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

24th
Apr 2018
Dana Lewis
@danamlewis
Apr 24 2018 01:08
@cbrenner as @scottleibrand said - it's a long path. We all know it's technically feasible. But it's really about co's figuring out the legal and liability aspects, which are not clear yet. (And also deciding there's a business case to have an iPump approach, similar to the iCGM de novo approval that just came out.)
@cbrenner and no- haven't had a chance to ifttt test yet, sorry
@garykidd last I heard they were still testing the batch. If that's successful, I assume you'll get your shipment notification when it's done if you preordered. It's still April ;)
@jimrandomh today's meeting was the beginning of those discussions re figuring out a way forward for interoperability of all components( pump, CGM, algo, and/or controller).
@CarnageMarkus have you checked the NS troubleshooting page and checked all those steps? Might be some way future data throwing things out of whack on NS
Scott Leibrand
@scottleibrand
Apr 24 2018 04:44
Is there anyone here who's actively (within the last 36h) using UAM+SMB w/o carb announcement?
James Babcock
@jimrandomh
Apr 24 2018 04:45
Not on purpose, but I sometimes forget carb announcements and I've done that within the last 36h
Whoa, I see something bizarre in my Nightscout visualization over the past ~2.5h
WeirdNightscoutScreenshot.jpg
James Babcock
@jimrandomh
Apr 24 2018 04:51
Those giant boluses do not appear in my pump history and did not disappear from my pump's cartridge
I'm on my own dev branch. Checking logs for weirdness...
Scott Leibrand
@scottleibrand
Apr 24 2018 04:51
I've seen that occasionally as well: not sure what it is, but it seems to be a NS visualization thing
I've seen the same boluses show up correctly at other times
@jimrandomh I sent you a PM re: UAM
Martin Haeberli
@mhaeberli
Apr 24 2018 05:00
@jimrandomh @scottleibrand - except for the .9999 part of the equation, this looks very similar to something we saw maybe 72 hours ago. NS showed (per memory) 18 g 12 U but reservoir history luckily showed much smaller bolus … OpenAPS Pill IOB and standalone IOB Pill reflected something like the larger (12 U) dose on board. Actual CGM trace over over next 6 hours luckily worked out well even though forecast was for - 800+ BG
@scottleibrand likely but let me look - that is, running UAM enabled, SMB enabled, sometimes do actual UAM
James Babcock
@jimrandomh
Apr 24 2018 05:25
A little time plus a refresh later, and the weird NS visualization from the earlier screenshot is gone, replaced by correctly rendered supermicroboluses. Hmm.
Scott Leibrand
@scottleibrand
Apr 24 2018 06:01
openaps/oref0#985
Samuchco
@Samuchco
Apr 24 2018 13:59

@danamlewis I have a question regarding the docs. Under "Understanding your preferences and safety settings", it states that

*If the temporary basal rate setting recommended by OpenAPS (as determined in oref0/lib/determine-basal/determine-basal.js) exceeds either:

the user’s max basal rate setting (which is set in the user’s pump), or
max_daily_safety_multiplier the highest programmed basal rate (as specified by the basal rates in the user’s pump), or
current_basal_safety_multiplier
the user’s current basal rate (as specified by the current basal rate programmed in the user’s pump)*

Does the autotune settings trump the pump settings? So for example, If highest pump basal is 0.5u/hr and autotune changed it to 0.35u/hr, the max_daily_mulitplier at 3 will allow a max temp basal of 1.05u/hr not 1.5u/hr (3 X 0.5). Am I seeing it wrong?

