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

14th
Aug 2018
lyntonr
@lyntonr
Aug 14 2018 00:13
0.7.0 , logs said it was 11 commits behind. I have a nightscout screenshot I will put up a bit later , can't right now though ..
Dana Lewis
@danamlewis
Aug 14 2018 00:14
@cluckj any strong reaction/pushback to me merging it, and you can PR additional edits?
or anyone else who has a strong opinion :)
lyntonr
@lyntonr
Aug 14 2018 00:21
Screenshot (1).png
Screenshot (2).png
out for a little while now back in 3 hours. thanks for looking.
Jon Cluck
@cluckj
Aug 14 2018 01:05
@danamlewis :thumbsup: merge now, edit later?
Dana Lewis
@danamlewis
Aug 14 2018 01:06
ya, I will. thx
Jon Cluck
@cluckj
Aug 14 2018 01:09
no bit.ly link?
Dana Lewis
@danamlewis
Aug 14 2018 01:09
took it out for now until the page was live and we could create it
Jon Cluck
@cluckj
Aug 14 2018 01:09
ah
Dana Lewis
@danamlewis
Aug 14 2018 01:10
Zach Gohr
@zgohr
Aug 14 2018 02:16
@lyntonr i assume the carbs and bolus you’re referring to were Bolus Wizard? if so - i had the exact same issue multiple times and is what pushed us faster to move to manual boluses (non-wizard) with carbs entered via IFTTT buttons. My understanding is this is a semi-known issue reading from the pump. have not had the same issue reading boluses since
Garrett Webb
@garetis
Aug 14 2018 02:25
Quick question -- on my log the "Raw" value comes in as a long string of zeros. Is this something I should look into or leave it be and not worry about it? I copy & pasted into excel and it was 128 of them, if it matters
Dana Lewis
@danamlewis
Aug 14 2018 03:11
@garetis where are you getting raw data from?
Garrett Webb
@garetis
Aug 14 2018 03:39
@danamlewis, it's an Edison/Explorer/G5 using Lookout
@danamlewis It looks like it stopped. I just installed Lookout/xdrip and was playing with plugging in the receiver with Lookout running, so maybe that was it. I will keep an eye out and if it comes back I'll report back
lyntonr
@lyntonr
Aug 14 2018 03:50
@zgohr yes bolus wizard meal bolus but is detected as a carb correction without insulin only. So there is missed insulin here. I think I have seen it before too. I believe this to be a serious problem .
lyntonr
@lyntonr
Aug 14 2018 05:07
Screenshot (3).png
just checked nightscout again and that incorrect 70 carbs "correction" has been updated to correctly reflect the 10.0 units missed earlier. All too late of course but maybe this will help figure this out. Here is screenshot of last 24 hours from now. Note it is showing a meal bolus and not a correction now. (8:22pm last night)
lyntonr
@lyntonr
Aug 14 2018 05:13
Screenshot (4).png
Screenshot (5).png
closer look at the pills reveals there are now two entries one as a correction and one as a bolus, see the pills below at 8:22
Scott Leibrand
@scottleibrand
Aug 14 2018 05:14
@lyntonr which oref0 version/branch?
lyntonr
@lyntonr
Aug 14 2018 05:17
0.7.0 dev at the time it was 11 commits behind. I have since done git pull but not global install since. I noticed there was one recent commit of yours in the area of the checking for duplicate entries that I did not have that may be relevant here.
Scott Leibrand
@scottleibrand
Aug 14 2018 05:23
Hmm, I was hoping it was an 0.6.2 thing and fixed in 0.7.0. Sounds like we have the problem with the Go based pump comms too, though. I wonder how we’d go about reproducing this... I suspect it’s probably some sort of race condition...
lyntonr
@lyntonr
Aug 14 2018 05:35
interesting that the correct meal bolus finally did get in there although long after the event. Let me know if I can do anything to help debug. I thought the log messages I posted showing the times are interesting too, only 11 seconds after the bolus, (edit actually 36s after) so perhaps as you say a race condition. here are those relevenat log bits again,
Aug 13 20:22:00 openaps pump-loop.log:  Retry 1 of refresh_pumphistory_and_meal
Aug 13 20:22:47 openaps pump-loop.log:  Pump history updated through 2018-08-13T20:22:11+10:00; meal.json Skipping bolus wizard entry 0 in the pump history with 70 g carbs and no insulin.
Aug 13 20:22:47 openaps pump-loop.log:  Timestamp of bolus wizard: 2018-08-13T20:22:11+10:00
Zach Gohr
@zgohr
Aug 14 2018 12:59
@lyntonr this is my experience exactly. a few times only took an hour or two. Another time took upwards of 4 hours
Diabetic_Dude
@diabetic_dude_twitter
Aug 14 2018 13:37
does anyone know the general range of the new PI Hats? They seem to be pretty darn good.
