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

2nd
Jan 2016
Scott Leibrand
@scottleibrand
Jan 02 2016 00:00 UTC
look at (paste) the requestedtemp.json
as @dm61 mentioned earlier, I need to write up what everything in those files means
diabeticgonewild
@diabeticgonewild
Jan 02 2016 00:01 UTC
I don't have a requestedtemp.json file
maybe it's in /enact/
idk
let me look. There really needs to be some housekeeping done on this (and a checklist)...
Scott Leibrand
@scottleibrand
Jan 02 2016 00:02 UTC
I put my output of oref0-determine-basal into enact/, ya
Dana Lewis
@danamlewis
Jan 02 2016 00:02 UTC
PR away, @diabeticgonewild :)
diabeticgonewild
@diabeticgonewild
Jan 02 2016 00:02 UTC
{"temp":"absolute","bg":255,"tick":"+4","eventualBG":237,"snoozeBG":237,"reason":"$
Scott Leibrand
@scottleibrand
Jan 02 2016 00:02 UTC
we're not trying to exactly specify how everyone has to do it the same. everyone sets up their loop slightly differently, which is fine
that looks incomplete: it's not valid json
diabeticgonewild
@diabeticgonewild
Jan 02 2016 00:03 UTC
I know, but it's not clean....It may be awhile till I can PR this (a few days)
{"temp":"absolute","bg":255,"tick":"+4","eventualBG":237,"snoozeBG":237,"reason":"basal_iob 1.386 > max_iob 0","duration":0,"rate":0}
Scott Leibrand
@scottleibrand
Jan 02 2016 00:03 UTC
if we wanted something clean, we would make a setup script that does everything the same for everyone. but that wouldn't be safe, because people wouldn't be building it themselves and wouldn't learn how to troubleshoot
diabeticgonewild
@diabeticgonewild
Jan 02 2016 00:03 UTC
There we go...
Scott Leibrand
@scottleibrand
Jan 02 2016 00:03 UTC
cool
diabeticgonewild
@diabeticgonewild
Jan 02 2016 00:04 UTC
No, but the docs are disorganized.
Scott Leibrand
@scottleibrand
Jan 02 2016 00:04 UTC
so it's saying it's not high-temping you, because you haven't configured a max_iob, so it's defaulting to zero
diabeticgonewild
@diabeticgonewild
Jan 02 2016 00:04 UTC
It said to set it to : 0
Scott Leibrand
@scottleibrand
Jan 02 2016 00:04 UTC
with this setup, it will low-temp as needed to avoid lows, but won't high-temp IOB above zero
Dana Lewis
@danamlewis
Jan 02 2016 00:05 UTC
@diabeticgonewild if you'd like to suggest reorganization, please do so with a PR when you get to it so we can discuss any suggestions for improvement that way. otherwise, it's not very efficient to just complain about what's there - that doesn't help us all move forward as a community.
diabeticgonewild
@diabeticgonewild
Jan 02 2016 00:06 UTC
That should be in the glossary too...
Jason Calabrese
@jasoncalabrese
Jan 02 2016 00:06 UTC
what version of decocare has the year fix, updated openaps and saw it pulled in decocare 0.0.21
Scott Leibrand
@scottleibrand
Jan 02 2016 00:06 UTC
if you want it to high-temp you, you can set max_iob (which is just basal IOB, not bolus IOB) to something higher. right now your basal_iob is 1.386, so 1.5 or 2 might be a good place to start if you think it should high-temp in this situation.
0.0.20 I believe had the fix. @bewest did another release (0.0.21) immediately after
diabeticgonewild
@diabeticgonewild
Jan 02 2016 00:07 UTC
K
Ben West
@bewest
Jan 02 2016 00:31 UTC
0.0.20 has the y2k fix
0.0.21 has outstanding dev
Jason Calabrese
@jasoncalabrese
Jan 02 2016 00:33 UTC
great
Ben West
@bewest
Jan 02 2016 00:51 UTC
the parent directory thing would be a good first-timer issue
add code there to make sure the path/directory to report.name is doable
something like mkdir_p(path.dirname(report.name))
Dana Lewis
@danamlewis
Jan 02 2016 01:00 UTC
Just added a couple of things to issues re: things we know need to be documented in the future. @scottleibrand is assigned to #57 about documenting the variables and determine-basal output. I also logged a couple that people can tackle in the future related to various sensors and receivers that we know work, but just need documentation added. You can see them (and add more if you know of them) here: https://github.com/openaps/docs/issues
Ben West
@bewest
Jan 02 2016 01:01 UTC
productive day for openaps
Dana Lewis
@danamlewis
Jan 02 2016 01:01 UTC
:) indeed.
Just submitted #58 (openaps/docs#58) to re-iterate having people create issues if they find something that is not documented (or well) and don't yet feel comfortable creating a full PR to document it themselves.
Andrew Warrington
@andrew-warrington
Jan 02 2016 01:16 UTC
@sulkaharo, just deployed Nightscout to Heroku per your advice,+
went off without a hitch. Thank you!
happy to be closing my Azure account.
