These are chat archives for juttle/juttle

13th
Jan 2016
Rodney Lopes Gomes
@rlgomes
Jan 13 2016 00:22
@demmer failure on master after your last merge https://travis-ci.org/juttle/juttle/jobs/101986379 is already being taken care of here juttle/juttle#152
Michael Demmer
@demmer
Jan 13 2016 01:02
Ah ok thanks. That change looks good to merge to avoid this flakiness.
Rodney Lopes Gomes
@rlgomes
Jan 13 2016 01:07
Its in now.
David Cook
@davidbcook
Jan 13 2016 18:28
Hey everyone!
Daria Mehra
@dmehra
Jan 13 2016 18:28
hey - cook’s in the house!
David Cook
@davidbcook
Jan 13 2016 18:29
woo wooo!
The company I just started working at uses Orchestrate to store their data. https://orchestrate.io/ Looks a lot like certain aspects of the old version of Jut did
Daria Mehra
@dmehra
Jan 13 2016 18:31
got use cases for juttle over there?
David Cook
@davidbcook
Jan 13 2016 18:33
Oh yeah! We're using biometric data to understand the emotions people are feeling. My job is to analyze the data to determine what emotions like fear, elation, stress, and tension look like. I'd love to use Jut for that
Daria Mehra
@dmehra
Jan 13 2016 18:34
well the OSS projects are out
David Cook
@davidbcook
Jan 13 2016 18:34
I'm gonna work with our team here to figure out the best way to incorporate those into our infrastructure in the next couple days
Daria Mehra
@dmehra
Jan 13 2016 18:34
if your data is in orchestrate, do you have to use their rest API to extract it, or can you also talk directly to elastic search?
David Cook
@davidbcook
Jan 13 2016 18:35
It looks like we have to use their API
Daria Mehra
@dmehra
Jan 13 2016 18:35
we have read http supported out of the box, so you can do that
David Cook
@davidbcook
Jan 13 2016 18:35
cool
Daria Mehra
@dmehra
Jan 13 2016 18:35
and maybe keep some of the logic inside the orchestrate query, so you don’t get tons of raw data coming back, but then further processing would be in juttle.
if your team would like a demo of juttle, we can arrange for that. or you can be the ambassador :)
Matthew Nibecker
@mnibecker
Jan 13 2016 18:37
Can’t wait to see what things the Dave will Cook up with Juttle + Emotions
Mark Stemm
@mstemm
Jan 13 2016 18:37
can markdown emit emojis
also welcome
!
David Cook
@davidbcook
Jan 13 2016 18:38
I'm trying to be the ambassador for now, but we'll probably run into questions I can't answer especially with the new way things work
Daria Mehra
@dmehra
Jan 13 2016 18:38
emojis are tricky in markdown
even in non-juttle one
you have to load extra plugins for them
David Cook
@davidbcook
Jan 13 2016 18:39
but I'm excited to hopefully use Juttle and work with you guys (in a way) again
Daria Mehra
@dmehra
Jan 13 2016 18:39
awesome!
Mark Stemm
@mstemm
Jan 13 2016 18:39
how do you feel about being able to edit juttle in your own editor
Daria Mehra
@dmehra
Jan 13 2016 18:39
@mstemm like in the docs, we have to use markdown_include python extension and it looks at local png files for certain emoji, it doesn’t “just work”.
Michael Demmer
@demmer
Jan 13 2016 18:39
Indeed — @davidbcook as i mentioned before please feel free to point your developers here. We’re eager to enable users at this point (and to help find the problems that people will run into with the new world of juttle).
David Cook
@davidbcook
Jan 13 2016 18:42
I would miss the syntax error detection of the explorer, but I don't think it would be that much of a problem
Mark Stemm
@mstemm
Jan 13 2016 18:43
I think our editor hooks will highlight errors, although it will be at execute time not on the fly
David Cook
@davidbcook
Jan 13 2016 18:43
One question: do realtime and historical queries work in the same way as they used to? It seems like there's more of a separation between the two now
@demmer will do
Michael Demmer
@demmer
Jan 13 2016 18:45
The first question is what led you to think that?
As to the actual answer it’s a bit nuanced — it’s true that in the new world, we no longer have juttle components involved with the ingest pipeline so we don’t have a tap into the incoming “live” data like we used to. However the plan is that for adapters like elasticsearch, we would still support combined historical + live queries but would implement it with a “pseudo-live” mode where the read command would first do the historical part, then query every N seconds for new arrivals.
Now - when accessing something like orchestrate with the http adapter, my thought is that you would need to tell the read command something about how the API to which you’re communicating handles time filters (which we don’t yet have). At that point you could map the -from and -to arguments to corresponding API requests, and then the support for combined historical and live would come for free
Now that’s not something we have yet but would be pretty easy to add on and has been in my head as something we should do.
David Cook
@davidbcook
Jan 13 2016 18:47
I got that impression because it seems like the data engine no longer exists and that was what allowed Juttle to easily tap into both live and historical data feeds
Mark Stemm
@mstemm
Jan 13 2016 18:48
it’s up to each adapter to do what it takes to handle a live vs historical vs mix of both query
David Cook
@davidbcook
Jan 13 2016 18:48
makes sense
David Cook
@davidbcook
Jan 13 2016 20:31
So I'm running outrigger and juttle locally for now, and I just had some trouble running a simple program. I had 2 syntax errors, but outrigger didn't tell me I had any syntax issues, it just gave me a black screen with the run button and this in the console:
Failed to load resource: the server responded with a status of 400 (Bad Request)
(you're probably aware of this but just thought I should bring it up)
Michael Demmer
@demmer
Jan 13 2016 20:33
Yeah thanks — that’s high on our list of things to fix imminently:
juttle/outrigger#29
David Cook
@davidbcook
Jan 13 2016 20:34
awesome, I'll check github issues in the future
Michael Demmer
@demmer
Jan 13 2016 20:37
np — if you run into other issues and are up for it, feel free to add a comment or an upvote to indicate the issues you are hitting that we should focus on
David Cook
@davidbcook
Jan 13 2016 20:38
Definitely
Daria Mehra
@dmehra
Jan 13 2016 20:43
if searching issues across multiple repos proves hard, look at our waffle board, which combines them all: https://waffle.io/juttle/juttle
David Cook
@davidbcook
Jan 13 2016 20:44
Cool! You all must be stoked to be rid of Atlassian ;)
Daria Mehra
@dmehra
Jan 13 2016 20:44
no comment :)
David Cook
@davidbcook
Jan 13 2016 20:44
haha
Daria Mehra
@dmehra
Jan 13 2016 20:45
it wasn’t a very practical option for OSS. we tried Trello, which was nice, but didn’t integrate with GitHub, and then settled on Waffle, which does.