Diabetic_Dude
@diabetic_dude_twitter
Aug 14 2018 13:57
@kelseyyearick I work at a school also, I can try to help if you would like, I have had to do many things to get my connection to the internet solid while at Work/School. Schools can be very odd with the open networks. I don't even use our open internet here anymore, because it is not reliable for me at all. I will be in and out today between 8 am and 5 PM (CST) .....
Scott Leibrand
@scottleibrand
Aug 14 2018 16:19
One workaround might be to always do a full pumphistory refresh at the end of every loop...
renegadeandy
@renegadeandy
Aug 14 2018 19:15
Hi all, where do i add security so I lockdown my nightscout? As per : https://github.com/nightscout/cgm-remote-monitor/wiki/API-v1.0.0-beta-Security Where do I actually find and configure this key, and the auth_default_role?
Scott Leibrand
@scottleibrand
Aug 14 2018 19:20
search through the OpenAPS docs for token auth and follow the links
renegadeandy
@renegadeandy
Aug 14 2018 19:54
Hi Scott, thanks. I have enabled this, so if you have auth_default set to denied, then you will always be shown the API Secret entry page from a browser which has no key right?
Scott Leibrand
@scottleibrand
Aug 14 2018 19:55
I believe so, but I don't use it
renegadeandy
@renegadeandy
Aug 14 2018 19:58
Its odd that the token_auth docs are in openaps, and not in the nightscout readme or wiki :/
Dana Lewis
@danamlewis
Aug 14 2018 20:20
there's not been a lot of upkeep over on NS side. I was asking James about this this morning. if we have volunteers, I can help connect for people to get added to be able to help update the NS docs. It'll be a lot of work to update the docs, but does need to be done.
renegadeandy
@renegadeandy
Aug 14 2018 20:27
Happy to do my bit for this
Because I went looking and got quickly confused. It will however need somebody to check what I right, as I am very new to this whole thing
Dana Lewis
@danamlewis
Aug 14 2018 20:28
just asked James about moving to an open source back end, which allows for peer review as part of the update process. (that's what we use in OpenAPS, too)
renegadeandy
@renegadeandy
Aug 14 2018 20:29
Perfect
renegadeandy
@renegadeandy
Aug 14 2018 20:40
I also just setup the nightscout app on my wife's phone, but it uses the appsecret for the security instead of the token
is that normal? Also xDrip+ uses appsecret instead of token....so is it only openaps which uses the token for security at this point? Nightscout itself doesn't even appear to use the token
@scottleibrand ?
Scott Leibrand
@scottleibrand
Aug 14 2018 20:47
I don't use token auth or require an API secret, so you'll have to ask someone who does
the implementation of NS "security" features has been rather inconsistent: no one seems to stick around long enough to address all the interoperability and usability concerns, they just get it working well enough for themselves and leave it there
renegadeandy
@renegadeandy
Aug 14 2018 20:51
I would be interested in assisting and improving the docs and code. I would like somebody to confirm my findings first though :)
Zach Gohr
@zgohr
Aug 14 2018 20:54
@renegadeandy you can use token for xDrip+ i believe. use token=mytoken instead of the secret
renegadeandy
@renegadeandy
Aug 14 2018 21:29
@zgohr you are given an input URL, so currently it's http://<secret_key>@hostname. Are you saying now it would be http://hostname:port/api....?token=mytoken
?
renegadeandy
@renegadeandy
Aug 14 2018 21:36
btw, what does "smb" stand for?
smart bolus wizard? If so, it's not acutally called "smart bolus wizard" in my paradigm, its just called "bolus wizard"
MarvelousMarvin
@MarvelousMarvin
Aug 14 2018 21:37
@renegadeandy super micro bolus
Garrett Webb
@garetis
Aug 14 2018 21:52
@danamlewis It looks like that long string of zeros in my "raw" reading is part of the enact/bolused.json file, so whenever that comes up is when I see it. I'm guessing that's because I'm using Lookout and it's looking for other cgm data?
renegadeandy
@renegadeandy
Aug 14 2018 22:48
SMB is very cool, interesting because i sometimes actually do this manually myself!
Zach Gohr
@zgohr
Aug 14 2018 23:18
@renegadeandy that might work, but i take back what i said. i have my xdrip set up with the secret key. i’m not sure why i thought token worked
Dana Lewis
@danamlewis
Aug 14 2018 23:25
@garetis might want to post that in the xdri-js channel, too, and see if anyone else has seen that