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 01:20 UTC
Instead of connecting dexcom directly to RPi, can I use NS data? I mean is there already a "vendor" for this or should I ude process -> bash -> curl for this?
Dana Lewis
@danamlewis
Jan 02 2016 01:22 UTC
yes, you can use NS data. (:: makes note to add this to the docs::)
(My note: openaps/docs#55)
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 01:25 UTC
...I should have been looking to oref0 before asking :) thank you
Dana Lewis
@danamlewis
Jan 02 2016 01:26 UTC
it's not well documented, though.
Scott Leibrand
@scottleibrand
Jan 02 2016 01:28 UTC
we need someone to volunteer to take things out of that script and put them into the appropriate places in the docs.
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 01:31 UTC
I'd live to but my engligsh skills are under sink level :(
Scott Leibrand
@scottleibrand
Jan 02 2016 01:32 UTC
it's mostly not english, just code to copy-paste into the appropriate places. and it's a lot easier to get someone to fix your english than to do the hard work of figuring out what needs documented where and how
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 01:34 UTC
hm, ok, I'll try, but before doing this I must finish understanding the system. But I promise that after I will be familiar with the system I'll try to dispatch your setup script and will put pieces of it in documentation
Scott Leibrand
@scottleibrand
Jan 02 2016 01:34 UTC
sounds like a plan!
Dana Lewis
@danamlewis
Jan 02 2016 01:35 UTC
:)
Scott Leibrand
@scottleibrand
Jan 02 2016 01:35 UTC
fwiw, your english is actually quite good. a lot better than my spanish, and that's the only other language I can speak at all. :)
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 02:45 UTC
Is there a already-made way to upload pump basal-profile to NS?
Scott Leibrand
@scottleibrand
Jan 02 2016 03:03 UTC
I'm not aware of one. If there is we should add it to our loop.
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 03:04 UTC
ben has written something: mm-format-ns-profile
now I am trying to refresh my bash memories to try to make it work :)
Ben West
@bewest
Jan 02 2016 03:06 UTC
mm-format-ns-profile monitor/settings.json monitor/carb-ratios.json monitor/active-basal-profile.json monitor/insulin-sensitivities.json monitor/bg-targets.json
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 03:07 UTC
and all these json files are results from openaps reports, right?
Ben West
@bewest
Jan 02 2016 03:08 UTC
yeah, they are all reports from medtronic vendor
or medtronic uses if you like
could do add another use specifically to aggregate these things
I didn't know they were all needed together until I knew what worked in NS
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 03:10 UTC
as you need them basically only once it's ok to leave them as they are :)
Ben West
@bewest
Jan 02 2016 03:10 UTC
yeah...
could be part of tutorial or something
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 03:11 UTC
yes, kind of initial se up, if I will manage to get it working I'll update docs
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 03:57 UTC
hm....
SETTINGS=${1-monitor/settings.json}
CARBS=${2-monitor/carb-ratios.json}
BASALRATES=${3-monitor/active-basal-profile.json}
SENSITIVITIES=${4-monitor/insulin-sensitivities.json}
TARGETS=${4-monitor/bg-targets.json}
OUTPUT=${2-/dev/fd/1}
ok, so it is that old notation with n-th argument or default value
but why in this case 4 are for sensitivities and for targets?
Scott Leibrand
@scottleibrand
Jan 02 2016 03:58 UTC
Looks like someone forgot to update the last two
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 03:58 UTC
or it is just an error?
:)
Scott Leibrand
@scottleibrand
Jan 02 2016 03:58 UTC
targets should be $5 and output $6 I'd think.
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 03:59 UTC
yes, I think too :) let me try
and yes, it means reinstalling from sources...
Peter Miller
@ochenmiller
Jan 02 2016 04:03 UTC
Hank III is almost done (Grandson to Hank the Pank, Son of Hank Jr) - just need to make a case, 3"x2"x1" including the battery! Should travel well in pocket or on belt. Very very excited.
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 04:22 UTC
@bewest In NS what endpoint (/api/v1/???) is able to receive profile data?
Ben West
@bewest
Jan 02 2016 04:23 UTC
profile i tink
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 04:43 UTC
I don't understand it.... I receive "api-secret Request Header is incorrect or missing."
curl when launched with -v shows my api secret in headers
uploader works with same api and same secret
Host: ktomy.herokuapp.com
Accept: /
API-SECRET: <12 chars>

content-type: application/json

Content-Length: 3368
Expect: 100-continue

< HTTP/1.1 100 Continue
< HTTP/1.1 401 Unauthorized

Alin Artiom Kenibasov
@ktomy
Jan 02 2016 05:15 UTC
Just modified verify-token.js to check what is compared to my api_secret and found out that my api_secret is compared to something completely different, maybe with a hashed key or something
Yes! I managed to upload the profile to NS! I'll PR my changes in the evening... it was a really long night :)
Peter Miller
@ochenmiller
Jan 02 2016 05:23 UTC
@ktomy, Congrats!
