Player is loading...
Interactive Computing with Jupyter and Almond
Sören Brunk, USU Software AG
Thursday, June 13, 2019 · 3:46 p.m. · 47m 41s
Jupyter notebooks, interactive, web-based documents that integrate code with rich output and documentation, have become tremendously popular in data-science and other fields. I'd like to convince you that Jupyter notebooks are a powerful tool for prototyping and teaching, useful to you even if you're not a data-scientist. You'll learn why notebooks combine many strengths of REPLs and worksheets, and how the almond project integrates Jupyter with Scala, Ammonite, Spark and more. To illustrate these points and the interactive nature of Jupyter, the talk will be presented through running notebooks. Required Knowledge: Basic Scala knowledge and experience using the REPL is beneficial to understand the examples but not required. Sören is a software engineer at USU Software AG where he works on the Katana platform, helping companies to understand their industrial big data. He's a contributor to almond and maintains the almond examples, a growing collection of Scala based Jupyter notebooks you can try in your browser. He also writes about machine Learning with Scala in Journals and on his blog.
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:00
we can get started now
00:00:03
sell welcome to the session
00:00:06
um and so on um and i'm a software engineer on the cat and the
00:00:10
data sense that from a a interactive computing is something i'm very excited about so
00:00:17
i really hope i can share some of this excitement you today so let's get started
00:00:25
learning is a fundamental part of our daily work
00:00:28
uh software developers problem solving is essentially learning right
00:00:33
so we need to understand the problem and then learn about the right approach how to solve it
00:00:39
and then the technology we using is conch constantly evolving
00:00:43
languages libraries tools and we wanna keep back so we learn
00:00:49
but how do we learn
00:00:51
turns out that often we can look at learning as
00:00:54
a three phase process we usually start with observing things
00:01:00
and then we try to replicate um that the the things
00:01:03
we've observed like i'm trying i'm trying some example code ourselves
00:01:11
and finally we start exploring start changing things and and try something new
00:01:16
and um that exploration phase is essentially a an interactive feedback loop so
00:01:23
in that loop we start was trying trying something new um
00:01:28
and then we observed the feedback we get from the rolled and um
00:01:32
based on that feedback we update our internal model um so we learn
00:01:38
and then we try the next thing i'm observe new
00:01:41
feedback uh and so on and um this feedback is
00:01:47
is very important because this is essentially what enables us to learn and um we really wanted to be fast
00:01:55
um in some situations um because it's essential to our survival um
00:02:02
no one was aren't memories but more generally because only with
00:02:07
fast feedback we get a truly interactive a feedback loop
00:02:10
and that gives us a much better learning experience of
00:02:15
so what about exploring coder beta um one approach for
00:02:20
for creating this this fast feedback loop is a interactive computing
00:02:26
so interactive computing basically means we have the system which lets us um
00:02:32
right a block of code immediately returns a immediately runs that code and to inspect the
00:02:38
results and we can do that in the loop giving us this this fast feedback cycle
00:02:44
and if you think about it that definition essentially describes the rubble so
00:02:50
how many of you use of apple at least from time to time
00:02:55
okay that so that's almost hundred percent almost everyone cool um
00:03:01
so yeah the rap lettuce and an expression evaluates it immediately
00:03:06
and prints the result value and so we can do this in the loop and continuously we find things
00:03:12
huh or we can use previous results and we don't have
00:03:15
to explicitly print things because the rubble does that for us automatically
00:03:20
and this is really great for for exploring the language or or
00:03:24
or trying out things because yeah because we have this fast feedback loop
00:03:29
um but the rattle that's got to wrap all also has a few pinpoints
00:03:35
um there's um no support for for for editing multi line expressions um
00:03:43
there was no no pretty printing and limited syntax highlighting and all the completion
00:03:50
and most importantly there's no native support
00:03:53
for for loading maybe dependencies uh for
00:03:56
solving dependencies and that makes it much harder to exploring the new libraries right
00:04:03
um fortunately there's another apple implementation i'm from howie
00:04:07
i mean i'd so how many of you use them a night at least from time to time
00:04:13
okay i'd say still for fifty percent that's all that's
00:04:18
cool um so i mean i'd address many of the
00:04:22
of the issues that none of the default grapple it
00:04:24
has nicer nicer formatting and and and syntax highlighting um
00:04:30
um proper multiply next uh expression editing um but most most importantly
00:04:36
it has um support for dynamic a dependency resolution um with this um
00:04:44
import so id syntax so we can basically um import any j. v. m. library
00:04:51
and use it immediately and in the same um in the same rabble session
00:04:56
which is really cool for yeah exploring new libraries and
00:04:59
trying uh and and learning new libraries trying out new things
00:05:05
regardless of whether we use the default grapple or i mean i'd though um there
00:05:09
are some limitations that are that are inherent to the to the rapid design and um
00:05:16
this is mostly because uh the rap was optimised for for expiration
00:05:20
from scratch for for writing writing some code in seeing results fast
00:05:26
um it's it's but can cut out again is is
00:05:31
harder reading copying code um is is not not as easy
00:05:36