Dana Lewis
@danamlewis
Apr 24 2018 14:01
Yes, Autotune replaces pump profile basals.
Samuchco
@Samuchco
Apr 24 2018 14:02
@danamlewis then docs need to be updated
James Babcock
@jimrandomh
Apr 24 2018 14:02
Separate from the pump profile basals, the pump contains a max basal rate, which temp basals can't exceed
That limit will only be changed if you change it yourself through the pump's UI
Dana Lewis
@danamlewis
Apr 24 2018 14:51
@Samuchco please PR away!
Michael Spradling
@M1Sports20
Apr 24 2018 14:55
scottleibrand: I also do not purposely use UAM+SMB. I get my insulin to late(novalog). However, it is nice if I every eat something and forget to account for it.
scottleibrand: I guess I do use it for small snacks. If I eat something with less than 10 carbs I let it handle it on its own
garykidd
@garykidd
Apr 24 2018 15:24
@danamlewis thanks. Yup...still April! Is the best place to follow enhancements to efficiency with pi0 hat setup in hardware-dev?
Not sure if this is something that could be in the cards for a web interface to manually adjust pump basal rates...I know there has been some discussions about a high level interface of some kind...?
live4sw
@live4sw
Apr 24 2018 15:29
@garykidd Is there a good summary of the improvements the new boards offer? The visualization is nice, of course. Is the antenna superior? Will the docs need to be updated?
Debating whether to preorder now or wait
James Babcock
@jimrandomh
Apr 24 2018 15:31
The reason for work on the Pi Hat is that the Intel Edison is discontinued, and secondary market supplies will eventually run out. The Pi is no better than the Edison overall.
Dana Lewis
@danamlewis
Apr 24 2018 15:32
@live4sw yes the docs will need additional updates
garykidd
@garykidd
Apr 24 2018 15:32
In fact, it's quite inferior---except for cost.
Dana Lewis
@danamlewis
Apr 24 2018 15:34
Erm, and actually availability on the market
@live4sw antenna is same wire whip we are using on edison rigs
Differences are mainly: Pi vs Edison. Cheaper, but less power efficient. Pi0 or Pi3 are both still on the market. has screen. YDMV on all things including preferences for rigs.
@garykidd yes hardware-dev for discussions related to HAT dev
garykidd
@garykidd
Apr 24 2018 15:47
So here is quite an embarrassing question....where can I see the changes that are included with latest dev release? For instance, I noticed that I'm about 20 commits behind...what's the easiest way to see those last changes? thx-
You can also view the Pr if you prefer to view it via desktop
James Babcock
@jimrandomh
Apr 24 2018 15:54
@garykidd You can see code changes on the dev branch at https://github.com/openaps/oref0/commits/dev . If you're looking for a written-up English summary of changes, that usually only happens on full releases.
Dana Lewis
@danamlewis
Apr 24 2018 15:54
There is a summary on the current PR
garykidd
@garykidd
Apr 24 2018 15:55
thx follks
Dana Lewis
@danamlewis
Apr 24 2018 15:55
we are prepping for release, so there are also notes there about where to focus tests, and feedback welcome (even a :+1: if all is well) so we can get 0.6.1 released soon.
We can't release without help testing all the setups and scenarios, so for those wondering where to help out, please grab a fresh dev and chime in at openaps/oref0#816
garykidd
@garykidd
Apr 24 2018 15:58
thx. will try to do so this evening
Dana Lewis
@danamlewis
Apr 24 2018 15:59
Thanks!
Lorenzo Sandini
@lsandini
Apr 24 2018 19:13
Hello all, I haven't been here for a while, and I need some help with a little feature I am adding to a rig. It's a script that makes the rig upload the preferences.json file to my ftp server. I does this by creating a new folder on my ftp server's home folder every day. The it uploads the current preferences.json file, and finally it renames it inside the destination folder adding the current date and time, so it looks like preferences -2018-04-24-21-22-13.json.
the script works fine when I run it manually, but it fails when it's triggered in cron...
the daily folder is created on the ftp server, but preferences.json is not uploaded or renamed...
Scott Leibrand
@scottleibrand
Apr 24 2018 19:17
how are you authenticating?
Lorenzo Sandini
@lsandini
Apr 24 2018 19:19
curl -T preferences.json ftp://username:password@myftpserver.mine.nu/uploads-$(date +%d-%m-%Y)/
Scott Leibrand
@scottleibrand
Apr 24 2018 19:20
try echo curl -T preferences.json ftp://username:password@myftpserver.mine.nu/uploads-$(date +%d-%m-%Y)/ and see if it correctly does all the $() stuff
Lorenzo Sandini
@lsandini
Apr 24 2018 19:21
yes it does
curl -Q "RNFR /uploads-$(date +%d-%m-%Y)/preferences.json" -Q "RNTO /uploads-$(date +%d-%m-%Y)/preferences_$(date +%d-%m-%Y-%H-%M).json" ftp://myftperver.mine.nu/uploads-$(date +%d-%m-%Y)/ --user username:password does the renaming
Scott Leibrand
@scottleibrand
Apr 24 2018 19:26
Rather than issuing remote FTP commands like that, I would cp the file locally to the new name before uploading it
Lorenzo Sandini
@lsandini
Apr 24 2018 19:28
... and then delete the locally created copy to avoid filling the rig with copies of the pref file :)
I'll try your approach Scott, but do you have any idea why it works fine when I issue the bash command, and it fails half-way when called by cron ?
Scott Leibrand
@scottleibrand
Apr 24 2018 20:01
for one thing, cron doesn't know about your PATH unless you explicitly specify it in crontab
if you just put the copied file in /tmp/, it'll get cleaned up automatically on reboot
Lorenzo Sandini
@lsandini
Apr 24 2018 20:12
thanks a lot ! I had not thought about the path ! I now added ~/myopenaps/ where appropriate and with this line in my script, the rigs calls home and uploads the file as intended, before it renames it :) curl -T ~/myopenaps/preferences.json ftp://username:password@myftpserver.mine.nu/uploads-$(date +%d-%m-%Y)/
Scott Leibrand
@scottleibrand
Apr 24 2018 20:16
ah, yeah. not even PATH in that case, but working directory.
you probably do have a PATH added by oref0-setup
Lorenzo Sandini
@lsandini
Apr 24 2018 20:18
:+1:
Scott Leibrand
@scottleibrand
Apr 24 2018 20:22
the oref0-setup crontab lines do a lot of cd ~/myopenaps && whatever-else