Ben West
@bewest
Jan 02 2016 07:17 UTC
congrats @ktomy, sorry for confusion RE hashed vs plain API_SECRET
Sulka Haro
@sulkaharo
Jan 02 2016 09:18 UTC
I recall even the source confuses the naming between the hashed and secret somewhere.
About to try to get a Wixel going today. Should PR a little addition to the soldering instructions - I bet most people creating one are trying to solder without flux and using lead-free solder, which is pure hell on manual soldering. Flux + good old leaded tin makes even HM-11 an easy task. :)
Toby Canning
@TC2013
Jan 02 2016 15:01 UTC
I have a question about the watchdog service. I haven't been enabling it, I suppose because our loop is running really well. Do others here believe it is necessary?
Toby Canning
@TC2013
Jan 02 2016 15:28 UTC
Also, it appears I need to move away from Azure. Should I use digital ocean or heroku?
Toby Canning
@TC2013
Jan 02 2016 15:49 UTC
Okay, I decided on heroku and wow! 15 minutes to setup and install. That was amazingly simple.
Andrew Warrington
@andrew-warrington
Jan 02 2016 16:08 UTC
@TC2013 , I had the exact same experience yesterday. what an improvement, right? the website also seems more responsive. that could be bias though, I didn't measure. :)
To all: I'm working on reducing my power consumption. Found this which allows port-by-port power control for USB and Ethernet ports on the Pi 2. Can switch ports on at the beginning of a loop, and then off at the end.
plus leave some unused ports off all this time if you like
a thread about using this with the the RPi2 specifically:
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 16:44 UTC
One more thing to do for profile import from pump is that in SN values are stored in mg/dl and on pump values can be stored in mmol (as it is in my case). Will try to make an issue on NS sourcetree nad I hope my JS skills will permit me to solve it. But in any case its priority is low as there are only a few values which need to be corrected by hand.
Tony Zarro
@tzarro
Jan 02 2016 16:46 UTC
Hoping someone can help with this one: nearly every report and alias test is working thus far, except the alias enact. I am now getting this error: "openaps-report invoke: error: argument report: invalid choice: 'enact/enacted.json'". Any ideas?
is there supposed to be an enact/enacted.json report definition, and if so what are the parameters recommended?
Toby Canning
@TC2013
Jan 02 2016 16:49 UTC
what does the cat of enacted.json show?
I think this may be similar to an error I received until the first time the loop set a temp
Hichame Yessou
@hichameyessou
Jan 02 2016 16:49 UTC
Sorry for the o.t./newbie question, but I was thinking if there is any way to use the cc1111 with a Minimed 754 firmware 3.1 for an openaps
Toby Canning
@TC2013
Jan 02 2016 16:50 UTC
Firmware has to be 2.6A or lower on the 754 EU
Tony Zarro
@tzarro
Jan 02 2016 16:50 UTC
@TC2013 I don't have an output of enacted.json, I'll create one...that makes perfect sense, thanks!
@tzarro Once you have a temp run once, it will always save the last one
    {
  "bg": 115,
  "temp": "absolute",
  "snoozeBG": 69,
  "recieved": true,
  "rate": 0.0,
  "reason": "Eventual BG 5<110, no temp, setting -0.68U/hr",
  "eventualBG": 5,
  "timestamp": "2016-01-02T10:48:44.077329",
  "duration": 30,
  "tick": -3
}
Hichame Yessou
@hichameyessou
Jan 02 2016 16:54 UTC
I know, but there is no way to make it with this firmware? Its not possible to do a firmware downgrade?
Toby Canning
@TC2013
Jan 02 2016 16:55 UTC
As far as I know only the factory has the ability to change firmwares
However, this would be awesome if we developed this capability... though potentially very dangerous too.
I've often thought about trying to find video game console hackers because this would be right within their expertise
Hichame Yessou
@hichameyessou
Jan 02 2016 17:05 UTC
I was thinking that maybe could be possible to reverse engineer the firmware from an "old" pump if its not availabe and burn it with a kind of jtag or uart cable
Or anyone has ever tryed to request medtronic to have a pump with an older firmware?
Tony Zarro
@tzarro
Jan 02 2016 17:10 UTC
@TC2013 apparently I'm missing the report configuration for enact/enacted.json...trying to add it now, but what is the device I should cite in the parameters?
Sulka Haro
@sulkaharo
Jan 02 2016 17:37 UTC
Hey @scottleibrand and @jasoncalabrese I looked at porting the OpenAPS iob algorithm over to CGM-remote-monitor. Copy-paste would be easy, but would result in code going out of sync. Creating an OpenAPS npm dependency makes me itchy for some reason. Would it be silly to create an OpenIOB npm module and have both depend on that?
Scott Leibrand
@scottleibrand
Jan 02 2016 17:50 UTC
@jasoncalabrese wanted to create a "dcalc" thing (not sure if it would be an npm module or what) to allow them to share code...
@tzarro take a look at openaps pump use -h to find the command to enact a temp, and then set up a report to do it. The input should be the output of oref0-determine-basal.
