading to and showing up on NS now, but the openaps pill is still says unknown and the pump pill is empty :(
Jason Calabrese
@jasoncalabrese
Jul 15 2016 01:45
that's from devicestatus, using ns-status
it's a separate upload from treatments
sarahspins
@sarahspins
Jul 15 2016 01:47
honestly I feel like the docs are a huge jumbled mess at this point.... I am trying to follow along and it's just NOT working like it should
things are missing, I'm getting all kinds of errors and there is little to no troubleshooting help at any stage if it doesn't "just work"
Jason Calabrese
@jasoncalabrese
Jul 15 2016 01:47
docs never work for me
in general
sarahspins
@sarahspins
Jul 15 2016 01:48
I am sure @lollydaggle would agree with me
Chris Oattes
@cjo20
Jul 15 2016 01:48
For me the docs are more useful as an "how do I set up this small part"
sarahspins
@sarahspins
Jul 15 2016 01:48
well that's all well and good but I'm not a developer
@cjo20 that's nice, but then it sholdn't be referred to as a walkthrough
Jason Calabrese
@jasoncalabrese
Jul 15 2016 01:49
are you at the point where you're running determine-basal
sarahspins
@sarahspins
Jul 15 2016 01:50
I was earlier, but I started over
Jason Calabrese
@jasoncalabrese
Jul 15 2016 01:50
I'd wait on the ns-status upload till you get determine-basal going
Hannah Crabtree
@lollydaggle
Jul 15 2016 01:50
@sarahspins I'm working on my loop most of the weekend so message me and we can struggle together
Jason Calabrese
@jasoncalabrese
Jul 15 2016 01:51
the results from that are a big part of what gets uploaded to devicestatus
sarahspins
@sarahspins
Jul 15 2016 01:51
I get that
openaps do-everything doesn't finish without errors
and I can't figure it out
the gist of what I see is error: argument report: invalid choice: 'oref0-monitor/profile.json' and I can't figure out how to fix that one
the others I've managed to be able to fix so far
sarahspins
@sarahspins
Jul 15 2016 01:57
now I "get" that some of this stuff you only get once you are looping, but again let me point out that these instructions are in phase-1, and looping starts in phase-2
but knowing that does me little good if I can't get any of it workign at all
and the docs are not clear on any of it, and I can't find any more information by searching, which is why I am here, but it's not helpful if everyone is just telling me that I'm doing it wrong :(
I'm doing what the docs say to do
so I am not wrong, the docs are
basically I've gotten it working and looping 3 different times/ways now, basically by just moving on when I got stuck, but that really doesn't work when it leaves things BROKEN that I wold really prefer to have working
Chris Oattes
@cjo20
Jul 15 2016 02:00
I think that error means that you don't have a report oref0-monitor/profile.json set up
sarahspins
@sarahspins
Jul 15 2016 02:00
@cjo20 yes, but my point is that I don't know how to set that up and I can't find anythign about it anywhere
Chris Oattes
@cjo20
Jul 15 2016 02:01
profile.json should be generated by oref0-get-profile
Jason Calabrese
@jasoncalabrese
Jul 15 2016 02:01
you might have a different path
what do you get with openaps report show
sarahspins
@sarahspins
Jul 15 2016 02:02
hold on, my pi locked up
sarahspins
@sarahspins
Jul 15 2016 02:04
pi@openAPS:~/sarah $ openaps report show
cgm://JSON/oref0_glucose/monitor/glucose-oref0.json
cgm://JSON/iter_glucose/raw-cgm/glucose-raw.json
tz://JSON/rezone/monitor/glucose.json
ns://JSON/shell/nightscout/recent-missing-entries.json
ns://JSON/shell/nightscout/uploaded-entries.json
pump://JSON/read_bg_targets/raw-pump/bg-targets-raw.json
units://JSON/bg_targets/settings/bg-targets.json
pump://JSON/read_insulin_sensitivities/raw-pump/insulin-sensitivities-raw.json
units://JSON/insulin_sensitivities/settings/insulin-sensitivities.json
pump://JSON/read_clock/raw-pump/clock-raw.json
tz://JSON/clock/monitor/clock.json
pump://JSON/read_temp_basal/monitor/temp-basal-status.json
determine-basal://text/shell/oref0-predict/oref0.json
pump://JSON/iter_pump_hours/raw-pump/pump-history-raw.json
tz://JSON/rezone/monitor/pump-history.json
pump://JSON/model/model.json
pump://JSON/reservoir/monitor/reservoir.json
pump://JSON/read_status/monitor/status.json
pump://JSON/read_battery_status/monitor/battery.json
pump://JSON/set_temp_basal/oref0-enacted/enacted-temp-basal.json
oref0://JSON/shell/settings/settings.json
pump://JSON/read_settings/raw-pump/settings.json
oref0://JSON/shell/settings/selected-basal-profile.json
pump://JSON/read_selected_basal_profile/raw-pump/selected-basal-profile.json
pump://JSON/read_temp_basal/raw-pump/temp-basal-status.json
pump://JSON/reservoir/raw-pump/reservoir.json
pump://JSON/read_status/raw-pump/status.json
pump://JSON/read_battery_status/raw-pump/battery.json
ns://JSON/shell/nightscout/recent-treatments.json
ns://JSON/shell/nightscout/uploaded.json
ns://JSON/shell/nightscout/preflight.json
I haven't started on phase 2 yet
as I mentioned before I'm stick in phase 1
Chris Oattes
@cjo20
Jul 15 2016 02:05
It might be worth reading through all of it
sarahspins
@sarahspins
Jul 15 2016 02:05
i mean I guess I can ignore this stuff for now and move on to phase 2 and get those other devices set up, but there's absolutely nothing that assures me that is what I am "supposed" to do
@cjo20 I've done so a million times already
this is not the first openaps instance I've set up in the past few days
I've managed to break most of them and had to start over
Chris Oattes
@cjo20
Jul 15 2016 02:06
How did you break them?
sarahspins
@sarahspins
Jul 15 2016 02:07
various ways...
usually by breaking a report (or several reports in one case).. the last time I broke things by trying to zone data that I had skipped over zoning
that one was looping, but not updating NS
in my frustration I ended up deleting it when it would no longer loop and just started over again
Chris Oattes
@cjo20
Jul 15 2016 02:12
it might be better to ask for help in that situation rather than starting over. As the docs are at the moment, you can't really follow through from the top of the contents to the bottom and have everything working 100% at every step.
yiyoMX
@yiyoMX
Jul 15 2016 02:15
Hi, I don´t know if this is the right group to show this but I think it may be of use to some of you, its a BLE bridge for the CC1111 stick. Its compatible with the great work of Pete (RileyLink). If anyone is interested send me a PM.
20160712_175122.jpg
Scott Leibrand
@scottleibrand
Jul 15 2016 02:18
At one point the walkthrough worked end to end, but it was not organized the way it needs to be to encourage people to get NS set up and open loop before going fully closed loop. So we started a project to reorganize them, and then that ended up getting merged prematurely. Since then, no one has been willing to go through everything end to end 6 times and make sure it's all perfect. We've had sporadic efforts to fix some things, but need someone willing to go through from scratch and fix everything. It doesn't require a lot of skill, just some dedication and attention to detail. Any one of you could do it, and you don't have to wait until you have a working rig to get started. If you're taking notes already, you might as well make the actual edits at the same time, and PR them as soon as you finish each page/section.
Dana Lewis
@danamlewis
Jul 15 2016 02:20
Also, there are links to older versions of the docs if you prefer those.
sarahspins
@sarahspins
Jul 15 2016 02:22
@danamlewis that might be less frustrating - how do I find those?
