@davidkeddydb The docs have some detailed instructions on pulling from NS. Have you used the autoconfigure-device-crud?
@davidkeddydb I don't think the instructions are clear on this point--there's a big part missing from what I could see.
@mccgm yes. I did the autoconfigure-devce-crud. I am looping without issue. I apparently goofed something up with it pulling direct from NS vs the cgm being plugged in.
@garykidd I would love you if you give me the missing part :smile: This is like candy. You get going on looping then you want to move to the edison so you have it 24/7. This all would have gone way faster for me if I had known any linux before starting.
@davidkeddydb are you using @bewest aliases? BTW check your private messages too.
I am using the aliases. I pretty much followed the direction for the ns setup I likely just missed something
You can have NS get entries feed directly into monitor/glucose.json
dont need any processing from raw to monitor/glucose.json
either
I am wondering if I just dump the raw cgm out of do everything it will work.
are you using g5?
no g4
OK. well I cant speak to that for the moment but I can help you get it working form NS
in your instance directory enter
openaps report show
ok. you want the list?
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
cgm://JSON/oref0_glucose/raw-cgm/glucose-raw.json
tz://JSON/rezone/monitor/glucose.json
ns://JSON/shell/nightscout/recent-treatments.json
ns://JSON/shell/nightscout/uploaded.json
pump://JSON/model/settings/model.json
ns://JSON/shell/nightscout/recent-missing-entries.json
ns://JSON/shell/nightscout/uploaded-entries.json
get-profile://text/shell/oref0-monitor/profile.json
calculate-iob://text/shell/oref0-monitor/iob.json
units://JSON/bg_targets/settings/bg_targets.json
pump://JSON/read_carb_ratios/settings/carb-ratios.json
pump://JSON/read_insulin_sensitivities/settings/insulin_sensitivities.json
ns://JSON/shell/nightscout/preflight.json
get-profile://text/shell/settings/profile.json
ns-status://JSON/shell/nightscout/openaps-status.json
ns://JSON/shell/nightscout/uploaded-recent-devicestatus.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
cgm://JSON/oref0_glucose/raw-cgm/glucose-raw.json
tz://JSON/rezone/monitor/glucose.json
ns://JSON/shell/nightscout/recent-treatments.json
ns://JSON/shell/nightscout/uploaded.json
pump://JSON/model/settings/model.json
ns://JSON/shell/nightscout/recent-missing-entries.json
ns://JSON/shell/nightscout/uploaded-entries.json
get-profile://text/shell/oref0-monitor/profile.json
calculate-iob://text/shell/oref0-monitor/iob.json
units://JSON/bg_targets/settings/bg_targets.json
pump://JSON/read_carb_ratios/settings/carb-ratios.json
pump://JSON/read_insulin_sensitivities/settings/insulin_sensitivities.json
ns://JSON/shell/nightscout/preflight.json
get-profile://text/shell/settings/profile.json
ns-status://JSON/shell/nightscout/openaps-status.json
ns://JSON/shell/nightscout/uploaded-recent-devicestatus.json
just.....
openaps report show monitor/glucose.json --cli
openaps use tz rezone --date dateString --date display_time raw-cgm/glucose-raw.json
so that's what essentially is feeding the report.
I would just overwrite the monitor/glucose.json report
what should I have it pull from? or should I just have it do something completely arbitrary
try
openaps report add monitor/glucose.json JSON ns shell get entries.json count=24
ok I just did that and now I invoked it
cgm://JSON/oref0_glucose/raw-cgm/glucose-raw.json
raw-cgm/glucose-raw.json raised 'NoneType' object has no attribute 'iter_records'
raw-cgm/glucose-raw.json raised 'NoneType' object has no attribute 'iter_records'
I think you may have to change another alias that is claled in `openaps do-everyting' and then it should work
actually--you just hit the nail on the head above
ok, so I basically need to go and change all of these as they come up to that same and it should work
I think so.