Sulka Haro
@sulkaharo
Jan 02 2016 18:11 UTC
Oh right. And indeed this already exists and the old iob code is in there.
Scott Leibrand
@scottleibrand
Jan 02 2016 18:13 UTC
Well, the old IOB code from master doesn't do temp basals. And the wip/iob-cob code is now incompatible. So sounds like you're thinking along the same lines.
Dana Lewis
@danamlewis
Jan 02 2016 18:22 UTC
@hichameyessou Medtronic will not sell the older pumps in the US.
Oskar Pearson
@oskarpearson
Jan 02 2016 18:36 UTC
@hichameyessou You might find that there's a way to read the firmware from a pump and write it to another pump with jtag - but nobody's tried it so far as we know. http://eddiesamuels.tumblr.com/post/57996415497/minimed-insulin-pump-teardown has a teardown. @bewest had some pictures of the inside of a pump too somewhere, with focus on the model numbers
Dana Lewis
@danamlewis
Jan 02 2016 19:57 UTC
Just shared the winter update post on openaps.org - http://bit.ly/1OArmTj
If there's anything I left off, or someone didn't see my post a few days ago about sending me their update, I can always add to it.
Tony Zarro
@tzarro
Jan 02 2016 21:20 UTC
@danamlewis loved the winter update! Especially interested in the work folks are doing on the hardware side...
Andrew Warrington
@andrew-warrington
Jan 02 2016 21:24 UTC
yes, thanks @danamlewis for all the work you do on communication. OpenAPS would die without someone diligently putting in the effort to keep everyone up to date and engaged.
Dana Lewis
@danamlewis
Jan 02 2016 21:47 UTC
:)
Dana Lewis
@danamlewis
Jan 02 2016 21:53 UTC
Fun to see the evolution over the past year and how all the activity has blossomed (I was going to say "blown up" but that could be taken incorrectly ;)) in last month or two in particular.
Dragan Maksimovic
@dm61
Jan 02 2016 21:58 UTC
Couple of quick questions: during a square-wave bolus, the pump won't accept temp basal commands; in other words, a square-wave bolus effectively shuts down the OpenAPS loop; is this correct? second, are square-wave boluses taken into account in IOB calculations?
@danamlewis great update, thanks!
Rachel Sandlain
@audiefile
Jan 02 2016 22:02 UTC
Correct, no temp basal commands while bolusing, including square wave
Dana Lewis
@danamlewis
Jan 02 2016 22:09 UTC
And square waves are not yet in IOB calc
Scott Leibrand
@scottleibrand
Jan 02 2016 22:19 UTC
Well, they are, but they're counted as if they were manual boluses at the time the bolus record entry is recorded, which I believe is when the extended bolus ends.
Most people using OpenAPS don't use square wave boluses (since they disable the loop) and instead set an appropriate initial bolus and then let the loop take over. The meal-assist code we just added to dev really helps with that.
Tony Zarro
@tzarro
Jan 02 2016 22:21 UTC
Getting another error I'm hoping someone can help with...when running the enact alias:
pi@aps:~/openaps $ openaps enact
determine-basal://text/shell/enact/suggested.json
{"delta":13,"glucose":189,"avgdelta":13}
{"duration":0,"rate":3,"temp":"absolute"}
{"iob":0,"activity":0,"bolusiob":0}
{"max_iob":6,"type":"current","dia":3,"current_basal":2.5,"max_daily_basal":3,"max_basal":7,"min_bg":90,"max_bg":100,"sens":10}
reporting enact/suggested.json
{"temp":"absolute","bg":189,"tick":"+13","eventualBG":228,"snoozeBG":228,"reason":"max_iob 6, adj. req. rate:14.5 to maxSafeBasal:7.0, no temp, setting 7U/hr","duration":30,"rate":7}
pump://JSON/set_temp_basal/enact/enacted.json
enact/enacted.json raised byte must be in range(0, 256)
Traceback (most recent call last):
File "/usr/local/bin/openaps-report", line 4, in <module>
import('pkg_resources').run_script('openaps==0.0.9', 'openaps-report')
File "build/bdist.linux-armv7l/egg/pkgresources/_init.py", line 745, in run_script
File "build/bdist.linux-armv7l/egg/pkgresources/_init.py", line 1670, in run_script
File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/EGG-INFO/scripts/openaps-report";, line 82, in <module>
app( )
File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/openaps/cli/init.py", line 44, in call
self.run(self.args)
File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/EGG-INFO/scripts/openaps-report";, line 75, in run
output = app(args, self)
File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/openaps/cli/subcommand.py";, line 50, in call
return self.method.main(args, app)
File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/openaps/reports/invoke.py";, line 40, in main
output = task.method(args, app)
File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/openaps/uses/use.py";, line 45, in call
output = self.main(args, app)
File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/openaps/vendors/medtronic.py";, line 311, in main
results = self.upload_program(program)
File "/usr/local/lib/python2.7/dist-packages/openaps-0.0.9-py2.7.egg/openaps/vendors/medtronic.py";, line 320, in upload_program