Scott Leibrand
@scottleibrand
Jul 15 2016 02:22
But those won't have a lot of the new template stuff etc. that is a lot easier than it used to be.
sarahspins
@sarahspins
Jul 15 2016 02:22
well I'm working through phase 2 again, and just setting the ns stuff aside
Scott Leibrand
@scottleibrand
Jul 15 2016 02:23
unfortunately the template stuff hasn't gotten quite fully ironed out yet, so we need the help of people using it to fix that.
sarahspins
@sarahspins
Jul 15 2016 02:23
I want to get to where I was with it last night, and if I can get there, I may be in a better place to get help I think
Dana Lewis
@danamlewis
Jul 15 2016 02:24
@sarahspins go to the RTD version of the docs, bottom right corner has a little v drop down and you can pick another date for the docs. Probably will be frustrating in other ways, though, so keep that in mind.
Paul Martin
@Spazholio
Jul 15 2016 02:28
@inv4der That’s definitely interesting.
sarahspins
@sarahspins
Jul 15 2016 02:29
@danamlewis thanks, I'm going to keep going for now, but if I get stuck again it might be useful to be able to go back and see an older version
David
@dakago
Jul 15 2016 02:29
hello @sarahspins i just got everything working a few days ago and as you said, it is hard to follow the docs. What I had to do was to start with openaps report show oref0-monitor/profile.json --ini and work backwards to make sure that all of the corresponding reports were accurately pulling data and were named and zoned correctly. Same thing for oref0-predict/oref0.json and oref0-enacted/enacted-temp-basal.json. There are inconsistencies between the naming of reports in the old docs and some of the naming in the latest aliases and reports that ben has put up.
Paul Martin
@Spazholio
Jul 15 2016 02:29
@bewest https://github.com/openaps/docs/issues/132#issuecomment-215565256 <— if I were to set up a dev rig and start completely from scratch, would this be a good place to start if I wanted to use your new templating?
Dragan Maksimovic
@dm61
Jul 15 2016 02:39
Ben's templates is what I followed when I setup my latest instances about a month ago. Worked pretty well, but it still required some fixes, and filling in some missing reports etc
Julie
@julieraines
Jul 15 2016 02:41
@scottleibrand I had trouble doing PRs until I fully understood what I was doing. So, like @sarahspins I'm running across some of the same issues. So, I skipped anything that didn't work after troubleshooting and now my goal is to go back, troubleshoot, and do PRs. I had to understand the whole enchilada before I could do that. I will be on here a lot tomorrow through Sunday working on this.
jim-andrews
@jim-andrews
Jul 15 2016 02:47
@Spazholio Paul, I'm setting up a few Edison rigs and would welcome the opportunity to walk down that new path as well. My first run at this took about 4 days using the old docs (about 2 weeks ago... it seems like only yesterday) and while "I am" a developer and I stare at json files every day, I did manage to find my way. I think it's all pretty much documented and what is not one can pick up off of gitter. Is the purpose of this project to make it easy enough than "anyone" can do it? If so, I'm confused about liability and the potential hazard to cookbookers and their children. No disrespect intended. I understand it's a tightrope.
Paul Martin
@Spazholio
Jul 15 2016 02:49
I’m gonna redirect that question to @scottleibrand or @bewest or @danamlewis as they’re the people to speak to things like that.
Chris Oattes
@cjo20
Jul 15 2016 02:54
When I've seen this discussed before, there was a definite choice to avoid making things so easy to do that you could put together something that gave you insulin without having a good idea of what you're doing
sarahspins
@sarahspins
Jul 15 2016 02:56
@dakago I coudln't even figure out what reports oref0-monitor/profile.json was dependent on that was causing it to fail... hard to work backwards when you can't find anythign at all.
jim-andrews
@jim-andrews
Jul 15 2016 02:56
And I got that impression when I made my way thorough the docs, up to a point where there was a well designed gap. I thought it was incomplete until I realized that was there for a reason. That was my Ah...Ha. Moment. That for me was when the project became fun!
sarahspins
@sarahspins
Jul 15 2016 02:57
@jim-andrews I'm not referring to the well-designed gap, I had a nice aha moment with that one as well and it really wasn't that hard to figure out... some of the problems I'm running into just shouldn't be so difficult :(
David
@dakago
Jul 15 2016 03:01
@sarahspins do you mean you don't know which reports it's dependent on or why those particular reports weren't working?
sarahspins
@sarahspins
Jul 15 2016 03:02
@dakago I can't figure out what it is dependent on.. that's how I've always fixed the other broken things but with that one I have no idea
Martin Haeberli
@mhaeberli
Jul 15 2016 03:02
@sarahspins I completely relate - I also could not figure out exactly the same thing … time for me to work forward through the docs again ...
David
@dakago
Jul 15 2016 03:02
openaps report show oref0-monitor/profile.json --ini
Martin Haeberli
@mhaeberli
Jul 15 2016 03:03
@dakago - I continue to look for an example of how to add this report ...
sarahspins
@sarahspins
Jul 15 2016 03:04
@dakago nope.... error: argument report: invalid choice: 'oref0-monitor/profile.json' (that's in the middle of a whole lot of junk)
Jason Calabrese
@jasoncalabrese
Jul 15 2016 03:05
@mhaeberli which report can't you add?
Martin Haeberli
@mhaeberli
Jul 15 2016 03:06
@jasoncalabrese - I could probably add it, but I don’t seem to be able to find the invocation - for example, oref0-monitor/profile.json
Dragan Maksimovic
@dm61
Jul 15 2016 03:07
@sarahspins profile.json dependences can be found in the docs at this point: https://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-2/Using-oref0-tools.html#the-get-profile-process, except the file names are not consistent with the names used in phase 1.
Jason Calabrese
@jasoncalabrese
Jul 15 2016 03:07
this issue need to be completed, the old docs are making things so hard
openaps/docs#132
the comment here: https://github.com/openaps/docs/issues/132#issuecomment-215565256 has all the steps to get you 90% of the way done
sarahspins
@sarahspins
Jul 15 2016 03:10
ok, well I already got get-profile working from that page... I just don't know how to get it working as described on the bewest aliases page
and I have a working loop, just without the extra NS stuff I want
Martin Haeberli
@mhaeberli
Jul 15 2016 03:12
@jasoncalabrese - thanks - that looks like a step in the right direction
Jason Calabrese
@jasoncalabrese
Jul 15 2016 03:12
maybe there's a bug in the templates that generates a bug in the aliases
David
@dakago
Jul 15 2016 03:15
@sarahspins I know that I had to change names. The aliases that ben uses do not correspond to the older docs. So for example (I don't know exactly at the moment) the aliases might be using oref0-monitor/profile.json which probably does the same thing as monitor/profile.json or whatever.
sarahspins
@sarahspins
Jul 15 2016 03:21
@dakago ok, that makes sense, and I'll mess with that more later :)
Dragan Maksimovic
@dm61
Jul 15 2016 03:25
@dakago is correct; the names of reports and aliases in phase 1 (Ben's templates, new style) and phase 2 of the docs (old style) are not consistent. For example, oref0-monitor/profile.json from Ben's template has the same purpose as settings/profile.json report in the phase 2 docs. Once you figure out what does what, it should not be too difficult to make name adjustments
Dana Lewis
@danamlewis
Jul 15 2016 03:42
:+1: @cjo20
minus the “sufferers” language ;)
@jim-andrews tight rope is a good way to describe it. it’s meant to be possible for someone with determination who is willing to learn and has a deep understanding of diabetes. and, it’s possible - many of the 95 looping currently had zero technical background.