and that's because we have these interleaved inputs and outputs
00:05:40
and um all these refinement attends in in our history
00:05:44
and also changing all the expression can be up a bit hard we we have to find it
00:05:50
and if we have dependent expressions um we have to find them all and
00:05:54
um um and then to run by one and that can quickly become tedious
00:06:01
yeah sharing salsa difficult because we don't don't really have a system document all only the the history file
00:06:09
so because of this this focus on exploration um the observed
00:06:12
in the replicate faces are are not part of the rapid design
00:06:17
um but from learning perspective we rally to
00:06:20
start exploring um we usually need some input first
00:06:25
so what often happens is that we have something like ah like here on the screen shot so we have um
00:06:32
we have a browser window some some tutorial of for the observed face
00:06:37
and then we have a rapport window um where we have to copy the code over
00:06:41
for for the replicate phase and this is just at some overhead so can we do better
00:06:49
let's look at the captain of the rubble worksheets
00:06:54
so i'll worksheets are quite similar to the rabble but
00:06:58
we had this full editor window where we can write expressions
00:07:01
and uh when we get run or save the worksheet uh the expressions will be evaluated
00:07:06
it evaluated and um and the results are shown on the on the right hand side
00:07:13
and um yeah because we have this editor window um without the the interleaved inputs and outputs
00:07:21
i'm reading popping in and changing code um it's
00:07:23
much easier and changing is also easier because um
00:07:29
because when he'd run all the expressions are evaluated in one go
00:07:33
so in contrast to the rapid we don't have to to manually track dependent expressions
00:07:39
um the drawback is that we lose control over um uh only went to execute which expression
00:07:46
and therefore worksheets aren't aren't very well suited for
00:07:50
for longer running computations like like loading large data sets
00:07:55
um so it's also nice about them is the nice id integration
00:08:00
especially in and in the the doggy worksheets like like we see here
00:08:05
and um
00:08:07
and that they are actually back i file so we can save them put them in version control and ensure that
00:08:14
and then we also have these online worksheets with uh which make sharing in easier so
00:08:20
because they don't have the nice id integration um but it's
00:08:26
um yeah chang's really nice you can just create a link and send it to someone and
00:08:31
um they don't have to to they don't need any local set up so we have um
00:08:37
we have a scares the forty grams gala and um stella fill force coloured yes um
00:08:45
so looking at a worksheet from learning perspective um what's different compared to the rubble
00:08:51
well replication works a bit better um because it could provide worksheets
00:08:56
refilled with examples and copy and paste is also a bit easier um
00:09:03
but it's still not in a very integrated experience
00:09:06
we still have these these separate things right so
00:09:12
looking again at this this ballot for example um we
00:09:16
already have an an example code code snippet here right so
00:09:22
what is this joe wasn't just just static um what if you actually could
00:09:27
run the sink and turns out we can so if i hit control into here
00:09:33
um this will be evaluated a ride in my slides and we
00:09:37
get our double salary which three nice um but we can
00:09:42
also um at the sink so you have a real editor here
00:09:48
um so double salaries nice but
00:09:53
what about more
00:10:02
so let's give us a triple salary and now we can run the thing again and you will go more money
00:10:12
so
00:10:14
how does this work um turns out we actually need to put a note here
00:10:20
uh it's just shown as a slide show i'm using a plug in cold
00:10:24
rice and usually it looks much but it looks more like my more like this
00:10:30
which is the the classic uh jupiter interface um which lets us interact
00:10:35
with with these these notebooks of um but but what what are these notebooks
00:10:45
notebooks are interactive web based documents so their web based the run in the browser
00:10:49
the interactive we can we can added things and and and one code and we have
00:10:54
um they're made of cells with two types of cells uh one s. documentation and one is code
00:11:01
so um documentation cells are just just marked down
00:11:06
we we have all the usual marked on formatting options
00:11:10
um we can include images skilled fences um lake equations um
00:11:16
but the cool thing is that we can um added them right in the browser
00:11:21
so if i double click on this this uh will change to turn editor window
00:11:27
and i can write something here
00:11:37
and um when i had control enter again and this will be
00:11:42
the updated um you get updated rendered you and then we have
00:11:49
got cell three for it already seen one so they're basically added
00:11:52
first that lead us right right code similar to to work sheet
00:11:58
and we can of course evaluate that so let's try this again
00:12:03
and oops i we get a compile error so we get all these not the
00:12:07
error messages uh if something something is wrong and so let's just fix that tough
00:12:14
and when it began and here we go and um but we aren't restricted to single single
00:12:22
cells we can uh have any number of cells in the notebook and we can freely um um
00:12:28
makes coats else's documentation sell so here we have one talk talk sell a full by
00:12:36
but to called cells and of course we can evaluate them
00:12:40
and with black and we can even put these things side by side so displayed nice um
00:12:49
so instead of having these two separate things documentation and and and code um
00:12:56
we now have an integrated a document containing both of
00:13:02
and that's really nice for for teaching an interactive learning um
00:13:07
