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

24th
Feb 2016
tazitoo
@tazitoo
Feb 24 2016 00:11
changed max_bg=min_bg=120 ...current BG level at 103
Thanks for your suggestions.
Jason Curry
@mccgm
Feb 24 2016 01:24
Can anyone tell me what the "Offset" attribute is for in bg_targets.json?
What is it for and what does it control?
Dragan Maksimovic
@dm61
Feb 24 2016 03:02
I am not 100% sure, but I think Offset is equal to when a particular setting for bg targets starts, expressed in minutes from midnight=0. If you have just one setting, Offset=0. If you add another setting at some other time, you may check Offset.
Clayton McCook
@cdmccook
Feb 24 2016 03:19
Hey all I'm stuck trying to pull the data from my daughter's Nightscout site. I saw a thread the other day where Ben was trying to help someone but I can't remember what to put in the command line to add it as a device and then how to access it. Any help appreciated. Thanks.
Rachel Sandlain
@audiefile
Feb 24 2016 03:35
Agree with @dm re offset
Rachel Sandlain
@audiefile
Feb 24 2016 03:58
@cdmccook try something like openaps device add ns-glucose process bash -c "curl -s $NIGHTSCOUT_HOST/api/v1/entries/sgv.json | json -e \"this.glucose = this.sgv\""
Requires you define NIGHTSCOUT_HOST wherever you run that command. So in cron and in your bash shell to run it manually
export NIGHTSCOUT_HOST=https...
Matthias
@ceben80
Feb 24 2016 04:49

@/all
I have configured my smartphone hotspot und home WiFi.

Both network connections are working fine, but if I am gonna leave my home I'd be happy if the pi would be able to dynamically connect to the hotspot.
Currently I'm using only the hotspot, but the mobil battery consumption is enormous.

Any idea or how do you do this?