I'd just edit the
do-everything alias to remove that raw-cgm report
what is the easiest way to do that?
so if you ....
cat openaps.ini
you can see your aliases and report and the like
yes
I think what you need to do is look for your...I think...monitor-cgm ?alias
see that there?
that i don't see
wait try
openaps monitor-cgm
or you could
openaps alias show
yeah, sorry I found it. [alias]
monitor-cgm = report invoke raw-cgm/glucose-raw.json monitor/glucose.json
last_four_hours = report invoke last_four_pump_hours.json last_four_cgm_hours.json
monitor-cgm = report invoke raw-cgm/glucose-raw.json monitor/glucose.json
last_four_hours = report invoke last_four_pump_hours.json last_four_cgm_hours.json
right so you can always rever back by reinstalling the @bewest aliases...but for now jus tto test and see if we can get it running....
just
openaps alias remove monitor-cgm
ok, done
now I get this in the openaps do-everything: PREFLIGHT OK
[Errno 2] No such file or directory
openaps: 'monitor-cgm' is not an openaps command. See 'openaps --help'.
[Errno 2] No such file or directory
openaps: 'monitor-cgm' is not an openaps command. See 'openaps --help'.
I am also getting a time issue with clock settings determine-basal://text/shell/oref0-predict/oref0.json
BG data is too old, or clock set incorrectly. Your CGM time is Sat Sep 24 2016 20:39:23 GMT-0400 (EDT) but your system time is Sat Sep 24 2016 21:01:42 GMT-0400 (EDT)
BG data is too old, or clock set incorrectly. Your CGM time is Sat Sep 24 2016 20:39:23 GMT-0400 (EDT) but your system time is Sat Sep 24 2016 21:01:42 GMT-0400 (EDT)
right...now you just need to add back the alias but instead of having it be fed with raw-cgm and the like have it come from monitor/glucose.json
openaps alias add monitor-cgm "report invoke monitor/glucose.json"
tdid that work?
yeah, I just did that. let me retry it
I think you still need to scrub someting else out
Nope. That did it. AWESOME!!! Thanks for the help.
Boo Yaaa!
let me check that the cron is going ok without it being attached
I dont think it's be that hard to add the raw back----had you been actively using that?
it was running but I gather it was just chucking nothing in it so it was sort of pointless anyway
You know what you really want to do is add the ability for BGs to come from cgm or NS if cgm isn't accessible...
that way you can loop without internet access for your rig
I gather that I likely have it set now to just run off NS
I'm gonna step away but will check back in case there may be something else that you need...that i can also help you with.
I think you're running exclusively from NS now, yes.
it's pretty easy to change back though
Thanks @garykidd I will work on seeing if I can get that figured to add that in. Thanks
I haven't added both NS and cgm myself---but if you use @scottleibrand latest script...I understand it has both available. so easy. i
anytime bro. check one-on-ones too.
Good work @garykidd in troubleshooting this! @davidkeddydb Gary is right that there is code in the new setup from Scott's latest script. But if you just want to add it in a simpler way with your existing setup, here's how I do it:
pi@raspberrypi:~/openaps_sor $ openaps alias show monitor-cgm
monitor-cgm ! bash -c "if [ -e /dev/ttyACM0 ]; then openaps monitor-cgm-usb && cp monitor/glucose{-usb,}.json; else openaps monitor-cgm-ns && cp monitor/glucose{-ns,}.json; fi"
So to translate that: I have setup my
monitor-cgm alias to do an if-then.
If the file
/dev/ttyACM0 exists, that means the Dexcom receiver is plugged in and recognized.
On your system, it might be different, but you should be able to figure out what it is by doing
tail -f /var/log/messages and then watching the output as you plug in the receiver. (Type ctrl-c to get out of the mode of seeing the log messages.)
If the receiver is plugged, then it runs
monitor-cgm-usb, then copies the report glucose-usb.json over to the name glucose.json.
If not, then it runs
monitor-cgm-ns, to produce glucose-ns.json, then copies that file to the name glucose.json.
Either way, it should end up with a
monitor/glucose.json report that the other tasks can use.
The
glucose-usb is just a report produced from openaps use cgm iter_glucose 5 and the glucose-ns is a report from running openaps use ns shell get entries.json count=5 (similar to what Gary suggested). Of course you can change number of entries if you want.
I was glad today that I had previously set this up, since my Dexcom receiver just died. Not the end of the world, though, since it's G5 and I can still use my phone to get values and send to NS via Share, and I can pull from NS to loop. :smile:
The Dexcom tech support guy was looking at my record and said "Wow, you've been with us a long time - I've never seen anyone with a 4-digit customer number, and I've never seen some of these products you had in the past." I had one of the first-gen Dexcom units, so it has been about 10 years now I guess.