return self.pump.model.set_temp_basal(program)
File "/usr/local/lib/python2.7/dist-packages/decocare-0.0.21-py2.7.egg/decocare/models/init.py", line 188, in set_temp_basal
result = self._set_temp_basal(
basals)
File "/usr/local/lib/python2.7/dist-packages/decocare-0.0.21-py2.7.egg/decocare/models/init.py", line 25, in call
self.response = inst.session.query(self.msg, **kwds)
File "/usr/local/lib/python2.7/dist-packages/decocare-0.0.21-py2.7.egg/decocare/session.py";, line 101, in query
self.execute(command)
File "/usr/local/lib/python2.7/dist-packages/decocare-0.0.21-py2.7.egg/decocare/session.py";, line 98, in execute
return super(type(self), self).execute(command)
File "/usr/local/lib/python2.7/dist-packages/decocare-0.0.21-py2.7.egg/decocare/session.py";, line 35, in execute
self.transfer( )
File "/usr/local/lib/python2.7/dist-packages/decocare-0.0.21-py2.7.egg/decocare/session.py";, line 61, in transfer
return self.stick.transmit_packet(self.command)
File "/usr/local/lib/python2.7/dist-packages/decocare-0.0.21-py2.7.egg/decocare/stick.py";, line 856, in transmit_packet
result = self.process( )
File "/usr/local/lib/python2.7/dist-packages/decocare-0.0.21-py2.7.egg/decocare/stick.py";, line 485, in process
raw = self.send_force_read( )
File "/usr/local/lib/python2.7/dist-packages/decocare-0.0.21-py2.7.egg/decocare/stick.py";, line 601, in send_force_read
self.link.write(reader.format( ))
File "/usr/local/lib/python2.7/dist-packages/decocare-0.0.21-py2.7.egg/decocare/stick.py";, line 400, in format
return bytearray(packet)
ValueError: byte must be in range(0, 256)
No action required
Never seen an error like that before (byte must be in range), any ideas?
Scott Leibrand
@scottleibrand
Jan 02 2016 22:23 UTC
Not sure, but your max basal is way too high.
Hmm or maybe not. Your ISF is really 10 mg/dL?
That's 4x lower than I'm used to.
Do you get that error every time?
Dragan Maksimovic
@dm61
Jan 02 2016 22:26 UTC
@scottleibrand I am not sure that's the case; I'll check and get back to you on how I think square-waves are recorded. I understand the logic behind not using square-waves when openAPS loop is working. However, there could cases when the user simply 'knows better' and would prefer to issue a square wave and disable openAPS loop for some time. Not a high priority, but I would suggest looking into this some more at some point.
Scott Leibrand
@scottleibrand
Jan 02 2016 22:26 UTC
@tzarro if not, it may just be a communications error. If so, it might be an upper limit on temp basal rate that decocare knows how to decode.
Dana Lewis
@danamlewis
Jan 02 2016 22:27 UTC
@dm61 yea, worst case you do what you need to do, and once it's done the loop can kick back in. big concern would just be if you had late dinner and went to bed with a super long square running? rare probably, though.
Scott Leibrand
@scottleibrand
Jan 02 2016 22:28 UTC
@dm61 yeah. If there is someone who gets better results with square waves than meal-assist and wants to use them, properly decoding them and incorporating them into IOB wouldn't be that difficult.
Just haven't had any demand for that yet.
Dragan Maksimovic
@dm61
Jan 02 2016 22:28 UTC
@danamlewis I think if square-waves were included in IOB calcs everything would be just fine
Scott Leibrand
@scottleibrand
Jan 02 2016 22:29 UTC
More likely that'll come first from Nightscout, where most people aren't looping, but may be entering their extended boluses.
Once we have dcalc in place, we'll want to add extended bolus support into NS, which we will then get in oref0 as well pretty much for free.
Dragan Maksimovic
@dm61
Jan 02 2016 22:30 UTC
What is dcalc?
Scott Leibrand
@scottleibrand
Jan 02 2016 22:31 UTC
A proposed common diabetes calculations library to be shared between NS and oref0
and any other js-based d-tools
Dragan Maksimovic
@dm61
Jan 02 2016 22:35 UTC
ok, thanks. Is 'meal assist' in the current master?
Dana Lewis
@danamlewis
Jan 02 2016 22:37 UTC
no, just in dev right now
Dragan Maksimovic
@dm61
Jan 02 2016 22:38 UTC
ok
Dana Lewis
@danamlewis
Jan 02 2016 22:44 UTC
@TC2013 do you still say the same about over-bolusing, with meal-assist?
Toby Canning
@TC2013
Jan 02 2016 22:45 UTC
Well, that's all we've ever used
Dana Lewis
@danamlewis
Jan 02 2016 22:46 UTC
ah, gotcha. wasn't sure when you switch to oref0 and if you had done dev or just master
Toby Canning
@TC2013
Jan 02 2016 22:46 UTC
Just dev and previously the meal-assist branch
Because of how it functions of the deltas (BGI), it seems to be great for our situation
Christian forgot to bolus yesterday and it couldn't keep up with that
Dana Lewis
@danamlewis
Jan 02 2016 22:47 UTC
Do you have an example with carbs and type of bolus you would do? @scottleibrand is wondering if meal-assist is not agressive enough. I'm not the ideal target for meal-assist usually, so I'm not testing the same way I have other features we've built.
