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

27th
Apr 2018
Tore Bjørndalen
@BjoTore_twitter
Apr 27 2018 00:00
Skjermbilde 2018-04-27 01.38.28.jpg
Eric
@ecc1
Apr 27 2018 00:00
@BjoTore_twitter I assume that huge bolus wasn't actually delivered / is not listed in the pump's Bolus > Bolus History menu?
Tore Bjørndalen
@BjoTore_twitter
Apr 27 2018 00:01
No, reservoir didn't change
Somehow they get into NS?
Plenty of the as you can see
Eric
@ecc1
Apr 27 2018 00:02
Do you still have the pump history JSON file from around one of those times?
Tore Bjørndalen
@BjoTore_twitter
Apr 27 2018 00:03
I can connect and fetch it I guess. Which one? the 24h zoned from monitor?
katie disimone
@Kdisimone
Apr 27 2018 00:42
@BjoTore_twitter did you have the rig automatically changing the pump time? Can you find logs from around the time of the large bolus records?
Jon Cluck
@cluckj
Apr 27 2018 00:44
@Kdisimone I think we fixed it, the explanation is pretty wild :) the pump creates history entry at midnight every day that contains the total daily dose & timestamps it for 24h prior
it was making a large number of duplicates with the way pumphistory is handled in 0.7.0-dev
katie disimone
@Kdisimone
Apr 27 2018 00:46
:+1:
Tore Bjørndalen
@BjoTore_twitter
Apr 27 2018 10:48
    "timestamp": "2018-04-27T00:45:22+02:00",
    "_type": "Bolus",
    "amount": 0.1,
    "programmed": 0.1,
    "unabsorbed": 4.8,
    "duration": 0
  },
  {
    "timestamp": "2018-04-26T00:00:00+02:00",
    "_type": "DailyTotal523"
  },
  {
    "timestamp": "2018-04-26T00:00:00+02:00",
    "_type": "DailyTotal"
  },
  {
    "timestamp": "2018-04-26T23:51:59+02:00",
    "_type": "Bolus",
    "amount": 0.6,
    "programmed": 0.6,
    "unabsorbed": 6.025,
    "duration": 0
  },
So these 2 "DailyTotal" entries are creating the havoc since they have timestamp 00:00?
I've seen related issues when timestamp at midnight was recorded as "24:00:00" back in the "year 2000" days :)
Tore Bjørndalen
@BjoTore_twitter
Apr 27 2018 10:53
Never underestimate date/time calculations. Never expect, inspect! :)
@ecc1 Do you need more of the pumphistory?
Jon Cluck
@cluckj
Apr 27 2018 11:23
yep, they have a timestamp which is much earlier than the records around them
so the pumphistory since request was getting 24 hours of history instead of just a few minutes
Tore Bjørndalen
@BjoTore_twitter
Apr 27 2018 11:24
Not much info?
Jon Cluck
@cluckj
Apr 27 2018 11:25
it was making lots of duplicate entries, and also messing with calculate-iob
Tore Bjørndalen
@BjoTore_twitter
Apr 27 2018 11:25
(In the dailyTotal records)
Purpose?
Jon Cluck
@cluckj
Apr 27 2018 11:25
for total daily dose
the data was filtered out, but the timestamp wasn't :laughing:
Tore Bjørndalen
@BjoTore_twitter
Apr 27 2018 11:27
Ok, I see. Need to try understand what openaps.jq is actually doing.
Jon Cluck
@cluckj
Apr 27 2018 11:30
oh, it translates the raw history pages into something that openaps can parse
the bugfix filters out any & all records that aren't explicitly defined in openaps.jq
Tore Bjørndalen
@BjoTore_twitter
Apr 27 2018 11:36
:+1:
Andrew Kirkby
@kirkbysound
Apr 27 2018 13:56
Screen Shot 2018-04-27 at 14.52.38.png
I have a question regarding post meal entry predictions... Is it normal for the prediction to show a steep drop like that for the first few readings?
Or are my ratios off somewhere?
Carlb ratios seem to be ok here during the day based on the last few days
Dana Lewis
@danamlewis
Apr 27 2018 14:02
WDAS (what does Autotune say?)
But also yes - if you do a full bolus for a big meal, that often has predictions like that
Because that insulin is going to hit faster than food can be digested
Michael Spradling
@M1Sports20
Apr 27 2018 14:04
@kirkbysound: I have the same predictions
Eric
@ecc1
Apr 27 2018 14:06
@BjoTore_twitter just to be pedantic, openaps.jq is actually translating from one JSON form (corresponding almost directly to the internal data structures the Go program creates when it parses history records) to the JSON form that openaps expects. You can see those other forms with the "-f format" options to the pumphistory command if you're curious (for example, to see what other info is in the DailyTotal records)
Actually, that's wrong too: pumphistory always produces the "Go version" of JSON. I was confusing it with the "mdt" command that has multiple format options
Tore Bjørndalen
@BjoTore_twitter
Apr 27 2018 14:10
@ecc1 Thank you Eric I had in mind looking for some of the statistics that the pump has and retrieve that to make some history presentation for it. Don't think any of that exists?