yeah because we have we we know have integrated all these these learning phases in one day
00:13:15
um but not books are more and so
00:13:18
they also integrates a worksheets uh with freckles
00:13:24
and uh how this works is best seen in a little demo so let's switch to another notebook
00:13:32
uh_huh
00:13:38
so
00:13:41
if we look at a single code sell 'em that base
00:13:48
mostly like a worksheet so we have multiple expressions we have
00:13:51
this editor window and um we can about it this thing
00:13:59
oh i should be able to well we had a thing i let's give it another try
00:14:09
and he we go um
00:14:13
i liken worksheets when when we had run um would always uh evaluates all all
00:14:18
cells in one go so we we don't have to track a dependent expressions manually
00:14:24
it's also persistent um but it all uh it also saves the outputs so this
00:14:31
really cool for for sharing of putting these things online uh or exporting them into
00:14:36
into the p. d. f. or something like that um
00:14:43
on the other hand there are a little bit um limited for um
00:14:48
when it comes to to id features though there's not the nice lady integration
00:14:53
and i'm also the outputs are shown below the cell by default not on the right hand side
00:15:00
but it shouldn't be too easy to actually change uh the front
00:15:04
end to to have this the side by side you in notebooks to
00:15:09
um so if you look at school cells of note because the whole mm they behave more like a rap well
00:15:16
and that's because we we can run all the cells independently
00:15:19
um which gives us better control when to run specific code
00:15:24
and also all previous um results like s. to memory
00:15:27
so if we have this this um huge data set here
00:15:34
which takes a long time to load of the if we uh would run that's in the in the worksheet
00:15:41
and um if you want to to change something here we'd have to run the whole thing
00:15:46
again which uh takes takes some time right i'm gonna not okay we can just split the cells
00:15:57
and um run them independently so we can load that data
00:16:05
and then it's cached in memory and we can uh use it in our computations and
00:16:11
a change things and it's much faster
00:16:21
not books also give us um more uh flexibility for
00:16:25
for prototyping we can we can select multiple cells and um
00:16:31
run them in one go oh we can run all cells um we
00:16:36
can move cells around so we have this a drag and drop support here
00:16:41
ah cut copy and paste uh even between different
00:16:45
notebooks a split in merchant all these convenience features
00:16:49
um that make 'em in that that make this is a good thing for for prototyping
00:16:59
yeah so
00:17:03
what about exploring data
00:17:06
with something rattle like like like this part sell here of course we can do that
00:17:12
um but because this is eternal and cuts text based it's
00:17:18
usually not not the best option to show data or or
00:17:23
the lights things so what often happens is that
00:17:28
we have these these separate things so we we render
00:17:33
um we use an external graphics system to render our our our
00:17:37
visualisation so um and have them in a in a separate window
00:17:42
and yeah so this can not not the the best the most integrated experience
00:17:48
and with notebooks on the other hand um we can do you can do a little bit better
00:17:56
so let's stay with this um this very same
00:18:00
example just said and screenshot the titanic data set
00:18:05
and run this and now we have a nice um h. t. m. l. representation of of our data nice table
00:18:12
and we can even sorta things and um so i i'd argue this if this is a much better to read um
00:18:21
and the same for visualisation so we um also have an integration for this
00:18:26
for that i guess library here and notebooks and we can just on that
00:18:32
and he we go we have our our plots um to uh in in an
00:18:40
embedded in the notebook with our code and we can share these things together um and
00:18:48
we aren't even restricted to to static images because we have a browser we have javascript
00:18:54
available um when we use the the partly
00:18:58
library we um can create these these interactive um
00:19:03
oops these interactive um plots we consumed in
00:19:08
and uh you have basically interact with the plot
00:19:13
and um i we also not restricted to data we can also
00:19:19
visualise code which is a great for for for teaching for instance so
00:19:24
hey we use the the the ref three library which lets us um visualise they distract us um
00:19:33
and we prayed a simple a simple case class here and the a list of people
00:19:39
and then we use ref three to to render under this list so let's give it a try
00:19:48
and here we go so we have um this nice structure showing uh how
00:19:54
uh how how the list of how the elements of the list are connected um
00:20:00
this this i think this is really cool example for for actually explaining these things we can we can even
00:20:06
change things in one one this again that the the update and all that just within the notebook um so
00:20:16
like with this learning example um
00:20:20
um we have with the rapport we have this this this separate use um
00:20:27
but in notebook we have um we have our output and the
00:20:33
and the source code integrated so we have our output embedded with the code that produces it
00:20:38
so it stays in the same context and i think this is really cool cool feature uh
00:20:46
so
00:20:49
what i want to next is look a little bit uh at the technologies behind this so um
00:20:59
the examples here they were all run an egyptian notebook and tip there is uh
00:21:05
there's a huge project a devoted off to
00:21:09
develop open source software open standards and services for
00:21:14
for interactive computing and um the flexible is this as the jupiter
00:21:19
notebook inflected software but there's also a huge ecosystem around this um
00:21:26
and it's not restricted to a to a single programming language so we have um
00:21:32