Scott Leibrand
@scottleibrand
Feb 24 2016 04:57
With a Pi and a Bluetooth dongle you can do
Bluetooth tethering.
Unfortunately I haven't figured out how to make it work on the Edison.
Unlike wifi, Bluetooth tethering can be initiated by the Pi.
Matthias
@ceben80
Feb 24 2016 05:27
Sounds great, will try this, thx
eliotsdad
@eliotsdad
Feb 24 2016 06:00
So. . . am slowly wading back in to this. . have the Pi up and running, communicating with devices, able to read data from the pump and send commands. It's been a few months since we were on site with Scott and Dana. . . what's next?
Jim Matheson
@jmatheson
Feb 24 2016 07:57
great openaps meetup tonight in SF.
IMG_4469.jpg
Ben West
@bewest
Feb 24 2016 08:43
really great meetup, @jmatheson many thanks, and so great to meet people
was actually one of the biggest hacking sessions I've been to, if not the biggest
Jim Matheson
@jmatheson
Feb 24 2016 08:45
:thumbsup:
Ben West
@bewest
Feb 24 2016 09:09
I suspect only TI and suppliers have only ever seen so many cc1111 sticks in one place
Pete Schwamb
@ps2
Feb 24 2016 10:53
Jealous!
JaysonEwer
@JaysonEwer
Feb 24 2016 13:52
đź‘Ť +1 @ps2 @jmatheson
Scott Leibrand
@scottleibrand
Feb 24 2016 13:58
@eliotsdad the openaps/docs on Github have improved markedly since then. I would check those out again and pick up where you left off. Pop in here with any questions at all.
Colin Lennon
@colinlennon
Feb 24 2016 14:57
Hi all, newbie here from the UK. Could anyone tell me if you can have 2 xdrips both picking up signals from a single G4 transmitter simultaeneously? What I'm thinking of doing is building a similar rig to @ochenmiller (an rPi zero, wixel, HM-10 and carelink stick) which would all work offline, but at the same time, have a 2nd, separate xdrip paired to my phone so I can send readings to NS. Not sure if that's feasible?
Scott Leibrand
@scottleibrand
Feb 24 2016 15:00
No reason it shouldn't work. G4 doesn't pair, it just transmits blindly and receivers pick it up passively. We use multiple receivers all the time.
Peter Miller
@ochenmiller
Feb 24 2016 15:01
@colinlennon - I have several rigs picking up G4 signals. All updating nightscout. Works like a charm!
Scott Leibrand
@scottleibrand
Feb 24 2016 15:02
Careful with multiple differently-calibrated rigs all updating NS if you use that for looping: deltas will be all over the place. Better to use each rig's local receiver data only if you do that.
Peter Miller
@ochenmiller
Feb 24 2016 15:03
We also upload from the Dexcom receiver in parallel. More data points. Redundancy.
Scott Leibrand
@scottleibrand
Feb 24 2016 15:03
Even if they're calibrated the same, you'll get false +0 deltas.
oref0 assumes one data point per 5 minutes. Anything other than that will be suboptimal.
Someone was working on fixing it to use actual timestamps and relax that assumption, but not sure what ever happened to that effort.
Colin Lennon
@colinlennon
Feb 24 2016 15:04
Excellent, thanks! I don't think I would have that issue @scottleibrand as I'd only have one standalone xdrip transmitting to NS. The other wixel\HM-10 combo would be wired into the RPi zero purely to get G4 data into OpenAPS and allow the loop to work offline.
Scott Leibrand
@scottleibrand
Feb 24 2016 15:05
Perfect.
Peter Miller
@ochenmiller
Feb 24 2016 15:05
Our rigs check with Nighscout for a calibration every time they get a record from the g4. If there's no connection, then they use cached info
Scott Leibrand
@scottleibrand
Feb 24 2016 15:05
Are you using NS BG for oref0?
Colin Lennon
@colinlennon
Feb 24 2016 15:06
Off-topic but I'm amazed by the progress you have all made so far! I've got my xdrip and NS setup up and running now, very excited about making a start with OpenAPS.
Scott Leibrand
@scottleibrand
Feb 24 2016 15:06
:-)
Peter Miller
@ochenmiller
Feb 24 2016 15:15
Also, each rig only uses locally derived BG info, so no issues with the +0 offset issue that @scottleibrand was mentioning above. Much like @jasoncalabrese we're running three rigs. One pocketsized. On in the bedroom. One in the living room. We rarely have connectivity gaps of more than 30 minutes.