Scott Leibrand
@scottleibrand
Jan 02 2016 22:48 UTC
I don't want meal-assist to be able to keep up with an unannounced meal (though it will bring BG down eventually).
Toby Canning
@TC2013
Jan 02 2016 22:48 UTC
:) right!
Scott Leibrand
@scottleibrand
Jan 02 2016 22:49 UTC
but I'd like to figure out if we can ever safely do more insulin after a bolus
Dana Lewis
@danamlewis
Jan 02 2016 22:49 UTC
also wondering if @tghoward @sarhoward have tested meal-assist yet
Toby Canning
@TC2013
Jan 02 2016 22:50 UTC
Let me really pay attention to meal-assist over the next few days. We have been "over-bolusing" to keep BGs flat because OpenAPS does such a good job at removing any excess insulin
Sulka Haro
@sulkaharo
Jan 02 2016 22:51 UTC
Actually right now in process of adding the needed stuff to dcalc to make it an importable node module for iob. Something bugging out in module.exports though :)
Dana Lewis
@danamlewis
Jan 02 2016 22:51 UTC
Sounds good @TC2013 , thanks
Scott Leibrand
@scottleibrand
Jan 02 2016 22:51 UTC
Thx @sulkaharo !
Toby Canning
@TC2013
Jan 02 2016 22:51 UTC
We have more wiggle room than average. Christian's meal BGs are typicaly 10-12U and with an hourly basal rate average of 3U/hr, OpenAPS can quickly remove enough extra to come in for a soft landing... Previously all meals would take him up into the 180-220 range
Andrew Warrington
@andrew-warrington
Jan 02 2016 22:52 UTC
question for those on a closed loop: following a meal where you've given a bolus upfront, how soon do you get back to normal on average? e.g. 1 hour later, 2 hours, do you expect no peak at all? etc. I'm not asking about the best case where the carbs are perfectly estimated, but rather about real-world experience.
Dana Lewis
@danamlewis
Jan 02 2016 22:53 UTC
Wow. Yea. My avg meal bolus is probably 5-6u, basals are .85-1.3-ish. Nice that it works regardless of ratios and meal consumption :)
@andrew-warrington so, it's going to depend a lot on a pre-bolus or "eating soon" mode
if I have eating soon triggered properly, I either don't spike and/or am back within an hour.
Toby Canning
@TC2013
Jan 02 2016 22:53 UTC
Over course it can't do it, but I see OpenAPS recommend temps of 12U/hr
Dana Lewis
@danamlewis
Jan 02 2016 22:53 UTC
if I don't, might be more to 2 hours, and always spike.
and @andrew-warrington I think answers to your question will vary depending also on whether someone has tried meal-assist in dev yet or not?
Andrew Warrington
@andrew-warrington
Jan 02 2016 22:54 UTC
got it. right, so far I only have my pre-loop experience to look at, where it was frequently 2-3 hours.
Dana Lewis
@danamlewis
Jan 02 2016 22:55 UTC
big thing when looping is making sure you don't go into a meal with negative IOB. then you're going to spike for sure.
Andrew Warrington
@andrew-warrington
Jan 02 2016 22:56 UTC
actually I think that holds for diabetes overall, doesn't it?
Dana Lewis
@danamlewis
Jan 02 2016 22:56 UTC
correct
Sulka Haro
@sulkaharo
Jan 02 2016 22:56 UTC
Ha, our total daily insulin is usually under 10 units. Meals are 0.9 IU and basal goes all the way up to 0.3 IU :)
Dana Lewis
@danamlewis
Jan 02 2016 22:56 UTC
yes, @sulkaharo your PWD is also what percentage of our size? ;) :)
::feels like a giant in comparison::
Sulka Haro
@sulkaharo
Jan 02 2016 22:57 UTC
He's a big lad at.. around 16 kg :)
Andrew Warrington
@andrew-warrington
Jan 02 2016 22:57 UTC
:P oh to be young again...
Dana Lewis
@danamlewis
Jan 02 2016 22:57 UTC
lol
Tony Zarro
@tzarro
Jan 02 2016 22:57 UTC
@scottleibrand regarding my "byte must be in range" error, yes, unfortunately my ISF is very low compared to average...it has been decreasing over the years...
Rachel Sandlain
@audiefile
Jan 02 2016 22:58 UTC
A
Scott Leibrand
@scottleibrand
Jan 02 2016 22:59 UTC
@tzarro did it keep happening, or just that once?