we have support for um python and our
00:21:36
um c. sharp and uh and it's java
00:21:41
but also functional languages like f. sharp pascal and of course um of course canada i
00:21:50
um so jupiter came out of the type item shell um i pricing is yeah basically an
00:21:59
advance interactive python shell a similar to to what we have of i mean i'd first gotta
00:22:07
and um in two thousand eleven and
00:22:10
the eye python creators um released the the
00:22:16
the i price notebook interface which is yep still pretty similar to what we what we have today
00:22:23
so we have this this documentation and rubble code and uh
00:22:29
also rich output integrated and and and just the output document
00:22:34
um
00:22:36
but then i realise that this is actually something language agnostic um so
00:22:43
with project to putter they um added support for for other
00:22:47
languages basically moving the little language agnostic part out of high python
00:22:53
and um initially the support it um actually a a hyphen and our two guitar
00:23:00
and in two thousand eighteen um they released a jupiter lap so last year um which is
00:23:07
yeah basically the next generation user interface for notebooks and more
00:23:13
so let's have a brief look at at the uh at the components of the the church
00:23:18
stacked so we have the the user interface the notebook interface um which ones in the browser um
00:23:26
and lets users interact with with notebooks and then we have the notebook
00:23:31
server which um basically sets everything up
00:23:35
communication and delivers to the u. i.
00:23:38
and also manages the the notebook files and um we also have the carl
00:23:44
uh which is responsible for actually running running our code
00:23:49
and returning results uh and they're connected through the jupiter particle
00:23:55
um which does does this loosely coupled architecture um i um and that that
00:24:02
enables to to basically develop these components independently and controls the scale and distribute things
00:24:15
so there are um the bonds of how um different fronts available so we have the the classic notebook
00:24:22
and to to lap as it's a designated successor and
00:24:26
there's also an tracked from the net flicks fox um
00:24:30
and i'm up a plug in for for the atom and interpret
00:24:34
hydrogen which is a little bit more worksheet like but with the jupiter
00:24:39
with the jupiter back and and then uh the the major ideas
00:24:44
also have some limit that i'm not support but currently it's um restricted to pipe and
00:24:51
um but i've been hearing some rumours here but um
00:24:56
there's some ideas to to actually improve this colour story here
00:25:04
yeah and then when we have kernels and they're responsible for for running running the code or um
00:25:11
or also compiling or talking to the compiler for compiled languages
00:25:16
and they don't know anything about notebook so they just get a bunch of code and um
00:25:22
and uh on that code and and return results or error messages
00:25:28
and in principle that can be implemented in any language as
00:25:31
long as they support at the typical and nowadays they're around
00:25:36
one hundred different trials available for for all kinds of languages
00:25:42
so all these examples we've seen so far they um were executed on the common colonel
00:25:50
um this actually bans of of kernels available to to girls available for scholar
00:25:55
um and most of them are based on the on the default rabble
00:26:01
i mean i'd on the other hand is spilled on top of tonight and
00:26:05
we've seen that i mean i this was um all these nice convenience features home
00:26:11
and in in town and we inherent not inherit most of these features so for instance
00:26:19
um we have dynamic dependency resolution um which uses a course here in in in the
00:26:25
background so we can import any any um j. v. m. library any maven dependency basically
00:26:33
and it's just wondered and and or no notebook so here we have the discounts library
00:26:40
and we just import bad and use it for in our
00:26:44
calculations and so this really nice um we also get the
00:26:50
and syntax highlighting pretty printing
00:26:55
and um i
00:27:00
something that was recently at it is um uh
00:27:04
also uh tonight um his other completion for um
00:27:11
for for main um dependency so this really nice if you
00:27:15
don't know what the latest version of library you can just um
00:27:21
could um it shift and and it will show you what's what's available
00:27:28
um and of course we also have um on the completion on on on regular code
00:27:33
so here we uh knows that we're mapping over string and uses all those completions oh
00:27:41
on one string of so there's another feature type things and not across integration it
00:27:50
but it is a little bit unstable and and i'm gonna try them unless it's a
00:27:57
so if i hit um shifts shift tab i hope i don't kill the colonel multiple times
00:28:07
um can you see that yeah if um this
00:28:12
will give us a nice uh type hints here so
00:28:16
some some id like features and um
00:28:20
so what can you see from this from this type and from the map function
00:28:26
anyone
00:28:30
exactly so this is uh still not to to thirteen but it's coming uh i think we
00:28:38
waiting for cats or something like that and so and we when we click on that um
00:28:45
this actually fires up amanda browse server um which lets us um
00:28:51
look at the source code on this so this really cool to yeah to to see
00:28:56
uh to see the the code of of of something of a okay so let's go back
00:29:07
are you close that
00:29:12
we can also use the the i'm not i mean i'd a. p. i. in um
00:29:16
so not everything uh make sense but for instance we could take the um
00:29:22
the history so you you basically see the three uh
00:29:29
uh latest cells we've sent uh we we we for an
00:29:33
um which is sometimes sometimes handy horror to look at at
00:29:39
the repositories of calls at at new you may even repositories here
00:29:45
and um some and also gives us an a. p. i.