that being said, there is that well designed gap, and there’s other things that are well-designed.
the mis-naming between sections is not. but, each section is done by a volunteer who has spent hours dedicating themselves to trying to share knowledge.
so keep that in mind looking at the docs, and for everyone talking about the difficulty, thanks for trucking through and thank you in advance for sharing YOUR knowledge of what you’ve learned through this process, like many have done before you ;)
Jason Calabrese
@jasoncalabrese
Jul 15 2016 03:45
:+1:
Dragan Maksimovic
@dm61
Jul 15 2016 03:46
:+1:
sarahspins
@sarahspins
Jul 15 2016 03:50
@danamlewis I do appreciate all the work everyone has put in - I think my sticking point is that maybe the docs shouldn't reference themselves as a walkthrough if that wasn't the intention
Dana Lewis
@danamlewis
Jul 15 2016 03:51
Do you have an alternative name suggestion for that top level example section? (That was a recent add, and the way we make improvements/suggestions is to throw out some ideas and/or submit a pull request, which is basically a suggestion of exactly how to edit something :))
sarahspins
@sarahspins
Jul 15 2016 03:52
that I don't have, but I can certainly think on it when I'm a little less frustrated ;)
I really don't think anyone wants my suggestions right now :P
Dana Lewis
@danamlewis
Jul 15 2016 04:01
k. the last thing I’ll say in general is: keep in mind that people have poured their heart into this, and hundreds (thousands) of hours trying to make this a) possible and b) easier for everyone, ALL while balancing doing the right thing and trying to show people how to DIY safely. So I very much get the frustrations (more than you know - try building something with ZERO docs ;)), but also remember this is not meant to be a push-a-button-and-get-an-AP, thus the learning curve and the weeks spent building and learning new things and navigating the docs in whatever shape they’re in is part of the process and the nature of taking advantage of a somewhat documented DIY project. That’s why, after an issue is identified, I often try to encourage filing an issue or submitting PRs or even brainstorming, or even stepping away and coming back when someone might have more time and is in the frame of mind for wanting to fix it.
sarahspins
@sarahspins
Jul 15 2016 04:03
@danamlewis I agree with that last part... I'm pretty much done for right now, but I'll keep working at it tomorrow, and if I have some giant epiphany that might help others I can certainly contribute to the docs :)
Dana Lewis
@danamlewis
Jul 15 2016 04:03
:+1: sounds like a plan!
Scott Leibrand
@scottleibrand
Jul 15 2016 04:13
Even little suggestions are much appreciated. Lots of people seem to think their input isn't good enough to PR in yet. That is the wrong way to think about it. Any suggestion for revising things is more likely than not an improvement, and if there is some aspect that wouldn't be, we can flag that in the PR before we merge. So please, PR early and often.
Also, PRs of minor fixes are way easier to review and merge, so a lot less likely to be difficult to merge versus large wholesale PRs of all the things you've noticed over weeks. So one PR per day, or one PR per file, is about right. PR early and often. :-)
sarahspins
@sarahspins
Jul 15 2016 04:19
@scottleibrand :+1: I will certainly compile a list of minor things as I go.... when I am sure they need to be revised vs I just have no clue what I'm doing (and even if that is the case, maybe if I eventually figure it out, I can add insight that might help others), I can do a PR :)
Scott Leibrand
@scottleibrand
Jul 15 2016 04:39
Even if you end up being wrong, a PR can help, because we'll immediately comment on it and point out what you're missing. :-)
Kinda like asking on Gitter, but more productive. :-p
Martin Haeberli
@mhaeberli
Jul 15 2016 04:41
:+1: @sarahspins @scottleibrand @danamlewis did a great job turning around my first two PRs this week - the first - very small - the second - probably I bit off more than I could chew . And I’m still lost in the desert, but, candidly , less lost now than 4 hours ago, partly because you (re-)raised the issue I had tried unsuccessfully to raise a few times earlier - everyone here is busy and doing their best …
Scott Leibrand
@scottleibrand
Jul 15 2016 04:46
@mhaeberli's last PR is a great example. Even though he made a few "corrections" that had to be changed back before we could merge, it ended up improving the docs quite a bit. And it took way less effort on my part than finding things and fixing them myself, so there was no need to feel apologetic for anything. The whole thing was quite helpful.
Giving :+1: or :-1: on things is something I can do any time from my phone. Finding things to fix (and proposing fixes) takes way more concentration, so that is the part we need the most help with.
Martin Haeberli
@mhaeberli
Jul 15 2016 04:50
:+1:
sarahspins
@sarahspins
Jul 15 2016 06:08
okay, I lied... couldn't resist tweaking it a little more (after reading on openaps/docs#132 linked above) a bit more before bed, and now my loop is working and it's updating NS, which is all I was trying to do earlier when I broke everything (including a NS site, which is still blank, I had to make a new one to play with).. so that's success for the day, and I've backed it up so I can't break it and have to completely start over from scratch again, I can easily get back to this point :) Progress! :sparkles:
Dana Lewis
@danamlewis
Jul 15 2016 06:26
:+1:
Kim St. Dennis
@kimstdennis_twitter
Jul 15 2016 06:33
nicely done.
Martin Haeberli
@mhaeberli
Jul 15 2016 07:21
:+1:
berat
@beratkurar
Jul 15 2016 09:18
Thank you @cjo20 and @julieraines
cjo20 @cjo20 can't remember what that was for, but you're welcome :)
berat
@beratkurar
Jul 15 2016 09:22
It was the answer for what nightscout is?
Chris Oattes
@cjo20
Jul 15 2016 09:22
ah, yes :)
berat
@beratkurar
Jul 15 2016 09:26
I have Dexcom G4, so will I put xDrip Wixel to T1D's smart phone?
and what I will put in remote observer's smart phone?
berat
@beratkurar
Jul 15 2016 11:00
The link to download Raspbian Jessie for RPi2 is not working, does anyone have an idea to solve this?
More correctly it works very very very slow
berat
@beratkurar
Jul 15 2016 11:11
Ok, it works and I was mistaken
berat
@beratkurar
Jul 15 2016 11:19
where do we write the commands like " sudo bash"?
Chris Oattes
@cjo20
Jul 15 2016 11:23
what do you mean? Can you give an example of where you're being told to type that?
berat
@beratkurar
Jul 15 2016 12:12
For example: "To configure wifi: Type sudo bash and hit enter"
Sarah Hodson Grady
@shgrady
Jul 15 2016 13:16
From first glance, with a PREFLIGHT OK, doesn't it look like my pump not communicating?
pi@artificialpancreas:~/awesomepancreas $ openaps report invoke enact/suggested.json
determine-basal://text/shell/enact/suggested.json
{"delta":-9,"glucose":180,"avgdelta":-5}
{"duration":0,"rate":0,"temp":"absolute"}
{"iob":0,"activity":0,"bolussnooze":0,"basaliob":0,"netbasalinsulin":0,"hightempinsulin":0}
{"max_iob":8,"type":"current","dia":4,"model":{},"skip_neutral_temps":false,"current_basal":1.3,"max_daily_basal":1.3,"max_basal":2,"min_bg":100,"max_bg":100,"sens":40}
Setting neutral temp basal of 1.3U/hr
reporting enact/suggested.json
@beratkurar those commands go in the terminal, what type of connection are you using for the Pi (SSH, VPN, direct monitor/mouse/keyboard)?
if you don't have wifi up yet, I am thinking direct? but if you are working headless you won't do it in a command, you will edit the file
berat
@beratkurar
Jul 15 2016 13:19
Hardwares are not with me yet, I am reading the documentation to get ready.