Scott Leibrand
@scottleibrand
Feb 24 2016 15:26
Nice. Sounds similar to our setup, except that we only upload from one receiver to NS.
Peter Miller
@ochenmiller
Feb 24 2016 15:30
BTW, we avoid +0 deltas in nightscout by having each rig/source "upsert" into nightscout. That way there's only ever one NS record for a given BG.
That also allows us to re-run uploading for rigs that are off of the network for a few hours. Rerunning just fills in the gaps.
We're also shoe-horning some additional info into the BG records: iob, separate device fields for each source, what calibration was used.
Scott Leibrand
@scottleibrand
Feb 24 2016 15:38
Nice. Might be useful to write up how to do all that, particularly the downloading calibrations bit that gets them to stay in sync so upsert works.
Peter Miller
@ochenmiller
Feb 24 2016 15:39
The multiple sources allow us to easily derive calibration records from the dexcom receiver. I have a script that loads all the records for the last x hours, then does a linear regression on the dexcom receiver vs the raw.
Scott Leibrand
@scottleibrand
Feb 24 2016 15:40
Nice. Is that open source yet?
Is your pocket sized rig using Edison with xDrip?
Peter Miller
@ochenmiller
Feb 24 2016 15:40
I have to confess that I keep meaning to do that, then I look at my code which is a nightmare of spaghetti. Wish I did this for a living.
Scott Leibrand
@scottleibrand
Feb 24 2016 15:40
you should see mine. :-p
Peter Miller
@ochenmiller
Feb 24 2016 15:41
But we all have! Looks pretty good to me!
Scott Leibrand
@scottleibrand
Feb 24 2016 15:41
You didn't see DIYPS, though. PHP, with no functions and all variables global.
Peter Miller
@ochenmiller
Feb 24 2016 15:41
Pocket sized is RPi /Wixel/Carelink/HM-11.
Scott Leibrand
@scottleibrand
Feb 24 2016 15:42
Oref0 has benefited from a lot of refactoring by people who know what they're doing.
Peter Miller
@ochenmiller
Feb 24 2016 15:42
Who needs local variables?
Function schmunction
Scott Leibrand
@scottleibrand
Feb 24 2016 15:42
I had a rather interesting test failure because I was changing a global variable.
;-)
to the point where I named that branch wtf-indeed until I figured it out. :-p
Peter Miller
@ochenmiller
Feb 24 2016 15:43
Then there's passwords and keys baked right into the code...
Scott Leibrand
@scottleibrand
Feb 24 2016 15:43
heh ya.
Peter Miller
@ochenmiller
Feb 24 2016 15:43
lol
Anyhow, time to rehabilitate myself and post what I've done!
Cuz that's how we all move this thing ahead
Scott Leibrand
@scottleibrand
Feb 24 2016 15:44
:-) yep
Peter Miller
@ochenmiller
Feb 24 2016 15:45
You can publicly shame me if I don't post within a week
Scott Leibrand
@scottleibrand
Feb 24 2016 15:46
I will forget unless you mention it again. :-)
Peter Miller
@ochenmiller
Feb 24 2016 15:47
:smile:
Colin Lennon
@colinlennon
Feb 24 2016 15:51
@ochenmiller is it a Pi zero you're using in your pocket size rig?
eyim
@eyim
Feb 24 2016 15:56
Anyone have more details on Marius' set up on a single chip?
https://twitter.com/PancreasInGame/status/702345463593238528
Jim Matheson
@jmatheson
Feb 24 2016 16:27
This message was deleted
Scott Leibrand
@scottleibrand
Feb 24 2016 16:31
@eyim are you in the OmniAPS slack channel I referenced in my reply to that tweet?
Jim Matheson
@jmatheson
Feb 24 2016 16:50
@danamlewis @scottleibrand it would be nice to see traffic trends for openaps.org. I don’t see any GA tracking tags on the site. Do you have something set up with wordpress? Any data you can share?
Jim Matheson
@jmatheson
Feb 24 2016 16:58
and are you able to see the traffic tab (under graphs) for docs on github or is that only for private repos?
eyim
@eyim
Feb 24 2016 16:59
@scottleibrand Missed that about the Slack channel. Just joined now
Jim Matheson
@jmatheson
Feb 24 2016 17:04
nice
thanks for sharing
Dana Lewis
@danamlewis
Feb 24 2016 17:06
:+1:
Jim Matheson
@jmatheson
Feb 24 2016 17:16