00:29:48
to talk to the to the front end for the browser
00:29:52
and this is what powers um these rich outputs
00:29:57
we've seen um in in in the data exploration examples
00:30:02
so if you import the the element display a message methods on
00:30:06
on and on the display i think they're actually automatically imported um
00:30:12
we get all these these methods uh for output so we can
00:30:18
and send some h. t. m. l. or and as e. g.
00:30:24
or um load an image and don't send that to the browser
00:30:30
um i like take equations and even cellist tripped
00:30:37
um
00:30:39
but this is somewhat restricted in your front ends like jupiter lap because of um security concerns
00:30:49
and basically we can uh if the fountain supports
00:30:53
something we can we can basically send any mime
00:30:58
type and if it from the knows how to display it uh we get we get these which outputs
00:31:04
so we could also have plug ins for for other things uh and based on
00:31:09
on this we have um some ah high level which it's like a progress bar
00:31:17
and if we keep this handle here um we can also update
00:31:24
the progress bar and an later i'm like yourself so let's try this
00:31:34
so this gives us a way to get to to
00:31:37
basically show show some progress unloading data things like that oh
00:31:44
uh_huh we can also read user input um hum
00:31:51
so this just don't be um a written into into this well
00:31:58
and um i oh yeah we also have a um a nice feature um for seeing what's
00:32:05
going on when there's some crack calculation in the background so if we have this future here
00:32:11
um that that's just on this uh we can see um this animation while it's running
00:32:19
and as soon as it's finished it should actually update
00:32:23
on to get to the final value um and the same
00:32:29
we have the same for four hours so iffy up that that x. here
00:32:36
we get this this would be updated value in in in the in the output
00:32:41
yeah and we can use this uh these eight guys in in our own libraries which we
00:32:47
can then provide a notebook users so we we could depend on the on the skull acrylic guy
00:32:54
and then in in our library um um use it so we
00:32:59
we let's just let's just simulate that here in in the cell
00:33:04
so we'd have um this i implicit instance of jupiter a. p. i.
00:33:09
available um to use and we we can use that to send
00:33:14
output to the front and uh and so let's just run that here
00:33:19
and then in our library we we can uh uh sorry not not that we can just him
00:33:24
a import this this library and and use it so um i think that we can do better right
00:33:32
oh not better okay okay i like this one i
00:33:39
and so far we've most look look at this um from the learners perspective
00:33:46
but um what about teaching um so one
00:33:51
thing i hope has become clears that i'm not
00:33:55
books are nice way for actually demonstrating life
00:33:58
code in slides um but what about library documentation
00:34:07
or so of course we could just great notebooks uh from scratch
00:34:12
um but usually or a has some documentation available
00:34:16
and uh most of the time this is um
00:34:19
mark down was cut fences for the examples i um
00:34:25
and if you think about notebooks are are quite similar
00:34:27
we also have um the smart down cells and then um
00:34:31
i told cells and just that the that we can actually run that
00:34:36
and there are encoded differently so notebooks are basically it adjacent document um
00:34:43
um fortunately we have a tools like note down that led us convert um this
00:34:50
martin with code fences to notebooks and also
00:34:53
converts the decode fans the actual code cells
00:34:58
and i'm i'm currently working on something that works a little bet a little bit better with
00:35:03
uh with took over and dark um and all this modify is um but it's not finished yet
00:35:12
and and then if you had that these notebooks with the documentation on of course you want to
00:35:18
to share then let someone uh actually use them
00:35:21
um without without uh too much too much effort right
00:35:27
and this is where binder comes into play which is really cool service um
00:35:32
so basically what we can just give it a a it get people and
00:35:41
um what it does is it creates a doctor contain no with a notebook
00:35:46
server and our um our notebooks and that we pull and give suspect the link
00:35:54
and um we can and and when you click on that link we we get
00:36:00
uh i'm running notebook service are not books so this one summer in the cloud
00:36:05
that's it but it's a free service and i think we got a little bit of time so let's just try this
00:36:14
so here we have the binder website
00:36:17
and um we're using the the um and example three people
00:36:24
and um
00:36:26
just launching this thing
00:36:29
and the first time you add a a ripple over if it up they something this
00:36:34
this will build the the doctor and the doctor image but now it's already built
00:36:39
so it should be a little bit faster if the engine networks yeah looks good so
00:36:45
he uh he we have a running up big server with with jupiter lap
00:36:51