How does RPi2 connects to Dexcom, by bluetooth USB?
is this bluetooth USB an ordinary USB driver?
I have Dexcom G4
Sarah Hodson Grady
@shgrady
Jul 15 2016 13:24
idk, Bluetooth already on my Pi3, but I don't use it for anything
is it a Dexcom G4 with Share?
if not Share, BLUETOOTH won't do you any good, but you have options
you can either OTG it to the Pi, or you can OTG to an uploader phone to NightScout
Dave Ewall
@daveewall
Jul 15 2016 13:26
Apologies for the long message. I'm not sure if this is a known bug, but I've seen some references to it in the past, but no real answer. Last night my loop stopped working for no apparent reason. After looking at the logs, I was getting the "clock wrong or BG data too old" error, because the last known BG in the glucose.json log was from right when it stopped. If I look at the pump, the clock is fine (as is the looper), it's reading data just fine, but for some reason, the looping device isn't getting any new information after a certain point. This has happened to be a bunch of times, probably around a once per week average. The only fix I found was to do a finger-stick recalibrate. All of a sudden, the calibration point shows up on NS, and 15 minutes later (like clockwork), the glucose numbers start coming in again, but the data between the stopping and the recalibration are permanently missing, even though the pump has the data on its graph. I'm using a 523 with an Edison/TI, but had the issue on the RPi/Carelink stick too. Has anybody else had this issue, or does anybody have any ideas as to what's causing it?
live4sw
@live4sw
Jul 15 2016 13:41
Have you tried disconnecting and reconnecting the USB cable from your receiver to your uploader phone? that happens to me about 1-2x a week
live4sw
@live4sw
Jul 15 2016 13:46
Simply doing that fixes it like 99% of the time. It's annoying and I haven't found a good fix, but it works. Would be curious if others have experienced that and how they have dealt with it
Dave Ewall
@daveewall
Jul 15 2016 13:46
It doesn't have problems receiving data, just the BG info after a certain point, and the Edison uses a soldered serial connection to the TI stick, not USB. Since both the Edison and Pi both had problems with different transmitters, I'd have to think it's something else.
With the Pi, I did have to do an oref0-reset-usb once in a while, but that's a different issue. Maybe that's what you're experiencing.
I also don't have an uploader phone like most, since I'm not on Dexcom. I'm just wi-fi'd into my house or work, and tethered over wifi to the phone when I'm out.
dougtakeuti
@dougtakeuti
Jul 15 2016 14:02
Are we allowed to fly with a rig on?
berat
@beratkurar
Jul 15 2016 14:05
@shgrady thank you
Dana Lewis
@danamlewis
Jul 15 2016 14:05
@dougtakeuti yes
Hannah Crabtree
@lollydaggle
Jul 15 2016 14:09
@danamlewis and others - very curious how TSA reacts. I always opt out of the scanners for a pat-down and they're pretty good with the CGM and pump (one awful experience only). I would assume we don't put the rig through any scanner but I'm having a hard time imagining what to say about it.
live4sw
@live4sw
Jul 15 2016 14:10
This message was deleted
Hannah Crabtree
@lollydaggle
Jul 15 2016 14:10
any use for a "life with a closed loop" section to discuss stuff like travel?
live4sw
@live4sw
Jul 15 2016 14:11
@lollydaggle I am going on my first trip post-rig next month and am planning to build a new one and ship it to the destination. Obviously a hassle but I have had bad experiences with TSA. But on the other side @danamlewis has tweeted a number of times about ease of going through TSA with the rig, I'm pretty sure.
Chris Oattes
@cjo20
Jul 15 2016 14:12
@lollydaggle would you put a laptop through a scanner?
Julie
@julieraines
Jul 15 2016 14:13
@danamlewis I've been pouring over the OpenAPS docs and my first inclination is to suggest structural/order changes but that doesn't really lend itself to the pull request. How would you like that? Open discussion here first? PM?
Hannah Crabtree
@lollydaggle
Jul 15 2016 14:19
@cjo20 more worried about if it messes up the dexcom receiver or carelink - haven't traveled with either before
GabeA23
@GabeA23
Jul 15 2016 14:20
@cjo20 ... I'm wondering the same thing. I would just put my rig through the x-ray machine. If a laptop can go through, I don't see why a rig can't. I think you and your rig get more radiation just by being in a plane in the upper atmosphere than by going through an x-ray machine. The metal detector is a different story (I woudn't put my rig through that).
Chris Oattes
@cjo20
Jul 15 2016 14:27
I wouldn't have a problem putting my rig in a bag and putting it through the normal scans
pump / receiver I stick to manufacturers recommendations
JohnKiel
@JohnKiel
Jul 15 2016 14:27
@lollydaggle I've put my pump and dexcom receiver through the air port x-ray scanner multiple times with no ill effect. (Don't have a rig yet, but I couldn't imagine it being any different.) I generally avoid wearing them through the body scanner as it's a hassle dealing with the TSA.
Chris Oattes
@cjo20
Jul 15 2016 14:30
I've only been through TSA screening once, and that was just with Pump + CGM, and they were fine with it. I got a pat-down, and they swabbed my hands. That was it.
berat
@beratkurar
Jul 15 2016 14:30
I have a wi-fi adapter which is not from canakit, so can i insert it into rpi2 directly or shall i do some installations on it before i insert it to rpi2?
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jul 15 2016 14:33
I just came back from Middle East with no issues and My RPi rig looks like a b_ mb. Everything went in my laptop bag including pump as normal. In Amsterdam for secondary USA TSA Screaning they made me take it out of it’s LowePro Bag. and scanned it in a tote by itself.
JohnKiel
@JohnKiel
Jul 15 2016 14:33
@cjo20 I find the swab and pat-down a hassle. They often re-swap be because they mess up the first. (However, often the TSA agents do seem exited to have the opportunity to swap and test. Seems like a high point of their day when I forget to disconnect and put my devices trough the x-ray, at least at MSP.)
Jeremy M Lucas (GetRileyLink.org)
@jlucasvt
Jul 15 2016 14:33
Domesitc USA no problem
But I am TSA Pre Check as well
Chris Oattes
@cjo20
Jul 15 2016 14:34
@JohnKiel it wasn't any slower than my colleague who went through the normal scanners
JohnKiel
@JohnKiel
Jul 15 2016 14:34
(er, replace swap with swab.)
JohnKiel
@JohnKiel
Jul 15 2016 14:40
@cjo20 I guess it varies from airport to airport. I found KAIA (Jeddah) a real hassle, but perhaps that because I didn't speak Arabic. CDG and LHR have been a breeze.
Dragan Maksimovic
@dm61
Jul 15 2016 14:41
@julieraines I've also been thinking about some more substantial changes to the docs, and I'd be interested to hear about your ideas. I think the main issue in the present docs is that all of NS comes before oref0 is even introduced. I would split NS into two parts: first, in phase 1, allow user to upload treatments, so one could immediately visualize boluses or manual temps. Then, after oref0 is introduced, but before the loop is closed, we would have a second part of NS, which would upload the status into the NS openaps pill. The second open issue with the docs is related to inconsistencies in naming various reports and aliases. This should not be difficult to fix, although we may talk some about what directories and names should be to make most sense. I think all these changes could still be done incrementally, through small PR's, but I just have not had the time to sit down and do it. And I am not sure my ideas about docs are any good anyway :smile: We could use gitter or doc issues, https://github.com/openaps/docs/issues, to discuss details.