:older_man:
@elnjensen Thanks as always. That is crystal clear. You and me both with the number with Dexcom.
Quick question: what's the command to generate the device "units"? And does it have a distinct vendor, or does it use 'openaps.vendors.process'?
You could try
openaps device add units. Not sure if it needs more than that.
yeah, it needs more than that...not sure what though...I've searched previous posts and couldn't find anything
alternately, if someone could post their entry from their openaps.ini file (and the contents of their units.ini), I'd just enter it manually
Docs here say
openaps device add units units.
weird, no idea how I missed that...thanks!
In openaps.ini I just have
[device "units"]
vendor = openaps.vendors.units
extra = units.ini
and units.ini just has
[device "units"]
Not clear to me exactly what this is used for...
I wonder if, because I'm using mg/dl, this is actually a no-op (i.e. doesn't do anything) but would have to be configured differently if it were mmol/l.
I believe its a device that is used to convert raw pump data, for things like time zone correction, etc...
without it, I can pull raw data but cannot use the raw data to invoke the corresponding unit-converted report
I know auto sensitivity has a nice set up outlined on openaps docs. Is there something similar for Meal Assist?
@elnjensen thanks for your compliemnt. If you are around tomorrow at all--was hoping you could address some questions I have on your dual CGM approach described above. if not, anoither time maybe. thx
@Missdazzle no, you don't have to do new setup script to get temp targets, you can enable them manually/separateky
@davidkeddydb the meal assist stuff is documented in the wip/ama-docs branch of openaps/docs
@scottleibrand Thanks and I had already read it. The page is primarily dedicated to the set up of the automatic sensitivity with little on meal assist. I know that I got directed from my search to your set up here: https://gist.github.com/scottleibrand/63a8541397378bf9f4b8 which I think I can follow
there is also this from before we wrote it, explaining how we planned for it to work: openaps/oref0#68
@scottleibrand Thanks. I am excited to be looping. I will get meal assist figured here in the next few and get it working.
There probably aren’t a lot of people with a temp basal max higher than 6.375, but if there are, this is a pretty serious issue for them: ps2/rileylink_ios#254 decocare will need a patch as well.
I issued a safety warning for Loop users. Can someone verify that this is an issue for OpenAPS users as well (you can do a quick temp basal > 6.375, cancel it, and see how it decodes). I believe the line that needs to change in decocare is this: https://github.com/bewest/decoding-carelink/blob/master/decocare/history.py#L189
It is potentially a bigger impact to OpenAPS users, because history decoding is always used for calculating IOB.
Here is how I fixed it for RL/Loop: https://github.com/ps2/rileylink_ios/commit/9b0f0ac3f743ceef5c15c665b1fbf4da07a5288d#diff-c6ddd181aee3b0d96dbf26a136e8076bL40
I don't see an open issue, but I do recall this being a problem a limitation at some point. IIRC it would fail safely, though: it would bail out completely rather than decode the basal incorrectly. Will need to test to verify that, and get your fix in regardless.
Also, with expectedDelta, oref0 behaves much more safely than Loop when IOB is incorrect. We should probably add that logic to Loop at some point as a safety enhancement, and to prevent things like having to turn off Loop because it's temping to zero after a site change and a big bolus to replace missed insulin.
We've had cases when IOB calculation was totally broken and oref0 thought IOB was always zero, and it still managed to keep BG in range (just a bit more variable than otherwise) solely based on BG info.
"Mystery effect" (retrospective correction or whatever it's actually called) might do something similar there for loop now...
Although that probably doesn't have the logic of reducing temps proportionally to how much the delta is already moving in the right direction. It seems a lot more along the lines of the deviation stuff in oref0
I don't think it will bail out. It will decode a 6.4 basal rate as a 0.0 rate, for example.
K, that might have been a different overflow then. I'll test with our test pump in the morning.
Does this affect read_temp_basal and set_temp_basal too?
I think that's what I'm remembering: an inability to set really high temps.
Not as familiar with that code, but yes, it looks like setting temp basal might silently truncate data as well.