and and the the common examples and we can
00:36:54
just try them out so we can uh for instance
00:37:00
look at this colour matter tree guide
00:37:04
and
00:37:06
just want things and i think it takes a little longer here
00:37:12
because the the the resources are limited and the top has to
00:37:17
to uh compile under nite stuff first but after some time it should actually around
00:37:27
let's give it a few more seconds
00:37:29
oh yeah that it also has to download things of course because um it's not cached so
00:37:36
here we secrecy in action
00:37:41
oh
00:37:44
oh
00:37:47
and ah okay the internet is no no not around
00:37:52
so um now we can use it here and um
00:37:59
just try things are and there's also
00:38:05
um other examples like um spark
00:38:11
um so here we can see the devious park integration of
00:38:17
hum of um and an action so i encourage you to to give it a try and just
00:38:22
there's a link in the in the comment box and in the example report to on this
00:38:36
so if you're interested um there's the jupiter website um for
00:38:42
information about people in general and how to run the sings
00:38:46
and element also has a shiny website with documentation for for
00:38:50
installation and um and use it um so check it out
00:38:56
and then uh we've just seen the example three people so they can just
00:39:01
sh try try these things out without without any local settlement and get a feeling um how this works i
00:39:09
and we'd also laugh to see contributors so um we want to
00:39:14
have more librarian framework integrations um and of course more examples um
00:39:21
then there's the story of better id integration and we also need a logo so if you're
00:39:28
if you have better design skills than me i'm kind of interested just talk to to to like song
00:39:37
um or me or john added up channel um or check out the the on and people
00:39:46
so to recap we we've seen that um
00:39:51
that apples and worksheets are really great
00:39:54
tools for for exploration and rapid prototyping
00:39:59
but they don't really cover these these other learning phases and um
00:40:04
to to the notebooks with notebooks we get this
00:40:08
disintegrated to use so we have to integrate a documentation
00:40:14
uh ronald code and also which output in a
00:40:17
single document in a single in a single view um
00:40:23
and that makes them a very powerful tool for for teaching
00:40:28
and for interactive learning um but of course also forte designs and stayed expiration
00:40:35
um and element
00:40:39
yeah by combining the the power of jupiter and i'm a night i'm being said experience discover
00:40:47
so thank you very much and i still we i
00:40:52
think we still got some time for for a few questions
00:41:02
thank you um
00:41:06
uh_huh
00:41:09
thing
00:41:11
uh_huh oh yeah bye sorry uh_huh uh_huh so okay so that's all have
00:41:18
i just got a question at that i'm using element in a in
00:41:22
a teaching environment 'cause i think the problem right now that i'm ah
00:41:26
how the orgasms colour work shops and the number one problem is always
00:41:30
yeah getting to but it's but this at the moment correctly sorting um
00:41:34
element is actually quite a and that the that that solution for that but
00:41:39
what i'm thinking of is danny stuff like um solution and actually ah
00:41:45
thought the participants to have a separate sessions in say a lot but or
00:41:51
uh i mean i just using one several level unconnected one seven having their own stuff like session
00:41:57
uh yes there's actually a solution uh i didn't have that on the slides but it's called jupiter ha
00:42:04
which is basically a multiuser notebook server on
00:42:09
and has all kinds of integrations i'm still
00:42:14
you can but what what it basically does is it advance it's
00:42:18
all in a notebook server for uh for every user so they're isolated
00:42:24
and we also have a authentication um likeable
00:42:28
affected authentication and you can on that in
00:42:33
and in in the clouds um could native son and things
00:42:36
like set that so this is this might be useful solution
00:42:41
uh_huh uh so sorry so um this is something that can easily easily done so you just
00:42:46
plug kalman into this year how how yeah so this is this is basically separate because um
00:42:53
um yeah it runs the uh to jupiter server and
00:42:57
of course in that in that server you'd have to
00:43:02
you'd have to prepare a provide um almond up
00:43:05
front so you'd have to to install that there perhaps
00:43:08
provided doctor an inch or so depending on your setup with the formant installed already yeah okay awesome thank you
00:43:16
uh_huh but
00:43:19
it's an presently for that or i'm still questions first of
00:43:24
all if could you just used to using the presentation because
00:43:28
it's kind of interesting to see oh yeah um so the black and it's called a rise um
00:43:36
just go back
00:43:39
mm
00:43:42
yeah that's
00:43:48
uh_huh
00:44:07
and then when the slide was but it's called rice uh i think if you should just um
00:44:14
if you just cool it don't lose with the sun rises in the nose and we we should
00:44:22
because he's far is there some way of t. p. g.