Julie
@julieraines
Jul 15 2016 14:46
@dm61 OK - I agree there is a chicken/egg issue of what comes first with NS and just with generic introductory info that sometimes gets buried too far in (what is OpenAPS comes after it's installed but before I've set up RPi for example). And some are having difficulty with seeing the whole picture so they (me included) run through everything quickly to get a better understanding and then go back again to fix and now we're still confused a little so I think structure discussions will then show us where the actual smaller issues are but structural doesn't happen well in PRs so discussing here or in docs/issues is fine with me. Here has the advantage of getting others in the conversation
Dana Lewis
@danamlewis
Jul 15 2016 14:46
re: restructure, you can open an issue and outline your suggested flow changes.
Julie
@julieraines
Jul 15 2016 14:50
how do i open an issue?
There is a tab that says Issues
And should be a button that says create new issue
Julie
@julieraines
Jul 15 2016 14:53
when I go to issues it just says no results matched my search and there is no create new issue button from any dropdown or on the page
Dana Lewis
@danamlewis
Jul 15 2016 14:53
@dm61 that suggestion (splitting NS instructions) would work well. The big BUT: no one has been willing to dive in and do that work. So, laying that out there :) for someone who feels energized to make the docs better. It's totally doable, it just takes time and attention to detail.
Dragan Maksimovic
@dm61
Jul 15 2016 14:53
@julieraines would be nice to have a top-level diagram of the entire system up front, with a simple explanation of how the system is configured and how it works; steps that follow with various vendors, devices, uses, reports, aliases should then make more sense. Issues: go to https://github.com/openaps/docs/issues, and click on the green "New issue" button.
Dana Lewis
@danamlewis
Jul 15 2016 14:54
Top right corner should have a "new issue" button in green
By the way - look at the README on github. It may be some of the info there needs to be repeated in other places, since more people are diving into RTD view first
Dragan Maksimovic
@dm61
Jul 15 2016 14:56
@danamlewis if you think that would be a good idea, I'll try to find some time to figure out what exactly should go where, and then split it into small PR's to avoid making even more mess
Julie
@julieraines
Jul 15 2016 14:57
@dm61 the link got me there and yes a diagram would be delightful and @danamlewis I will look at the README too
Dana Lewis
@danamlewis
Jul 15 2016 14:58
Yup. The NS piece I would say go ahead and start doing as PRs. Any larger structural changes, if someone opens an issue would be good to aggregate suggestions and ideas there first before diving into that worj, since Gitter is so fluid
Dragan Maksimovic
@dm61
Jul 15 2016 14:59
well, got some other work to do now, will get back to docs hopefully this weekend
Dana Lewis
@danamlewis
Jul 15 2016 15:00
Sounds good
berat
@beratkurar
Jul 15 2016 15:01
I could not know what to do with my wi-fi adapter which is not from canakit. can i use it directly with rpi2?
Julie
@julieraines
Jul 15 2016 15:16
what adapter do you have? do you have any info on it?
berat
@beratkurar
Jul 15 2016 15:16
tp-link
now i connect everything to rpi2 and i have a gui in my tv screen. but i don't know how to configure my wifi adapter.
Julie
@julieraines
Jul 15 2016 15:18
is the adapter plugged into the rpi2?
or the wall?
berat
@beratkurar
Jul 15 2016 15:18
yes it is conected
tp-link is connected to rpi2
Julie
@julieraines
Jul 15 2016 15:19
ok in the upper right corner is there a wifi drop down menu next to the volume control?
you should see available networks when you click it once in the drop down
berat
@beratkurar
Jul 15 2016 15:20
yes i saw it
Julie
@julieraines
Jul 15 2016 15:20
looks like a couple computer screens - do you see your network that you log into at home?
berat
@beratkurar
Jul 15 2016 15:20
yes exactly
i saw it
Julie
@julieraines
Jul 15 2016 15:21
try logging into it
berat
@beratkurar
Jul 15 2016 15:21
ok i will try and inform
Julie
@julieraines
Jul 15 2016 15:21
i had trouble at first not being able to connect so i ended up connecting directly to my router with an ethernet cable instead which works well for now and i can fix the wifi later
this is just an option if that doesn't work
the wifi
berat
@beratkurar
Jul 15 2016 15:22
ok i got it hopefully, thank you very much
Julie
@julieraines
Jul 15 2016 15:23
no problen
Scott Leibrand
@scottleibrand
Jul 15 2016 15:25
@daveewall sounds like you're using MDT CGM? If so, that is unfortunately a common issue, but not something anyone has figured out a permanent fix for, just workarounds like you found. It might be helpful though to collect some raw pages of data from the pump to help with decoding what is causing decocare/openaps to get tripped up. Check through the decocare and openaps issues pages: there should be one open for this with instructions for collecting the data for someone like @bewest to analyze.
Hannah Crabtree
@lollydaggle
Jul 15 2016 15:27
@julieraines if you have issues with wifi/ SSHing over wifi, definitely check your router settings. My router needed to be switched from WEP to WPA2 before I could connect to wifi. Then, I had to tweak other settings to let all our network devices communicate so I could SSH in.
berat
@beratkurar
Jul 15 2016 15:28
now my rpi2 is connected to internet over wi-fi
JohnKiel
@JohnKiel
Jul 15 2016 15:33
Perhaps moving the OpenAPS docs, or at least some of the areas more subject to change, to a Wiki would make it easier for the community to keep them up-to-date?
Julie
@julieraines
Jul 15 2016 15:54
thanks @lollydaggle I will try that as I can't live on the ethernet forever! as much as I'd like to : ^ )
with wiki can there still be someone overseeing to make sure the changes are sound before accepted? the wiki I work in at work doesn't do that - it's a free-for-all
Dana Lewis
@danamlewis
Jul 15 2016 15:56
@JohnKiel the little changes aren't a problem - those happen all the time. The challenge is someone carving out time to do big structural changes to make things easier.
@julieraines wikis are free for all, and I think would make the issue worse because there wouldn't be review of changes
Scott Leibrand
@scottleibrand
Jul 15 2016 15:57
Yeah, a wiki would be wide open, so better for drafts of new stuff. Once we have something mostly working, PRs are a pretty lightweight way to get peer review.
Dana Lewis
@danamlewis
Jul 15 2016 15:57
The PR process catches a lot of things, but I don't think
..that it's a blocker to updating the docs. I think time is the biggest issue.
Scott Leibrand
@scottleibrand
Jul 15 2016 15:58
@jasoncalabrese was talking about using a wiki page to draft some docs on openxshareble and some other new stuff. Then maybe those might get copied over to the main docs repo?
sarahspins
@sarahspins
Jul 15 2016 16:00
probably an obvious question, but does anyone know at approximately what voltage is a low battery on the pump triggered?
Scott Leibrand
@scottleibrand
Jul 15 2016 16:03
I'd guess around 1.2 or so.
My voltage history doesn't go that low because we change ours at 1.3
which for a lithium is <12h until dying. Voltage curve for alkaline is smoother, but they last half as long.
sarahspins
@sarahspins
Jul 15 2016 16:06
@scottleibrand gotcha, mine is at 1.34v after running for a few days on alkaline, so I was mostly curious if I needed to be worried about changing it soon or if it might go another day :)
Scott Leibrand
@scottleibrand
Jul 15 2016 16:08
Also depends on how much of the day you're looping.
I would change before bed if you loop overnight.
Alkaline might hold out until morning, dunno.
sarahspins
@sarahspins
Jul 15 2016 16:09
that is a good suggestion - I'm not actually connected to my pump yet, so for the moment I can leave it to do it's thing and see what happens with little worry :)
Scott Leibrand
@scottleibrand
Jul 15 2016 16:10
:+1:
the pump shuts off the radio first, and keeps pumping for a long time once that drain is gone.