Andrew Kirkby
@kirkbysound
Apr 27 2018 14:13
@danamlewis that makes sense.. just curious as it normally levels back out but I wanted to make sure I’m not missing anything. Auto tune is a little out regardless as I have multiple ratios. I tend to need 3x the insulin in the morning per gram.
Eric
@ecc1
Apr 27 2018 14:13
@BjoTore_twitter What kind of statistics were you looking for?
Dana Lewis
@danamlewis
Apr 27 2018 14:21
That's probably influencing the predictions then. But if you're happy with your outcomes: :+1:
Jon Cluck
@cluckj
Apr 27 2018 14:41
hehe
Samuchco
@Samuchco
Apr 27 2018 14:54
Rig is working fine. But NS is showing an "Unknown" message on the Pill. Tried to clean mLab which has a size on disk of 490MB and size 343MB.
Shawn
@dyjaks
Apr 27 2018 15:00
Is it possible to reset autotune basal rates? Around 2 weeks ago I my rig started to have to constantly issue SMB just to keep me level but still too high (around 160mg/dl) and by my math my basal should need to be ~20% higher than it is
Diabetic_Dude
@diabetic_dude_twitter
Apr 27 2018 16:56
what CGM software do you use and why? I have tried Spike, and I love it, but it looses connection to the transmitter too often. I have tried xdrip, but my BGs seem to go up and down too much, I am only assume that the issue is the same that I am having with spike. I like the dexcom app but god the alarms! I don't need the people in the next room to know my BG is low and when I have headphones on. WOAH. ears blown...
Samuchco
@Samuchco
Apr 27 2018 17:20
disregard. it's working. Re-start rig probably did it.
Samuchco
@Samuchco
Apr 27 2018 18:44
spoke too soon. Back to rig working properly as seen on logs and main body of NS. However, OpenAPS pill shows "waiting" . Pump pill shows "undefined".
This message was deleted
Screenshot (29).png
Samuchco
@Samuchco
Apr 27 2018 19:03
Apr 27 14:59:32 loop2 pump-loop.log: enact/smb-suggested.json: {"temp":"absolute","bg":56,"tick":-4,"eventualBG":109,"insulinReq":0,"reservoir":"121.6","deliverAt":"2018-04-27T18:59:32.634Z","sensitivityRatio":1.2,"predBGs":{"IOB":[56,52,48,43,39,39,39,39,39,39,39,39,39],"ZT":[56,50,44,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39,39],"COB":[56,52,48,45,43,40,39,39,39,39,39,39,39,39,39,39,39,40,43,45,48,51,55,58,62,66,69,73,76,79,82,85,88,91,93,95,97,99,101,103,104,105,106,107,108,108,108,109],"UAM":[56,50,44,39,39,39,39,39,39,39,39,39,39]},"COB":24,"IOB":1.151,"reason":"COB: 24, Dev: 12, BGI: -6, ISF: 83, CR: 13, Target: 93, minPredBG 28, minGuardBG -7, IOBpredBG 39, COBpredBG 109, UAMpredBG 39; 5 add'l carbs req w/in 0m; minGuardBG -7<67","carbsReq":5,"duration":120,"rate":0}
Apr 27 14:59:35 loop2 pump-loop.log: Temp refreshed: monitor/temp_basal.json: {"duration":106,"rate":0,"temp":"absolute"}
Apr 27 14:59:38 loop2 pump-loop.log: enact/smb-enacted.json: "Rate: 0 Duration: 120"
Apr 27 14:59:41 loop2 pump-loop.log: Temp refreshed: monitor/temp_basal.json: {"duration":120,"rate":0,"temp":"absolute"}
Apr 27 14:59:41 loop2 pump-loop.log: No bolus needed. Settings less than 15 minutes old. Edison on battery: 59%. Pumphistory-24 < 30m old
Apr 27 15:00:21 loop2 pump-loop.log: Pump profile refreshed; Settings refreshed; Refreshed pumphistory and meal.json
Apr 27 15:00:27 loop2 pump-loop.log: IOB: 1.151
Apr 27 15:00:27 loop2 pump-loop.log: Your instance of oref0 [0.6.1-dev, dev] is up-to-date.
Apr 27 15:00:27 loop2 pump-loop.log: Completed oref0-pump-loop at Fri Apr 27 15:00:27 EDT 2018
Apr 27 15:01:02 loop2 pump-loop.log: Starting oref0-pump-loop at Fri Apr 27 15:01:02 EDT 2018 with 30 second wait_for_silence:

pump-loop.log