it’d be great to start tracking the funnel of interest in openaps and time lag between steps. Something like:

  • Visitors to the project
  • Interest in trying to loop
  • Have a pump that works (this will be a big drop off i assume for now)
  • Started building
  • Live

Could really help drive where efforts ar focused moving forward.
I am quite experienced in this area. If the group is interested in this type of thing I could help get it set up.

for those that are interested, last night @mariusae mentioned that his ping comms code is up on github @ https://github.com/mariusae/pingrf
Dana Lewis
@danamlewis
Feb 24 2016 17:49
@jmatheson If you have ideas on ways to track that, would love to hear them. I think having ppl pr themselves into a "loop in progress" list will be helpful in that regard.
Oskar Pearson
@oskarpearson
Feb 24 2016 18:54
Yay for a meetup-and-hack-around-with-cc111x-get-togethers :)
@colinlennon I'm in London - our setup is quite different from yours (No DexCom in our case), but if you need a hand with anything I can try help out :)
So everyone, I've a couple questions about getting Nightscout going if you don't mind.
Jim Matheson
@jmatheson
Feb 24 2016 18:55
@oskarpearson the problem is the radio clash had my loop down for a couple hours. once everyone left it started working again.
Oskar Pearson
@oskarpearson
Feb 24 2016 18:56
@jmatheson Ha! Terrible radio protocol the Medtronic ppl chose there!
So .. as usual, I'm "playing life on hard mode" and trying to set up Nightscout without Azure or any of the other things.
We're using openaps with medtronic pump and medtronic's sensors (no dexcom).
At the moment I'm just trying to send data straight from the Edison to the host over wifi. As few complexities as I can
The issue is I can't get any data to reflect on nightscout.
I'm running nightscout like this for testing:
API_SECRET=NONSHA1SECRETINPLAINTEXT DEVICESTATUS_ADVANCED=advanced DISPLAY_UNITS=mmol ENABLE='openspump' MONGO_CONNECTION=mongodb://ns:MONGOPASSWORDHERE@localhost/ns PORT=1337 node server.js
I'm then running ns-upload $NIGHTSCOUT_HOST $API_SECRET devicestatus.json upload-status.json on the pi. It seems like it's doing what's expected there.
(API_SECRET in this case is the SHA1 version of the SHA)
It returns some json - eg: {"device":"openaps://pan-edison-1.hq.deckle.co.uk","pump":{"battery":{"status":"normal","voltage":1.31
Peter Miller
@ochenmiller
Feb 24 2016 19:01
@ochenmiller is it a Pi zero you're using in your pocket size rig?
Oskar Pearson
@oskarpearson
Feb 24 2016 19:01
On the server, if I connect to the mongodb I can go:
Peter Miller
@ochenmiller
Feb 24 2016 19:01
Yes, using Pi zero in pocket rig.
Oskar Pearson
@oskarpearson
Feb 24 2016 19:01
use ns
and something like this: db.devicestatus.find()[12] and I get the json I just pushed to the server
Dana Lewis
@danamlewis
Feb 24 2016 19:02
::puts in a pitch for @oskarpearson to review and PR anything he learns today to the NS visualization segment of the docs:: ;)
Oskar Pearson
@oskarpearson
Feb 24 2016 19:02
:) Will do!
Dana Lewis
@danamlewis
Feb 24 2016 19:03
thank you! (and same for anyone else working through this part, since we know those docs need further love) :)
Oskar Pearson
@oskarpearson
Feb 24 2016 19:05
Does anyone know of a way to turn debug mode on in nightscout or something? Like show the queries it's trying to run against mongo or something?
enabling feature: openaps
enabling feature: pump
enabling feature: delta
enabling feature: direction
enabling feature: upbat
enabling feature: errorcodes
enabling feature: ar2

body-parser deprecated bodyParser: use individual json/urlencoded middlewares lib/api/treatments/index.js:10:17
body-parser deprecated undefined extended: provide extended option node_modules/body-parser/index.js:105:29
body-parser deprecated bodyParser: use individual json/urlencoded middlewares app.js:26:22