But obviously no looping without radio.
Dave Ewall
@daveewall
Jul 15 2016 16:38
@scottleibrand Thanks, Scott, I'll definitely do that the next time it happens.
dougtakeuti
@dougtakeuti
Jul 15 2016 16:54
I thought I had asked this question this morning, but something went wrong as I can't see the post. Are we allowed to have a rig working in an airplane?
Jason Calabrese
@jasoncalabrese
Jul 15 2016 16:56
I have, multiple rigs
dougtakeuti
@dougtakeuti
Jul 15 2016 16:57
no questions at security?
Jason Calabrese
@jasoncalabrese
Jul 15 2016 16:57
only about juice boxes
dougtakeuti
@dougtakeuti
Jul 15 2016 16:57
hilarious!
Jason Calabrese
@jasoncalabrese
Jul 15 2016 16:57
except last time I had a frio gel pack and they needed to call in an explosives expert
they saw something in the explosive tests
but never opened any of the boxes in my backpack
Dana Lewis
@danamlewis
Jul 15 2016 16:59
Jason Calabrese
@jasoncalabrese
Jul 15 2016 17:00
someone needs to make mini 5g juice boxes
Andrew won't use glucose tabs, that would be too easy
Dana Lewis
@danamlewis
Jul 15 2016 17:02
I used to take juicy juice through airports, but decided long run I didn’t like those and instead do skittles, starbursts (those are awesome for microcorrections), and other candy
and they stopped making the toddler size 15g juicy juice boxes, lame.
Jason Calabrese
@jasoncalabrese
Jul 15 2016 17:03
he won't eat any candy, we still get the kids size 15g juice boxes
sarahspins
@sarahspins
Jul 15 2016 17:03
I've become a big fan of transcend glucose gel.... can't explode in my bag the way a juice box can (and will, the moment anything that really shouldn't have juice on it is put in there too) and it's nothing my kids will try to steal :)
Dana Lewis
@danamlewis
Jul 15 2016 17:03
where do you find them, @jasoncalabrese? I can’t find them anywhere
sarahspins
@sarahspins
Jul 15 2016 17:04
@danamlewis they do still make them..... our walmart carries them
Dana Lewis
@danamlewis
Jul 15 2016 17:04
gah. I haven’t seen them in Washington or in Alabama
sarahspins
@sarahspins
Jul 15 2016 17:05
there is another company that still does too.. maybe Apple and Eve? They are organic though and about 3x the price
Dana Lewis
@danamlewis
Jul 15 2016 17:05
oh, yea. I specifically am talking juicy juice brand.
sarahspins
@sarahspins
Jul 15 2016 17:06
our local walmart sells a 3-pack of the 15 juicy juice with different flavors.. I think it's apple, grape, and berry
Chris Oattes
@cjo20
Jul 15 2016 17:12
The problem with skittles or starbursts is stopping eating them :P
Dana Lewis
@danamlewis
Jul 15 2016 17:13
hehe :)
Chris Oattes
@cjo20
Jul 15 2016 17:13
I use glucose tablets. Just had 2 because 0 temp didn't seem to be stopping my BG dropping :(
Hannah Crabtree
@lollydaggle
Jul 15 2016 17:22
Is there any way to turn off the Dexcom G5 receiver alarms before I have it attached to a rig? I HATE this thing right now because it won't shut up. I went Urgent Low last night (want to keep that alarm) but it should not wake me up again 20 minutes later when I go from Urgent Low to just Low. I also don't need it to yell at me during the day since I'm pretty on top of what my numbers are doing.
sarahspins
@sarahspins
Jul 15 2016 17:24
@lollydaggle you can turn off most of them under alerts in the menu... you can't however, turn off the urgent low alarm
Hannah Crabtree
@lollydaggle
Jul 15 2016 17:24
@sarahspins I don't see any profiles for silence. Just vibrate and that still annoys me.
sarahspins
@sarahspins
Jul 15 2016 17:25
@lollydaggle it's not in profiles, scroll down, you should see alerts right above settings
you can turn off the high and low alerts, but not the urgent low (<55)
advanced also contains your rise/fall/signal loss alerts if those were active... high and low repeat won't matter if you turn those alerts off
Hannah Crabtree
@lollydaggle
Jul 15 2016 17:28
Great! I'll try to find it under settings. Urgent low is useful for me but the others not as much
sarahspins
@sarahspins
Jul 15 2016 17:28
@lollydaggle yeah, I understand - I never have my alarms on since I wear a pebble, it just seems redundant :)
which reminds me I that had meant to ask.. I've seen some people using OpenAPS with a pebble watchface that seems to display way more info from NS... how are they doing that?
cdaccarett
@cdaccarett
Jul 15 2016 17:33
Hello im testing my open loop right now and when i use "openaps looop" for some reason it's not updating my current blood sugar levels and is using levels from days before
sarahspins
@sarahspins
Jul 15 2016 17:34
@cdaccarett do you have code in your loop to update those values?
overly simplified, but your loop should basically run like this: get cgm data > get pump data > decide what to do > do it
Dana Lewis
@danamlewis
Jul 15 2016 17:39
@sarahspins probably the URCHIN watchface by @mddub: https://github.com/mddub/urchin-cgm, or a custom one by using Pebble for Cards that’s reading json output direct from the loop
sarahspins
@sarahspins
Jul 15 2016 17:39
@danamlewis thanks :)
cdaccarett
@cdaccarett
Jul 15 2016 17:43
@sarahspins are you talking about this ?
cgm://JSON/iter_glucose/monitor/glucose.json
reporting monitor/glucose.json
sarahspins
@sarahspins
Jul 15 2016 17:45
if that is getting your cgm data, then yes... have you checked cat monitor/glucose.json after that to make sure it's getting current data?
cdaccarett
@cdaccarett
Jul 15 2016 17:45
yeah when i cat monitor/glucose.json its the current data
after going back and doing $ rm -f settings/* monitor/* enact/* and then running the commands again it has the current data
but im wondering am i going to have to $ rm -f settings/* monitor/* enact/* before each loop?
berat
@beratkurar
Jul 15 2016 17:47
do i need to have nightscout to connect to dexcom g4 and get the glocose data
sarahspins
@sarahspins
Jul 15 2016 17:47
you might need to, mine does
Hannah Crabtree
@lollydaggle
Jul 15 2016 17:48
That's another question that's been bugging me - do I need to remove the reports before re-invoking or does the invoking process override what's current there?
berat
@beratkurar
Jul 15 2016 17:48

how i do the following step from openaps documentation:
"Now you will do this for the Dexcom CGM receiver using a usb cable:

openaps device add cgm dexcom"

sarahspins
@sarahspins
Jul 15 2016 17:48
@beratkurar no, you can get it directly from the dexcom
berat
@beratkurar
Jul 15 2016 17:49
how?
by micro usb cable?
sarahspins
@sarahspins
Jul 15 2016 17:50
then use openaps device add cgm dexcom to set up the device in opeaps
Scott Leibrand
@scottleibrand
Jul 15 2016 17:51
@lollydaggle removing a report removes its config, and is not necessary
sarahspins
@sarahspins
Jul 15 2016 17:51
@lollydaggle they should overwrite the old data, but it's a good idea to rm them in case something goes wrong and they don't, that way you aren't using stale data
Scott Leibrand
@scottleibrand
Jul 15 2016 17:51
invoking the report updates the file with new data
and yes, removing the output files is a good safety check to make sure you're not stale
cdaccarett
@cdaccarett
Jul 15 2016 17:52
how do i add the removing the output files to the loop?