Hi, new to openaps and trying to setting up my first system on a Raspberry Pi 3, I have come to the stage of uploading data into nightscout but cannot get it to work... when running
nightscout autoconfigure-device-crud <https://my.nightscout.host> <averylongplainsecret>
I only end up with the following result
bash: nightscout: command not found
This message was deleted
Anybody have any recommendations on usb battery power packs with charge thru capability...otherwise just going to pull trigger on https://www.amazon.com/Zendure-2nd-Portable-Charger-10000mAh/dp/B01AXZUMOK/ref=sr_1_1?th=1
@garykidd the zendure is awesome I love it. It came with a soft bag for the battery and a flat cable . Tecknet also have charge trough battery , the correct model could be found on OpenAPS.org but don't remember where exactly .
@scottleibrand @jasoncalabrese Hi guys, as you know I'm using xDrip for BG upload, so far so good, but one Q what needs to be done to upload BGs from my Dexcom receiver from my openAPS :question: In the next time I will use the Dexcom receiver for offline looping so that it would be great to have the Dexcom BGs available in NS, please advice - thanks you!
@ceben80 check out the oref0 repo wiki
That method is somewhat overkill, but does include upload of Dexcom data to NS.
Oki @scottleibrand :+1:
I think you meant the docs wiki? My notes for openxshareble (but will work with usb also) https://github.com/openaps/docs/wiki/CGM-Loop-with-openxshareble-scratchpad
I haven't moved to the refresh loop script yet, but after I get back from NY it's on my list
Yes, that, sorry.
@scottleibrand sorry asking again didn't find the method, would you mind me pointing to that? Thanks a lot :+1:
sorry for asking again, for edison bluetooth tethering it is still the bleno approach from Monica or does jubilinux provide easier solutions? My offline solution using local NS and local Mongodb worked fine for a week, now I want to switch to BT rather then Hotspot connection
Ok...the link from Jason...will look at this...thx
Hey everyone. Hope the OpenAPS community is well!
Welcome @MHoskins2179_twitter
I am sitting in on a #OpenAPS Setup session with my mom and her non-D coding tech guru... They are "hitting heads on wall" trying to setup. To find correct info files for specific D-Tech being used...
@ceben80 if you don't want to do the jq merge or the openxshareble stuff, the steps in that wiki can also be adapted just to upload the info from a USB-connected CGM.
@MHoskins2179_twitter which d-tech?
MedT 722
Dexcom G5 with iPhone
What is the issue? What step in the docs are they at? (and what steps have they completed so far?)
(It's also a good idea for them to jump in here, too, so they can share info and Qs and be familiar with Gitter for future stages, too.)
And so whomever is doing the typing can copy-paste Qs and As between their terminal and gitter.
Currently, they're asking how to connect Wifi robot... A mobile hotspot?
Did you find the part in the docs that shows how to add your wifi network to the Pi config?
That works the same whether you use a home wifi network or a mobile hotspot.
Is there an illustration or graphic anywhere, as to how exactly OpenAPS works? (More for my own mental procession than anything...)
So, the OpenAPS docs readme is a good starting place - scroll down and read the phase 0-6 descriptions for a high level. https://github.com/openaps/docs/blob/master/README.md
And then the image at bottom of this page in the docs may also help you get big picture of how it flows, albeit has a lot of details:
But generally "how it works", I always recommend reading the Reference Design: https://openaps.org/reference-design/
@scottleibrand Found a small issue with the commit from a few days ago about requiring clock-zoned.json. It is now causing warnings in running the meal.json report. Comment and possible fix are openaps/oref0@3f4c205. Just verified in my local setup that just adding a ? after the colon in the regex fixes this, but could be other implications of that I suppose.
If I wanted to change my DIA in something other than full-hour increments (e.g. from 3 to 3.5 hours), is there a way in OpenAPS preferences to do that? The pump only does full hours.
right now there isn't an override in preferences. you could manually change the contents of the report that pull that from the pump I suppose...
An issue: It's not easy to find the specific documents with coding for devixes (G5, 723, Edison board), and then to know if those codes are good to go or must be modifed.
Scott -- will look thru those internet connectivity docs to see if Qs are answered. Thx.
@MHoskins2179_twitter what do you mean by “know if those codes are good to go or must be modified”? And again, which section of the docs are you in, what stage are you attempting to accomplish?
Scott, thx re DIA info. Not a big deal, and would rather not mess around with altering reports if it's not built-in already.