Sulka Haro
@sulkaharo
Jan 02 2016 22:59 UTC
Arght total core brainfart with the module.exports. Programming at 1 AM after a glass of wine is awesome. :D
Toby Canning
@TC2013
Jan 02 2016 23:00 UTC
wine got me through much of my pi setup
Scott Leibrand
@scottleibrand
Jan 02 2016 23:00 UTC
@TC2013 it wants 12U/hr when, during meal-assist? What's that, 4x normal?
Toby Canning
@TC2013
Jan 02 2016 23:00 UTC
Yes
Rachel Sandlain
@audiefile
Jan 02 2016 23:00 UTC
@danamlewis and others who pre bolus or eat soon mode and loop, how do you prevent the loop from undoing what you did by low temping?
Sulka Haro
@sulkaharo
Jan 02 2016 23:01 UTC
Uh oh, checked out cgm-remote-monitor IOB tests to steal stuff from there. Not good..
Scott Leibrand
@scottleibrand
Jan 02 2016 23:01 UTC
Bolus snooze will prevent low temping in most cases.
Tony Zarro
@tzarro
Jan 02 2016 23:01 UTC
@scottleibrand it has kept happening...and only started happening once I trued up the settings on my old pump (compatible with OpenAPS) to the settings configured on my current pump, including insuliln sensitivities...
Sulka Haro
@sulkaharo
Jan 02 2016 23:01 UTC
Or rather, could perhaps maybe be more extensive
Dragan Maksimovic
@dm61
Jan 02 2016 23:01 UTC
I have so far closed the loop only overnight, because to me it looked like it could actually do more harm than good during day; but, I have not tried the meal-assist yet - am curious to find out how it works (just for reference, my 'all manual' control is pretty good - not sure if openAPS would be able to do any better; well, except during night, when I am out of the loop :smile: )
Scott Leibrand
@scottleibrand
Jan 02 2016 23:01 UTC
I think the oref0 IOB tests do a bit more.
Jason Calabrese
@jasoncalabrese
Jan 02 2016 23:01 UTC
think they are about the same
very basic
Toby Canning
@TC2013
Jan 02 2016 23:02 UTC
Is there a doc that describes how to best determine max IOB?
Dana Lewis
@danamlewis
Jan 02 2016 23:02 UTC
@audiefile bolus snooze prevents it from temping and counteracting an eating soon/prebolus.
Scott Leibrand
@scottleibrand
Jan 02 2016 23:02 UTC
@tzarro try lowering your max_basal and see if it can do lower rates.
Tony Zarro
@tzarro
Jan 02 2016 23:02 UTC
will do
Dana Lewis
@danamlewis
Jan 02 2016 23:03 UTC
@dm61 I feel the same, loop still helps during the day when I'm busy doing other things, but meal-assist hasn't been as significant for me as it has been for others. will look forward to your feedback if you try it some time in dev.
Scott Leibrand
@scottleibrand
Jan 02 2016 23:03 UTC
We do 2 or 2.5 for max_iob. Not sure there's anything scientific about it. :-)
Toby Canning
@TC2013
Jan 02 2016 23:03 UTC
Is the 6.35/U per hour limit hard coded into OpenAPS?
Dana Lewis
@danamlewis
Jan 02 2016 23:03 UTC
@TC2013 I don't think we've written anything up on it.
Toby Canning
@TC2013
Jan 02 2016 23:03 UTC
Yay, 2 or 2.5 wouldn't do much for us
Rachel Sandlain
@audiefile
Jan 02 2016 23:03 UTC
Not always in my experience. I don't have notes but I've seen low temps attempting to undo if I pre bolus part or all of a bolus
Scott Leibrand
@scottleibrand
Jan 02 2016 23:03 UTC
@TC2013 what 6.35U/hr limit?
Toby Canning
@TC2013
Jan 02 2016 23:03 UTC
temp basal limit
Jason Calabrese
@jasoncalabrese
Jan 02 2016 23:04 UTC
max safe basal?
Tony Zarro
@tzarro
Jan 02 2016 23:04 UTC
@scottleibrand what is max_iob actually used for? That's one of the few data elements I'm not sure of...
Scott Leibrand
@scottleibrand
Jan 02 2016 23:04 UTC
Max safe basal is set based on your max daily basal and current basal
Rachel Sandlain
@audiefile
Jan 02 2016 23:04 UTC
Dana do you manually change your target for eating soon and then back up again after eating?
Jason Calabrese
@jasoncalabrese
Jan 02 2016 23:05 UTC
thats the limit I run into more, don't think I've hit max basal iob of 2
Scott Leibrand
@scottleibrand
Jan 02 2016 23:05 UTC
max_iob should be renamed to max_basal_iob. It's the highest that oref0 will let basal iob get.
Toby Canning
@TC2013
Jan 02 2016 23:05 UTC
OpenAPS/decocare can't issue a temp beyond 6.35 because of a bit limitation .05 x 256 = 6.4U
Scott Leibrand
@scottleibrand
Jan 02 2016 23:05 UTC
Ah
Is there another bit we haven't found yet?
or is that a MDT limit?