Scott Leibrand
@scottleibrand
Jul 15 2016 17:53
removing report = openaps report remove
removing output files = rm monitor/* or similar
Dave Ewall
@daveewall
Jul 15 2016 17:53
@scottleibrand A while ago, there was talk about merging JSON files so you don't need to get 24h of history every time. Was that ever realized? Maybe it's still on the wishlist...
Scott Leibrand
@scottleibrand
Jul 15 2016 17:53
there should be something in the walkthrough suggesting that
sarahspins
@sarahspins
Jul 15 2016 17:53
rm -f monitor/*
Scott Leibrand
@scottleibrand
Jul 15 2016 17:54
@daveewall no, @Spazholio tried it but got all tied up trying to use the not-quite-complete tools. not sure if he ever got anything working
Dave Ewall
@daveewall
Jul 15 2016 17:54
k, thx
Scott Leibrand
@scottleibrand
Jul 15 2016 17:54
@daveewall I just moved on to a loop config that just refreshes the 24h files when they're old enough to need it
Hannah Crabtree
@lollydaggle
Jul 15 2016 17:54
@scottleibrand that's exactly the explanation I was looking for - thanks!
Dave Ewall
@daveewall
Jul 15 2016 17:54
How old is "enough to need it"?
Scott Leibrand
@scottleibrand
Jul 15 2016 17:55
see my openaps-sh setup.sh on the refresh-loops branch
Dave Ewall
@daveewall
Jul 15 2016 17:55
It's really just for auto-sens, right?
it's also used by AMA
Dave Ewall
@daveewall
Jul 15 2016 17:58
That reminds me of another question I had... the last injection site I had (that I changed out this morning) had bad absorption, which caused a bunch of highs. This morning, auto-sens ended up doubling my insulin because it saw excess resistivity over the last 24 hours, causing me lows. I had to go in and remove the auto-sens report from my bash script (I don't use aliases much), so it would use my manually set sensitivity rating for today. Is there any other way to deal with an issue like that?
Jason Calabrese
@jasoncalabrese
Jul 15 2016 18:11
with my last pr to dev you can set the autosens limits in your prefs.json
same way you set max_iob
also the default limits were lowered from 2x to 1.5x
it would be nice to have a reset when there is a rewind event in the pump history, but there might be set changes without a rewind
@daveewall ^
Dave Ewall
@daveewall
Jul 15 2016 18:16
Limits would be great, I think that definitely helps. The rewind option sounds nice to me too. That definitely would have helped me the most for this instance. Thanks for the info! I was waiting to do any updating lately since I know you guys are merging in all of the AMA stuff, though.
Jason Calabrese
@jasoncalabrese
Jul 15 2016 18:19
I had the opposite happen the other week, a bad site, but to be extra aggressive I did a prime instead of a bolus to not effect iob, that caused auto sens to detect a bunch of sensitivity, I saw it low temping to try to get BG to 156
Dave Ewall
@daveewall
Jul 15 2016 18:20
A prime instead of a bolus? Wow, living on the edge. :)
Does it even tell you how much you've primed while you're doing it?
Jason Calabrese
@jasoncalabrese
Jul 15 2016 18:21
I used fill cannula for 1.2, so it shows in the history, but not used by autosens
was really distracted doing the previous site change and forgot to take the guard off the mio before inserting
Dave Ewall
@daveewall
Jul 15 2016 18:22
Oh, gotcha, not the actual "Hold to fill tube" option. Makes sense.
sarahspins
@sarahspins
Jul 15 2016 18:23
@daveewall you can do it as a cannula fill and have a little more control over the amount :) I do this sometimes when I don't necessarily want a small correction to show up as IOB, for example reconnecting after a shower, I use that to make up for missed basal
Dave Ewall
@daveewall
Jul 15 2016 18:24
I'll have to keep that in mind. I don't think too much about IOB. I'm one of those old-timers who is used to doing the basics in his head. Took me a while to even trust the Bolus Wizard. All these new variables are things I never had to think about until the last few months. Always something new to learn with this disease.
Jason Calabrese
@jasoncalabrese
Jul 15 2016 18:24
seems that doing that with autosens is a bad idea
I've been using autosens for a long time, that was the first time it became a problem, the new limits will help, but still feel like I need to watch it more
Dave Ewall
@daveewall
Jul 15 2016 18:27
It seems to track so much that I'm trying to figure out ways to fool it, so it'll just do what I want it to. It's making me have to do things its way... which is fine... just have to learn how it thinks. It seems that as you guys add more features, it just thinks better and better, and eventually, I won't have to fool it anymore.
Scott Leibrand
@scottleibrand
Jul 15 2016 18:58
@daveewall the preferred way to do that would be to use temporary targets: If you want openaps to be more aggressive, set a low temp target, or conversely set a high one if you have reason to believe it should be more conservative
@sarahspins what @danamlewis does for showers is temp to zero right before she suspends the pump, so that OpenAPS knows about the missed insulin and can adjust IOB accordingly. then she doesn't have to give any extra (which is good, because she often drops at the end of a shower due to the hot water releasing pooled-up insulin) but it knows about the lower IOB so it can high-temp later if she starts to rise as a result.
danamlewis @danamlewis re-posts the usability thing that talks about flying AND showering AND running AND etc FAQs ;) https://github.com/openaps/docs/blob/dev/docs/docs/walkthrough/phase-6/Usability-considerations.md
Dave Ewall
@daveewall
Jul 15 2016 19:02
I've actually been doing the same thing lately... I've seen her mention that a few times. I'll keep the temporary targets thing in mind. Thanks.
Dana Lewis
@danamlewis
Jul 15 2016 19:03
yea. anytime @scottleibrand sees me (even unconciously) doing something that would inadvertently fool the system, it means a new features needs to be built to acknowledge that unmet need.
Dave Ewall
@daveewall
Jul 15 2016 19:04
That's great thinking. In the last few months I've used it, I keep thinking, "It would be nice if...", and then it happens anyway a week later. :-)
Dana Lewis
@danamlewis
Jul 15 2016 19:05
haha. but still, ask about it here, or file an issue so we can start tracking discussion on the needs for the feature :) in case we haven’t already started working on it. we don’t always think of everything, so that’s where crowdsourcing is a benefit :)
Dave Ewall
@daveewall
Jul 15 2016 19:09
I think I've started to run out of questions. My last few issues are getting Bluetooth tethering working on an Intel Edison, and finding a way around the iPhone turning off its VPN whenever it sleeps, since I run NS in my house instead of on Azure, and the Pebble needs to see it over VPN. I'm looking forward to a login feature on NS, which would also fix the problem, now that I just got SSL working (which was harder than I thought for the openaps device). Most of those are issues related to me doing things in a non-standard way, though, so I'm not going to bother you guys about them.
John Sjolund
@sjolundjohn
Jul 15 2016 19:10

I am about to start working on a Edison rig. Any tips on hardware? I am going for no soldering route.

First question is whether CC debugger or GoodFET is a better option?

Dana Lewis
@danamlewis
Jul 15 2016 19:11
@sjolundjohn you might want to ask in the mmeowlink channel, too
and I think there were rumblings about someone doing an option comparison chart to help answer these first round of questions :) which would be awesome, but I haven’t seen it come to existence yet
the wiki also might give you some starter guidance: https://github.com/oskarpearson/mmeowlink/wiki
Dave Ewall
@daveewall
Jul 15 2016 19:13
I'd highly recommend building this (if it fits your usage): https://www.thingiverse.com/thing:1525618/apps/
It is the soldering version, though, but I think it was very worth the minor pain.