/root/cgm-remote-monitor/node_modules/browserify-express/node_modules/watchr/node_modules/typechecker/es6/lib/typechecker.js:29
                for ( let type of types ) {
                          ^^^^
Listening on port 1337
running data.update
query: { date: { '$gte': 1456167904275 } }
query: { created_at: { '$gte': '2016-02-16T19:05:04.275Z' } }
query: { created_at: { '$gte': '2016-02-20T19:05:04.282Z' } }
running websocket.update
browserify -- bundled [3.02s] /public/js/bundle.js
browserify -- writing file /root/cgm-remote-monitor/bundle/bundle.out.js
browserify -- watching /root/cgm-remote-monitor/bundle/../lib/
emitted clear_alarm to all clients
Jim Matheson
@jmatheson
Feb 24 2016 19:07
do you have CGM data in there?
before i had my dexcom i was trying to set up NS by pushing pump data only and i remember similar issues. I had also tried spoofing BGs to push in but i think i was having problems with dates. In general if the BGs aren’t there things don’t seem to display right.
i could see entries in the db like you did, but nothing visually
Oskar Pearson
@oskarpearson
Feb 24 2016 19:10
> db.devicestatus.find()[13]
{
        "_id" : ObjectId("...."),
        "device" : "openaps://HOSTNAMEHERE",
        "pump" : {
                "battery" : {
                        "status" : "normal",
                        "voltage" : 1.31
                },
                "status" : {
                        "status" : "normal",
                        "timestamp" : "2016-02-23T23:25:52.000Z",
                        "bolusing" : false,
                        "suspended" : false
                },
                "reservoir" : 125.2,
                "clock" : "2016-02-23T23:25:32+00:00"
        },
        "openaps" : {
                "suggested" : {
                        "bg" : ...,
                        "temp" : "absolute",
                        "snoozeBG" : ...,
                        "timestamp" : "2016-02-23T23:27:15.000Z",
                        "reason" : "...",
                        "eventualBG" : ...,
                        "mealAssist" : "Off: Carbs: 0 Boluses: ... Target: 101.5 Deviation: 16 BGI: 0",
                        "tick" : "+6"
                },
                "iob" : {
                        "timestamp" : "2016-02-23T23:27:14.000Z",
                        "bolussnooze" : 0,
                        "activity" : 0,
                        "basaliob" : 0,
                        "iob" : 0
                },
                "enacted" : {
                        "requested" : {
                                "duration" : 30,
                                "rate" : ...,
                                "temp" : "absolute"
                        },
                        "bg" : ...,
                        "temp" : "absolute",
                        "snoozeBG" : 166,
                        "timestamp" : "2016-02-23T23:24:05.000Z",
                        "reason" : "Eventual BG... ",
                        "rate" : ...,
                        "eventualBG" : ...,
                        "recieved" : false,
                        "duration" : 30,
                        "mealAssist" : "Off: Carbs: 0 Boluses: 5.2 Target: 101.5... ",
                        "tick" : "+..."
                }
        },
        "created_at" : "2016-02-24T19:06:29.513Z"
}
No glucose values, no - but I'd expect OpenAPS and pump information somewhere, right?
Jim Matheson
@jmatheson
Feb 24 2016 19:11
i never got it to work without the BGs
@jasoncalabrese may know more to the why
even the reporting would not show anything. same thing as you, only way i could see was to query data directly from db to verify it was in fact there.
Scott Leibrand
@scottleibrand
Feb 24 2016 19:13
NS is designed to show BG info. Without BGs, it's pretty useless.
Oskar Pearson
@oskarpearson
Feb 24 2016 19:14
Ah, ok - so if I get glucose data in, then the other features might suddenly start working? That's a great hint
Scott Leibrand
@scottleibrand
Feb 24 2016 19:14
Yep, should.
Oskar Pearson
@oskarpearson
Feb 24 2016 19:14
I was just starting somewhere :)
Let me refocus on that :)
Thanks!
Scott Leibrand
@scottleibrand
Feb 24 2016 19:14
Starting backwards apparently. :-p
Oskar Pearson
@oskarpearson
Feb 24 2016 19:14
haha!
Jason Calabrese
@jasoncalabrese
Feb 24 2016 19:14
yeah, the whole chart is sized (the domain) based on cgm data