@MHoskins2179_twitter keep in mind that these docs are a step-by-step walkthrough of how to set everything up, not a set of device-specific magic "codes". There are a few things that are device-specific, but they're clearly labeled. otherwise, the idea is that you start at the beginning and proceed linearly, and just skip over parts that don't apply to your hardware.
@MHoskins2179_twitter If I'm understanding your question correctly, for the most part the code is device-agnostic, e.g. nothing different for 522 vs. 523 vs 722 vs 723 pumps. Sometimes you pass flags when setting up devices (e.g. --G5 with the CGM).
At the very beginning pretty much. Trying to pinpoint the specific docs to follow for hardware we have. Not easy to locate them in the mass of documents/links on main site.
generally the answer is "follow the walkthrough". as @elnjensen mentioned, we don't have different docs for different hardware.
@elnjensen if we wanted to improve how we deal with DIA, we might want to explore using something better than my simple bi-linear activity curve (0 to peak at 75m, and back to 0 at DIA). I know Nate's mmhistorytools and Loop use the Walsh curves: you could explore that...
@scottleibrand So I've been testing the oref0-setup.sh script for mdt. Before I forget, let me just say how AWESOME this initiative is! I'm almost there, I think. I'm in the process of checking all aliases and reports. Most seem to work and I can upload all pills. However, no BG is uploaded. I also have one doubt about the
monitor/meal.json report. I couldn't find anything about it in the docs. Is the carbs = monitor/glucose.json input correct?[report "monitor/meal.json"]
profile = settings/profile.json
carbs = monitor/glucose.json
clock = monitor/clock-zoned.json
reporter = text
json_default = True
use = shell
pumphistory = monitor/pumphistory-zoned.json
basal = monitor/carbhistory.json
device = meal
remainder =
glucose = settings/basal_profile.json
the meal stuff is part of AMA, which is documented in wip/ama-docs: https://github.com/openaps/docs/blob/wip/ama-docs/docs/docs/walkthrough/phase-6/Configure-Automatic-Sensitivity-Mode-and-Advanced-Meal-Assist.md
ok, thanks!
and we haven't added support for uploading BG data to NS yet, so that's expected
that's one of the most-requested features right now, though, so if you want to help us get that working, it should be pretty straightforward
ok, I'll give it a try
But I do think the ordering there is wrong - before when I was checking some reports manually, I got a warning about ordering of arguments. Let me see if I can dig it up.
yes, that also needs fixed
Here it is
pi@raspberrypi:~/myopenaps_advanced $ openaps ns-meal-carbs
Warning: Argument order has changed: please update your oref0-meal device and meal.json report to place carbhistory.json after basalprofile.json
yep
But the good news is that it seems to recognize the inputs (yes?) and realize that they are in the wrong order, and deal with that.
Do you think that's right, @scottleibrand ?
yeah, I changed that in the code (and made it backwards compatible)
Great, thx.
just need to update the json to match
I think https://github.com/openaps/oref0/pull/219/commits/6f57758484b31442ba259091c500ff7c757ca8f6 will do it: care to test?
Can you remind me of the git magic incantation to pull down the latest?
cd ~/src/oref0 && git checkout oref0-setup && git pull
that's three very simple commands connected by &&
and the middle one is only necessary if you're on a different branch: if
git branch shows you're already on oref0-setup, then a simple git pull will do
Thank you!
Guess I need to set up a new installation to test, yes? Since it's in the device itself in
openaps.ini.
I think just re-running it on the same directory will re-create the devices
Hmm, having trouble testing at the moment, can't get other loop processes to die...
comment them out in cron and
killall -g openaps
Did that, but still had some things hanging around. Interesting watching it in
top - there would be nothing with openaps, then more new processes would keep popping up, even though commented out in cron. Don't really understand it.
Finally got rid of them all, now testing with a new installation.
you might have killed off the parent process some other way before doing
killall -g openaps?
if so, then it wouldn't find anything named openaps to kill, even though the child processes were still running
you can always just comment out the crontab entry and reboot
Could well be. Anyway, testing now...
The other funny thing was that after I Ctrl-C'ed out of a "listening for pump comms.." prompt it kept popping up those periods on the command line into other commands I was typing... Zombies!
heh yeah, I see that sometimes
OK, still getting that warning. Let me diff the openaps.ini files.
Looks like no changes to the file.
ok, try
openaps device remove meal and re-run?