John Sjolund
@sjolundjohn
Jul 15 2016 19:13
Thanks @danamlewis this is what I am working on - will find the mmeowlink channel and get going. What are you using yourself by the wy?
@daveewall why this suggestion? I it seems the difference between TI USB stick (no soldering) and Edison with ERF radio (soldering) is pretty minimal when it comes to performance, size etx
Dana Lewis
@danamlewis
Jul 15 2016 19:14
I have: Pi/TI; edison/TI; edison/rileylink; a rileylink; a Pi/slice of radio; at one point I had a Pi/ERF
some of them were soldered for me by someone else; one was glued by @scottleibrand, etc.
Dave Ewall
@daveewall
Jul 15 2016 19:15
This is the TI USB stick with soldering. I haven't seen the size of an ERF, but I can't imagine it getting any smaller than that. And the Powerboost makes it very wonderful. Lasts all day, no rebooting, etc.
Jason Calabrese
@jasoncalabrese
Jul 15 2016 19:27
I think those are the same and just have bad labeling, you want the kit with the breakout board and the compute module
even if you plan to start with the sparkfun board the kit is about the same price as the compute module by itself and the breakout board is hard to find by itself
If you're looking for an edison on its own, I think you want EDI2.SPON.AL.S
dougtakeuti
@dougtakeuti
Jul 15 2016 19:44
Thank you @jasoncalabrese and @danamlewis for the hints on flying, rigs and juice boxes. Jason, bomb squad for frio gel is certainly not ironic anymore... not funny :frowning: Off to the airport now, wish me luck!
John Sjolund
@sjolundjohn
Jul 15 2016 19:50
@cjo20 the link from rs-online that is only the breakoutboard right?
Chris Oattes
@cjo20
Jul 15 2016 19:52
Edison Breakout Board Kit (RS Part No. 833-0905)
Edison SoC Module
Exposes native 1.8 V I/O of the Edison module
0.1 in grid I/O array of through-hole solder pads
USB OTG with microUSB-AB connector
USB OTG power switch
USB to UART bridge with microUSB-B connector
Li-ion battery charger
Power supply connector: +7 V to +15 Vdc
further down the description
Jason Calabrese
@jasoncalabrese
Jul 15 2016 19:53
@dougtakeuti the explosives expert was nice, his theory on why some chemicals were detected is spray on sunblock, makes some sense since I soaked the gel pack in the sink and my daughter was playing with spray sunblock in that sink a few days before
sunblock = bad, juice = bad, tons of electronics = ok
good luck
Hannah Crabtree
@lollydaggle
Jul 15 2016 19:55
@jasoncalabrese reminds me of the time I thought my bag (with tons of syringes) got pulled for inspection. Nope, it was the guy ahead of me who had too much spare change floating around.
Chris Oattes
@cjo20
Jul 15 2016 19:56
@sjolundjohn I ordered 833-0905 and got an edison + breakout board. It was an older kit (EDI1BB.ALK), but the only difference between that and an EDI2BB is the countries it is licensed in
John Sjolund
@sjolundjohn
Jul 15 2016 19:58
Ok great thanks
syntaxerr66
@syntaxerr66
Jul 15 2016 20:07
If any of you guys have access to a MicroCenter, they have the edison + breakout on sale right now for $45. It's cheaper than the edison core on it's own :)
John Sjolund
@sjolundjohn
Jul 15 2016 20:12
Thanks all for the help - orders made. @syntaxerr66 that is a fantastic price, about 1/3 of what you pay in Sweden and 50% of what I just ordered from via the UK
John Sjolund
@sjolundjohn
Jul 15 2016 20:23
@daveewall could you share the URL to the PowerBoost you got?
syntaxerr66
@syntaxerr66
Jul 15 2016 20:24
@sjolundjohn yikes!
cdaccarett
@cdaccarett
Jul 15 2016 20:28
Anyone else have this problem? im currently using a G5
cgm://JSON/oref0_glucose/raw-cgm/glucose-raw.json
raw-cgm/glucose-raw.json  raised  Python int too large to convert to C long
Traceback (most recent call last):
  File "/usr/local/bin/openaps-report", line 6, in <module>
    exec(compile(open(__file__).read(), __file__, 'exec'))
  File "/home/pi/openaps/bin/openaps-report", line 82, in <module>
    app( )
  File "/home/pi/openaps/openaps/cli/__init__.py", line 51, in __call__
    self.run(self.args)
  File "/home/pi/openaps/bin/openaps-report", line 75, in run
    output = app(args, self)
  File "/home/pi/openaps/openaps/cli/subcommand.py", line 52, in __call__
    return self.method.main(args, app)
  File "/home/pi/openaps/openaps/reports/invoke.py", line 40, in main
    output = task.method(args, app)
  File "/home/pi/openaps/openaps/uses/use.py", line 45, in __call__
    output = self.main(args, app)
  File "/home/pi/openaps/openaps/vendors/dexcom.py", line 536, in main
    for egv, raw in itertools.izip_longest(iter_glucose, iter_sensor):
  File "/usr/local/lib/python2.7/dist-packages/dexcom_reader-0.1.10-py2.7.egg/dexcom_reader/readdata.py", line 300, in iter_records
    for x in reversed(xrange(start, end)):
OverflowError: Python int too large to convert to C long
Jason Calabrese
@jasoncalabrese
Jul 15 2016 20:36
saw someone else hit that yesterday
I think the the fix was to use iter_glucose, but there are options for glucose_oref0 that should make it work with the g5
the issue is that the g5 doesn't have raw data on the receiver
@cdaccarett ^
cdaccarett
@cdaccarett
Jul 15 2016 20:43
Looking through chat found bwest fix but for some reason doesn't want to work for me :/
cdaccarett
@cdaccarett
Jul 15 2016 20:50
@jasoncalabrese the iter_glucose worked
sarahspins
@sarahspins
Jul 15 2016 21:04
iter_glucose is what I had to use and it does work fine that way :)
JohnKiel
@JohnKiel
Jul 15 2016 21:51
@syntaxerr66 Thanks for the tip on getting an Edison from Microcenter! Was just about to order one from Amazon.
iananders
@iananders
Jul 15 2016 22:59
just getting started with autosens - is there anyway to explicitly assign a different ISF at certain blood sugar ranges?
or just autosens which looks back at the last 24 hours to "guess" or lack of a better term
Dana Lewis
@danamlewis
Jul 15 2016 23:08
@iananders read through openaps/oref0#58 for some background on autosensitivity. it doesn’t guess, it calculates based on observations of estimations compared to what actually has happened in last 24 hours (or really, 24hr+DIA)
iananders
@iananders
Jul 15 2016 23:22
yeah - i realize words matter here - "guess" was just my shorthand
Scott Leibrand
@scottleibrand
Jul 15 2016 23:26
no one has yet built an open source algorithm for figuring out the best way to reduce ISF when BG is high, but several people have discussed the possibility. @bustavo is testing one with #simpancreas, so he might be willing to share exactly what he did once he's happy with it. someone would still need to code it for OpenAPS though.
in the mean time, oref0 will increase high temp if BG is not coming down like it should, based on deviation and expected delta, which accomplishes much the same thing, just a bit more slowly.
we've never seen a case where BG stays high and refuses to come down despite high temps unless max basal or max IOB is too low, or there's a problem with the pump site (so extra insulin wouldn't really help regardless)
if you do have such an example while running oref0, I'd like to look at it
Hannah Crabtree
@lollydaggle
Jul 15 2016 23:59
blob