think we could change that now that we have so much other data
use the NS api is a good way to debug
/api/v1/devicestatus.json
/api/v1/entries/sgv.json
/api/v1/status.json
Great. devicestatus.json is returning as-expected, but sgv is an empty array (also as expected)
Rachel Sandlain
@audiefile
Feb 24 2016 19:18
I think nightscout needs 3 data points for everything to work/display properly
Jason Calabrese
@jasoncalabrese
Feb 24 2016 19:18
also you can do advanced finds like /api/v1/treatments.json?find[enteredBy]=/openaps/ or /api/v1/treatments.json?find[insulin][$gte]=2
Scott Leibrand
@scottleibrand
Feb 24 2016 19:24
The docs make a (potentially undocumented) assumption that you already have NS a working for CGM data.
eyim
@eyim
Feb 24 2016 19:28
Yep - my bad. When I put those docs together that was my assumption. I didn't think people would be setting up an Openaps system first before doing a NS implementation
Oskar Pearson
@oskarpearson
Feb 24 2016 19:29
:)
There's always one isn't there. In this case... it's me.
Jason Calabrese
@jasoncalabrese
Feb 24 2016 19:34
in your ENABLE above there's a typo
Jim Matheson
@jmatheson
Feb 24 2016 19:35
I actually think a lot of people are doing NS second. That was my approach. I had no need for NS until Openaps.
Jason Calabrese
@jasoncalabrese
Feb 24 2016 19:35
should be ENABLE='openaps pump'
Jim Matheson
@jmatheson
Feb 24 2016 19:35
i think i remember 1-2 other people in here with the same
As I mentioned last night, the Nightscout brand association for me was a product from two years ago for parents before Dexcom had share. And it required Android. I always thought “neat project”, but no value to me.
Jason Calabrese
@jasoncalabrese
Feb 24 2016 19:38
@oskarpearson you probably want BASAL_RENDER='default' too
Dana Lewis
@danamlewis
Feb 24 2016 19:38
@oskarpearson also, we want to not use gitbooks anymore, use the docs from github directly…not sure everything still renders in gitbooks
Jason Calabrese
@jasoncalabrese
Feb 24 2016 19:40
Nightscout means different things to everyone, got kind of overloaded I think
Jim Matheson
@jmatheson
Feb 24 2016 19:41
yea.
Oskar Pearson
@oskarpearson
Feb 24 2016 19:41
@jasoncalabrese Apologies, that was a cut-and-paste issue due to me running in screen and some weirdness with vertically-split-screens. It was actually right on the server.
Jim Matheson
@jmatheson
Feb 24 2016 19:42
@danamlewis in the screenshot above of github popular pages, is that first item listed the Git book?
Jason Calabrese
@jasoncalabrese
Feb 24 2016 19:43
I always thought of Nightscout as system
think the system finally has all the parts now
Oskar Pearson
@oskarpearson
Feb 24 2016 19:43
Aaand... now I have glucose data. Awesome. Still no pump or other data, but I'll figure that out next :)
Dana Lewis
@danamlewis
Feb 24 2016 19:43
@jmatheson not sure what you mean by screenshot above. But the link to use/give people for docs is https://github.com/openaps/docs
Jim Matheson
@jmatheson
Feb 24 2016 19:45
Screen Shot 2016-02-24 at 11.44.07 AM.png
just wondering if our efforts to point people elsewhere is not effective if that first, most popular page is the book.
Peter Miller
@ochenmiller
Feb 24 2016 19:45
There is a limit for max temp set by openaps at any time, which is equal to 3*max_daily_basal = 3.6 in your case. This limit overrides your pump setting for max_basal. There is also max_iob, which is a different limit - it applies to a limit for the total extra insulin delivered by openaps over past 4 hours (or whatever DIA is)
If you want a more agressive number, you can set a short basal at a higher rate sometime during the day.
Dana Lewis
@danamlewis
Feb 24 2016 19:46
not sure what timeframe he pulled that data from, but unless someone speaks up loudly and wants to claim ownership and fix all the renderings that break every time we update the docs.. I want to delete the gitbook..
Jim Matheson
@jmatheson
Feb 24 2016 19:46
i think deleting it would be an effective solution :)
Dana Lewis
@danamlewis
Feb 24 2016 19:47
I get the UI makes it look pretty. but I’m not sure it’s worth the time for those of us who can make a bigger impact actually improving the content of the docs rather than chasing formatting issues.
K.
Ben West
@bewest
Feb 24 2016 19:47
what do you mean delete the gitbook?
Dana Lewis
@danamlewis
Feb 24 2016 19:48
do whatever we need so people don’t end up on the gitbook version with outdated docs
Ben West