00:44:26
'cause she did it continues you that you just to do the sparkly
00:44:33
let's stick to the u. k. e. um uh years to
00:44:38
some points of you mean if you're for instance a display a
00:44:43
table uh data frame or something like that and and then
00:44:48
you wanna page instead of having all all the data or um
00:44:53
yeah there's no oh no built in support so this but this is something
00:45:00
you you you you could create a helper function that actually um
00:45:05
allows you to to to page two things but do it depends a little bit on how
00:45:12
i'm the user experience right so if you want something where
00:45:16
you can actually paid uh the paging through a button in uh
00:45:21
like having having a and an interactive table waken page two things
00:45:25
then you'd need to to build a a which it that allows the
00:45:28
so this is a little bit more involved in the twins couldn't
00:45:31
has its dues widgets okay uh so this not not available now and
00:45:36
distill it will you recently that it should use
00:45:40
really to load 'cause of the the the trendy to
00:45:43
you could a recluse like resume look mushy right
00:45:48
yeah so if you publish it locally yeah it shot
00:45:52
i think it should pick that yeah i to pick them up so if you if you just published this uh
00:45:58
if you do and uh s. p. t. for instance to publish
00:46:01
local problems that into a local i. b. ripple that it'll just work
00:46:07
uh which is the screen you are one more question
00:46:16
if
00:46:19
um first super quick question begins to twelve when
00:46:26
okay i actually is a very good question that i actually uh that uh
00:46:34
talk to uh to h. the men maintainer of
00:46:37
this um gives the day at the the community painter
00:46:43
and he promised to do it soon but okay
00:46:46
tomorrow at uh at tried ah start yes tight on
00:46:52
so many shows on it and then shall not job like you know call python so we can basically use
00:46:58
ah all tight and stuff and then destroy a um yeah so
00:47:06
there is no uh no native support for that at the moment but there's this um library
00:47:14
from what's it called high i i forgot the name but there's this library are independent of
00:47:21
of of notebooks we can actually call a whole private firms colossal but might be an option
00:47:31
that one
Conference Program
A Tour of Scala 3
Martin Odersky, Professor EPFL, Co-founder Lightbend
June 11, 2019 · 5:15 p.m.
8340 views
A story of unification: from Apache Spark to MLflow
Reynold Xin, Databricks
June 12, 2019 · 9:15 a.m.
1271 views
In Types We Trust
Bill Venners, Artima, Inc
June 12, 2019 · 10:15 a.m.
1569 views
Creating Native iOS and Android Apps in Scala without tears
Zahari Dichev, Bullet.io
June 12, 2019 · 10:16 a.m.
2233 views
Techniques for Teaching Scala
Noel Welsh, Inner Product and Underscore
June 12, 2019 · 10:17 a.m.
1296 views
Future-proofing Scala: the TASTY intermediate representation
Guillaume Martres, student at EPFL
June 12, 2019 · 10:18 a.m.
1159 views
Metals: rich code editing for Scala in VS Code, Vim, Emacs and beyond
Ólafur Páll Geirsson, Scala Center
June 12, 2019 · 11:15 a.m.
4697 views
Akka Streams to the Extreme
Heiko Seeberger, independent consultant
June 12, 2019 · 11:16 a.m.
1552 views
Scala First: Lessons from 3 student generations
Bjorn Regnell, Lund Univ., Sweden.
June 12, 2019 · 11:17 a.m.
577 views
Cellular Automata: How to become an artist with a few lines
Maciej Gorywoda, Wire, Berlin
June 12, 2019 · 11:18 a.m.
386 views
Why Netflix ❤'s Scala for Machine Learning
Jeremy Smith & Aish, Netflix
June 12, 2019 · 12:15 p.m.
5030 views
Massively Parallel Distributed Scala Compilation... And You!
Stu Hood, Twitter
June 12, 2019 · 12:16 p.m.
958 views
Polymorphism in Scala
Petra Bierleutgeb
June 12, 2019 · 12:17 p.m.
1113 views
sbt core concepts
Eugene Yokota, Scala Team at Lightbend
June 12, 2019 · 12:18 p.m.
1656 views
Double your performance: Scala's missing optimizing compiler
Li Haoyi, author Ammonite, Mill, FastParse, uPickle, and many more.
June 12, 2019 · 2:30 p.m.
839 views
Making Our Future Better
Viktor Klang, Lightbend
June 12, 2019 · 2:31 p.m.
1682 views
Testing in the postapocalyptic future
Daniel Westheide, INNOQ
June 12, 2019 · 2:32 p.m.
498 views
Context Buddy: the tool that knows your code better than you
Krzysztof Romanowski, sphere.it conference
June 12, 2019 · 2:33 p.m.
394 views
The Shape(less) of Type Class Derivation in Scala 3
Miles Sabin, Underscore Consulting
June 12, 2019 · 3:30 p.m.
2321 views
Refactor all the things!
Daniela Sfregola, organizer of the London Scala User Group meetup
June 12, 2019 · 3:31 p.m.
514 views
Integrating Developer Experiences - Build Server Protocol
Justin Kaeser, IntelliJ Scala
June 12, 2019 · 3:32 p.m.
551 views
Managing an Akka Cluster on Kubernetes
Markus Jura, MOIA
June 12, 2019 · 3:33 p.m.
736 views
Serverless Scala - Functions as SuperDuperMicroServices
Josh Suereth, Donna Malayeri & James Ward, Author of Scala In Depth; Google ; Google
June 12, 2019 · 4:45 p.m.