Toby Canning
@TC2013
Jan 02 2016 23:06 UTC
Well, it is still an issue in the decocare repo
I think it is bit related
Dana Lewis
@danamlewis
Jan 02 2016 23:06 UTC
@audiefile not in the pump, but that's one way you could do it. I do it in my head/or use DIYPS (still doesn't change it in my pump but just visualizes what I need as a result of switching my target to 80 before the meal)
Scott Leibrand
@scottleibrand
Jan 02 2016 23:06 UTC
so @tzarro set your max_basal to 6.35
Tony Zarro
@tzarro
Jan 02 2016 23:07 UTC
ok
Scott Leibrand
@scottleibrand
Jan 02 2016 23:07 UTC
Lots of questions. :-) Did I miss any?
Dragan Maksimovic
@dm61
Jan 02 2016 23:10 UTC
Here is another one :smile: : can min_bg be anything less than 90 in openaps?
Dana Lewis
@danamlewis
Jan 02 2016 23:11 UTC
Highly don't recommend it
Scott Leibrand
@scottleibrand
Jan 02 2016 23:11 UTC
I believe I hard-coded a minimum of 90 or thereabouts.
Dragan Maksimovic
@dm61
Jan 02 2016 23:11 UTC
Looks like it is hard coded - it ignored my pump setting at 80
Scott Leibrand
@scottleibrand
Jan 02 2016 23:12 UTC
You probably want it to be at least 100
Rachel Sandlain
@audiefile
Jan 02 2016 23:12 UTC
@danamlewis cool thanks. I've been working on taking at least some insulin earlier before meals to help with spiking. Still trying to find a method that works for me.
Scott Leibrand
@scottleibrand
Jan 02 2016 23:12 UTC
at 110 we saw it keep actual BG above 80
but if it were 90 it couldn't necessarily keep it above 60, which is too low.
Although the low glucose suspend would still kick in at ~75, it wouldn't take action early enough when dropping with iob.
Dragan Maksimovic
@dm61
Jan 02 2016 23:13 UTC
well, it must all depend on person to person, and on so many other external factors - it would be difficult to generalize
Scott Leibrand
@scottleibrand
Jan 02 2016 23:14 UTC
I don't think the loop is likely to be able to keep anyone above 80 with a min_bg of 90. That would require glucagon.
Dragan Maksimovic
@dm61
Jan 02 2016 23:14 UTC
But, I do agree that safer is better
Scott Leibrand
@scottleibrand
Jan 02 2016 23:15 UTC
(Or fast insulin with a DIA of ~1h)
Toby Canning
@TC2013
Jan 02 2016 23:15 UTC
Is that coming out?
Scott Leibrand
@scottleibrand
Jan 02 2016 23:16 UTC
Not that fast.
There's a faster novorapid in FDA approval.
Dragan Maksimovic
@dm61
Jan 02 2016 23:16 UTC
I've been waiting for that for 40 years now
Scott Leibrand
@scottleibrand
Jan 02 2016 23:17 UTC
But its DIA is similar to novolog, it just kicks in earlier (~15m instead of 30)
So it'll definitely help blunt meal spikes.
It's the same drug (insulin aspart) with some adjuvants that help it get out into the bloodstream faster.
Tony Zarro
@tzarro
Jan 02 2016 23:25 UTC
@scottleibrand enact is working again with my max_basal set to 6.35...thanks!
Sulka Haro
@sulkaharo
Jan 02 2016 23:27 UTC
Anyone here with a xDrip using HM-11? Having issues, no documentation available...
Rachel Sandlain
@audiefile
Jan 02 2016 23:29 UTC
Issues with what?
Sulka Haro
@sulkaharo
Jan 02 2016 23:38 UTC
The BLE module comms are maybe off. BLE shows up as "AnyFlite". Found a german page that indicated the baud rate of comms between the Wixel and the BLE unit might need adjusting.
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 23:49 UTC
@danamlewis jfyi, openaps.ini from your ns-upload-setup will not for for first time if you don't have any treatments in SN. But I think it shouldn't be fixed as it makes one who starts digging into openaps by your example to execute all reports manually to see what is the problem
and yes, it will not work because of API_SECRET needs hash and not plane password :)
Scott Leibrand
@scottleibrand
Jan 02 2016 23:56 UTC
@ktomy so the last-ns-treatment query fails if you've never added a treatment to NS? I guess that makes sense.
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 23:56 UTC
yes, it returns nothing, then ns-upload doesn't behave correctly
Scott Leibrand
@scottleibrand
Jan 02 2016 23:57 UTC
Everyone who's tried it before has treatments already, such as from care portal
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 23:57 UTC
except me :)
Scott Leibrand
@scottleibrand
Jan 02 2016 23:57 UTC
:-)
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 23:58 UTC
and in log you could see something like Uploaded; most recent treatment event @ PREFLIGHT OK
Scott Leibrand
@scottleibrand
Jan 02 2016 23:58 UTC
probably should fix it though, and/or add a note to create a care portal entry before trying uploader.
Alin Artiom Kenibasov
@ktomy
Jan 02 2016 23:59 UTC
it's up to you, I would make a not about running all aliases by hand at first