@bewest
Feb 24 2016 19:48
I've been wondering if we need to stand up our own build and not rely on their servers
ok, yeah, we can do our own build
should probaby file issue: move off gitbook builds
Jim Matheson
@jmatheson
Feb 24 2016 19:51
@danamlewis that data is the last 2 weeks
Dana Lewis
@danamlewis
Feb 24 2016 19:52
@bewest how will that result in a different outcome? and who does, and what permissions get locked in by who does it first, etc.?
@jmatheson k, thanks, good to know
but not sure if that’s just the book file on github or that’s actually people linking to the gitbooks view? @scottleibrand , do you know?
Scott Leibrand
@scottleibrand
Feb 24 2016 19:54
Not sure. Let me see if it links.
Jim Matheson
@jmatheson
Feb 24 2016 19:54
i think if @scottleibrand drills down on that page it will show referers to it. best bet would be to 301 it to the main github docs or worst case just get rid of the content and put a 1 liner to say go here for the docs
Ben West
@bewest
Feb 24 2016 19:55
that's purely the github views, not the book
Jim Matheson
@jmatheson
Feb 24 2016 19:55
cool, that answers the question then :)
Scott Leibrand
@scottleibrand
Feb 24 2016 19:56
Ben West
@bewest
Feb 24 2016 19:56
the renderings can live anywhere, it's a matter of making travis or wercker (or readthedocs) build the docs on every commit instead of gitbook
readthedocs is generally very good, also they allow versioning the docs nicely
so you have permalinks to all versions ever
Dana Lewis
@danamlewis
Feb 24 2016 19:58
TBH, one of my concerns is linking to old stuff. People get happy with their link and want to give it out..that’s why I like github.com/openaps/docs because I know people will get the current versions
Ben West
@bewest
Feb 24 2016 19:58
same deal with readthedocs and others, the latest link is always latest
I used readthedocs for some of the FDA presub builds:
Dana Lewis
@danamlewis
Feb 24 2016 19:59
is it the same URL, though?
Scott Leibrand
@scottleibrand
Feb 24 2016 19:59
I vote we delete the gitbooks.com one, and anyone who cares enough can set up others if they don't think github.com/openaps/docs is good enough
Ben West
@bewest
Feb 24 2016 19:59
ah, so proprosal is to use github itself as the main viewer?
Scott Leibrand
@scottleibrand
Feb 24 2016 19:59
yes
Dana Lewis
@danamlewis
Feb 24 2016 20:00
yes
Jason Calabrese
@jasoncalabrese
Feb 24 2016 20:00
I ran a little tool on the NS readme and other markdowns that generates a toc
with that the github view is better
Scott Leibrand
@scottleibrand
Feb 24 2016 20:01
nice. should use something like that to update summary.md or similar
Ben West
@bewest
Feb 24 2016 20:01
hmm, I like the ease of updating it
the toc there is generate
dunno, it's nice to have versioned docs with multiple downloads and available in multiple languages
Jason Calabrese
@jasoncalabrese
Feb 24 2016 20:03
I used this, it's a local install https://github.com/thlorenz/doctoc
I did like the gitbooks, too bad it didn't stay current
Ben West
@bewest
Feb 24 2016 20:04
that's nifty
here's another one a bit more like gitbook http://nightscout.github.io/tr-templates/
this build is done on wercker and is always current
Scott Leibrand
@scottleibrand
Feb 24 2016 20:08
if one of y'all want to manage that, feel free. pretty sure Dana is more interested in the actual content though.
Dana Lewis
@danamlewis
Feb 24 2016 20:12
:+1:
By the way - just sent out a note, planning a Seattle meetup for next week. If you’re a Seattle-area person lurking here and didn’t get it, please ping me so I can forward you details!
Chris Hannemann
@channemann
Feb 24 2016 20:19
I'm watching this docs conversation with interest, and if it was three months ago, I'd probably be pulling hard to keep gitbooks (or similar) alive. However, since I'm no longer actively working on the docs, I'm not going to argue in favor of something that I'm not willing to support. I agree with Dana that content is paramount, but I think usability is a not-so-distant second. I would strongly urge you to think of the broader intended audience and consider other UI options.
Dana Lewis
@danamlewis
Feb 24 2016 20:23
thx for chiming in, @channemann
Chris Hannemann
@channemann
Feb 24 2016 20:25
np
Nyadach
@Nyadach
Feb 24 2016 20:43
as a complete noob to all this, I'm not partial to anything as long as whichever wins is more current and fills in the gaps
William Fallon
@inform880
Feb 24 2016 21:37