936 views
How are we going to migrate to Scala 3.0, aka Dotty?
Lukas Rytz, Lightbend
June 12, 2019 · 4:46 p.m.
710 views
Concurrent programming in 2019: Akka, Monix or ZIO?
Adam Warski, co-founders of SoftwareMill
June 12, 2019 · 4:47 p.m.
1975 views
ScalaJS and Typescript: an unlikely romance
Jeremy Hughes, Lightbend
June 12, 2019 · 4:48 p.m.
1377 views
Pure Functional Database Programming‚ without JDBC
Rob Norris
June 12, 2019 · 5:45 p.m.
6389 views
Why you need to be reviewing open source code
Gris Cuevas Zambrano & Holden Karau, Google Cloud;
June 12, 2019 · 5:46 p.m.
484 views
Develop seamless web services with Mu
Oli Makhasoeva, 47 Degrees
June 12, 2019 · 5:47 p.m.
785 views
Implementing the Scala 2.13 collections
Stefan Zeiger, Lightbend
June 12, 2019 · 5:48 p.m.
811 views
Introduction to day 2
June 13, 2019 · 9:10 a.m.
250 views
Sustaining open source digital infrastructure
Bogdan Vasilescu, Assistant Professor at Carnegie Mellon University's School of Computer Science, USA
June 13, 2019 · 9:16 a.m.
376 views
Building a Better Scala Community
Kelley Robinson, Developer Evangelist at Twilio
June 13, 2019 · 10:15 a.m.
245 views
Run Scala Faster with GraalVM on any Platform
Vojin Jovanovic, Oracle
June 13, 2019 · 10:16 a.m.
1342 views
ScalaClean - full program static analysis at scale
Rory Graves
June 13, 2019 · 10:17 a.m.
463 views
Flare & Lantern: Accelerators for Spark and Deep Learning
Tiark Rompf, Assistant Professor at Purdue University
June 13, 2019 · 10:18 a.m.
380 views
Metaprogramming in Dotty
Nicolas Stucki, Ph.D. student at LAMP
June 13, 2019 · 11:15 a.m.
1250 views
Fast, Simple Concurrency with Scala Native
Richard Whaling, data engineer based in Chicago
June 13, 2019 · 11:16 a.m.
624 views
Pick your number type with Spire
Denis Rosset, postdoctoral researcher at Perimeter Institute
June 13, 2019 · 11:17 a.m.
245 views
Scala.js and WebAssembly, a tale of the dangers of the sea
Sébastien Doeraene, Executive director of the Scala Center
June 13, 2019 · 11:18 a.m.
662 views
Performance tuning Twitter services with Graal and ML
Chris Thalinger, Twitter
June 13, 2019 · 12:15 p.m.
2003 views
Supporting the Scala Ecosystem: Stories from the Line
Justin Pihony, Lightbend
June 13, 2019 · 12:16 p.m.
163 views
Compiling to preserve our privacy
Manohar Jonnalagedda and Jakob Odersky, Inpher
June 13, 2019 · 12:17 p.m.
302 views
Building Scala with Bazel
Natan Silnitsky, wix.com
June 13, 2019 · 12:18 p.m.
565 views
Diversity and Inclusion Panel: Bring the Thunder!
Panel
June 13, 2019 · 2:30 p.m.
245 views
Asynchronous streams in direct style with and without macros
Philipp Haller, KTH Royal Institute of Technology in Stockholm
June 13, 2019 · 3:45 p.m.
304 views
Interactive Computing with Jupyter and Almond
Sören Brunk, USU Software AG
June 13, 2019 · 3:46 p.m.
682 views
Scala best practices I wish someone'd told me about
Nicolas Rinaudo, CTO of Besedo
June 13, 2019 · 3:47 p.m.
2713 views
High performance Privacy By Design using Matryoshka & Spark
Wiem Zine El Abidine and Olivier Girardot, Scala Backend Developer at MOIA / co-founder of Lateral Thoughts
June 13, 2019 · 3:48 p.m.
754 views
Immutable Sequential Maps – Keeping order while hashed
Odd Möller
June 13, 2019 · 4:45 p.m.
278 views
All the fancy things flexible dependency management can do
Alexandre Archambault, engineer at the Scala Center
June 13, 2019 · 4:46 p.m.
390 views
ScalaWebTest - integration testing made easy
Dani Rey, Unic AG
June 13, 2019 · 4:47 p.m.
468 views
Mellite: An Integrated Development Environment for Sound
Hanns Holger Rutz, Institute of Electronic Music and Acoustics (IEM), Graz
June 13, 2019 · 4:48 p.m.
213 views
Closing panel
Panel
June 13, 2019 · 5:54 p.m.
400 views
Recommended talks
Keynote: Some Mistakes We Made When Designing Implicits (And Some Things We Got Right)
Martin Odersky, Professor EPFL, Chairman Typesafe
June 14, 2019 · 9:13 a.m.
303 views