Hi! I'm having some trouble setting up my nightscout uploader files to actually upload. I followed the documentation very carefully, but I'm getting this error:

Cannot find module '/home/pi/apsfallon/monitor/clock-zoned.json'

Jim Matheson
@jmatheson
Feb 24 2016 21:40
seems like it is saying that file is missing. It’s expecting it to be there. Have you created a clock-zoned.json ?
nightscout expects a certain date format, so you need to be sure to do the date conversion and save it in zoned files first. There’s a chance the documentation is not quite right yet.
Rachel Sandlain
@audiefile
Feb 24 2016 22:34
@oskarpearson I think you need careportal enabled for treatments to show
Oskar Pearson
@oskarpearson
Feb 24 2016 22:49
@audiefile thanks! I figured that out, but haven't got the data flowing yet. Enough for today though - will resume tomorrow. Thanks all for your help
Dragan Maksimovic
@dm61
Feb 24 2016 23:25
I think github.com/openaps/docs is good enough. Maybe add (previous section), (next section) and (toc) links on top and bottom of each section
William Fallon
@inform880
Feb 24 2016 23:26
@dm61 I totally agree with you. Navigating through it is kinda annoying.
Nyadach
@Nyadach
Feb 24 2016 23:29

am trying to figure the cron timing and it works sometimes, but not others...it will run it once every hour or two then ignore it a while, anyone mind looking it over and telling me what I messed up please?:

PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin/:/sbin:/bin

5 * * * * cd /home/pi/my_openaps && openaps retry-loop >> date && /home/pi/my_openaps/enact/enactlog.log 2>&1

thanks

Scott Leibrand
@scottleibrand
Feb 24 2016 23:33
you can't >> date
you probably want retry-loop; date >>
William Fallon
@inform880
Feb 24 2016 23:37
@Nyadach Are you trying to run it once an hour in your current configuration?
Scott Leibrand
@scottleibrand
Feb 24 2016 23:39
if you want to run it every 5 minutes (instead of every hour at 5 minutes after) you'll want */5 instead of 5
Nyadach
@Nyadach
Feb 24 2016 23:39
was hoping for 5 mins with it?
ahh
so thats 5hours or something?
William Fallon
@inform880
Feb 24 2016 23:39
That would be minute 5 of every hour
Scott Leibrand
@scottleibrand
Feb 24 2016 23:39
no, it's when minute = 5, hour = anything, day = anything, etc.
Nyadach
@Nyadach
Feb 24 2016 23:41
thanks
Nyadach
@Nyadach
Feb 24 2016 23:42
example there has it through a script, seen a few examples on here using a script also, better solution than putting it all in a line or no real advantage/disadvantage?
Scott Leibrand
@scottleibrand
Feb 24 2016 23:43
just a question of how long and hard to read your one-liner is. yours isn't very long
William Fallon
@inform880
Feb 24 2016 23:51

I'm still having trouble getting my info to upload to nightscout. I guess I don't understand where reports in this line:

openaps report add monitor/upload-status.json JSON ns-status shell monitor/clock-zoned.json monitor/iob.json enact/suggested.json enact/enacted.json monitor/battery.json monitor/reservoir.json monitor/status.json '

are coming from

specifically the battery.json and reservoir.json
Scott Leibrand
@scottleibrand
Feb 24 2016 23:57
you have to set up reports to query the pump for those things
check openaps use pump -h to figure out the use command for each
William Fallon
@inform880
Feb 24 2016 23:57
oh I understand
I completely missed the point in the documentation