►
From YouTube: Jupyter Community Call - August 2019
Description
Recording from Jupyter Community Call in August 2019.
The notes from this call can be found here:
https://jupyter.readthedocs.io/en/latest/community/community-call-notes/2019-august.html
Read more about these calls on Discourse:
https://discourse.jupyter.org/t/jupyter-community-calls/668
A
Right,
okay,
everybody
welcome
to
our
Jupiter
community
call
in
the
month
of
August.
Thank
you,
everyone
for
coming
great
turnout.
These
are
a
lot
of
fun,
which
I
do
these
every
month,
usually
the
last
Tuesday
of
the
month
for
future
note,
and
then
we
try
to
keep
you
updated
either
on
the
distress
Channel
or
on
the
Jupiter
mailing
list
or
Twitter
account,
so
keep
an
eye
out
for
that.
A
So
you
have.
The
format
of
these
are
kind
of
always
evolving.
The
goal
in
some
sense,
is
to
create
kind
of
a
space
like
a
virtual
Jupiter
con,
especially
since
we
don't
have
a
Jupiter
con
this
year
that
we
can
use
for
anyone
to
come
in
and
present
a
short
demonstration
of.
Maybe
something
they've
been
working
on
that
you
put
our
ecosystem.
It
doesn't
necessarily
have
to
be
directly
Jupiter
development.
It
could
be
just
something
cool
you're
doing
with
Jupiter,
and
so
this
is
really
open
to
anyone
and
everyone
to
share.
A
If
we
don't
get
through
all
of
them
before
the
end
of
the
hour,
then
we
can
either
ask
that
person
to
meet
next
week
month
or
keep
them
on
the
call
and
people
it's
free
to
leave
cool.
So
the
one
thing
I
really
wanted
to
point
out
in
the
short
reports
was
tracking
Jupiter
newsletter.
For
this
month
is
out.
You
should
go
check
that
out.
A
That's
enough
awesome
resource
put
out
by
a
Tony
Hearst
and
others
that
provide
a
lot
of
things
that
are
happening
in
Jupiter,
that
as
a
core
dev
of
Jupiter
I'm,
not
even
aware
of,
and
so
it's
really
fun
to
see,
just
new
things
pop
up
there
and
actually
a
few
of
the
things
we
have
down
in
our
agenda
items
get
shoutouts
there.
So
Tony
comes
across
some
amazing
stuff.
I,
don't
know
how
he
does
all
that
research,
but
definitely
check
that
out.
The
link
is
actually
in
the
back
MD
cool.
A
B
A
Okay,
with
that
I
think
we'll
go
into
our
agenda
items,
so
these
are
a
yen
demos.
So
goal
is
hopefully
that
the
people
showing
off
these
demos
can
share
their
screen
and
give
a
live
demonstration
of
what
they've
built
for
what
they're
working
on.
So
we'll
start
with
the
lift
to
demonstrations
from
the
Cal
Poly
intern
teams
for
the
summer,
and
so
we'll
start
with
that.
The
first
one
we
have
on
the
list.
If
they're
ready
is
the
rich
text,
Juba
lab
extension,
guys.
E
D
D
You
can
see,
there's
no
longer
reporters
on
the
cells
and
when
you
just
start
typing
instead
of
having
to
run
the
cell,
you
just
see
rich
text
right
away,
so
we
can
bold
italic,
strikethrough
and
all
those
things
we've
worked
really
hard
to
preserve
existing
markdown
features.
This
isn't
hiding
anything.
You
could
have
already
done
this
markdown,
but
instead
that
you
see
what
you're
typing
right
away
and
so
thing
to
run
and
then
go
back
and
edit
and
deal
with
that
experience.
C
E
E
F
E
D
D
We
have
those
in
line
and
code
blocks
like
currently
in
line
just
as
the
styling
we're
yeah
code
blocks.
You
can
still
define
the
language
through
the
interface
now,
which
is
nice,
because
it
will
also
tell
you
if
the
language
isn't
something
that
code
mayor
supports,
because
it
has
a
fuzzy
searching.
E
H
F
D
E
G
G
E
I
B
G
G
F
E
G
B
E
G
D
A
J
E
J
E
A
J
H
J
E
E
A
The
that
is
available
you
can
right
now
we
don't
have
it
actually
on
NPM
yet,
but
you
can
check
out
the
link
that
we
have
in
the
hack,
add
or
hack
MD.
Sorry
that
you
can
install
from
source
directly
will
have
releases
of
that
probably
this
week.
So
you
can
start
using
those
really
soon
all
right,
cool,
we'll
go
to
our
next
thing
on
the
list
and
that's
the
also
from
the
Jupiter
Cal
Poly
intern
team.
Marissa
is
going
to
lead
us
off
here,
and
this
is
the
Python
package
installer
to
the
lab
extension.
K
You
can
yo
see
my
screen:
okay,
cool
I'll,
just
show
you
a
basic
usage
and
then
I'll
do
a
few
more
things.
So
mine
is
the
package
installer.
It
installs
pipe
I
packaged
--is
into
your
current
active
colonel
here,
be
app.
It
so
I'm,
just
gonna
import
them
on
well.
I,
don't
have
yet
and
I
get
the
error,
and
then
you
can
go
ahead
and
use
the
input
box
to
install
that
package,
and
you
can
see
the
logs
in
case.
K
Anything
went
wrong,
but
now
we
can
rerun
it
and
you
can
see
that
it's
imported
it
here
and
this,
as
I
said
before,
is
installed
in
your
current
environment.
So
if
you
change
it,
that's
reflected
and
then
also
right
now,
I'm
working
on
a
feature
that
recognizes
the
package
you
don't
already
have
so
when
I
first
ran
that
so
and
I
didn't
have
names.
G
Yeah
I'd
love
to
be
able
to
Pippen
stall
EEE
some,
my
directory,
because
that's
the
best
practice
that
would
be
help
for
other
people
to
use
I
think.
K
K
L
G
H
K
K
K
A
Oh
yeah,
the
link
to
the
actual
package
is
also
an
act
in
D
file,
so
awesome
cool,
so
well.
Thank
You
Marisa
for
presenting
we'll
go
ahead
and
move
to
the
next
item
on
the
demo
list.
This
one's
from
Tony
sacks.
This
is
going
from
a
just
to
get
er.
Are
you
on
the
yeah
you're
still
on
the
call
Tony
you
want
to
leave
this
off.
A
G
G
Along
so
forth,
my
connection
doesn't
crash
here.
Okay,
so
one
of
the
things
that
I've
been
doing
is
I've
been
putting
notebooks
into
just.
G
My
connection
doesn't
be
all
that
great
okay,
so
one
of
the
things
that
I
do
often
times
is
I
share,
notebooks
and
with
gist
and
I've,
been
trying
to
teach
other
folks
how
to
do
that,
and
you
know
some
of
the
benefits
doing
that.
Also
so
I
want
to
click
and
work
this
so
some
of
them.
So
one
of
the
things
that
people
often
allude
to
is
that
reproducibility
is
is
hard
and
I
think
with
gist
and
for
small
enough
idea.
Does
it
necessarily
have
to
be
so?
G
If
we
look
at
this
notebook
that
Zack
just
shared
with
us,
it's
got
scheme
org
in
it
and
it's
got
pied
antic
in
it,
and
these
are
some
tools
that
I'll
talk
about
what
we
get
this
thing
built,
but
one
of
the
easiest
ways
to
go
and
take
Zach
is
made
and
make
it
reusable
is
to
go
and
do
quirements
that
text
identic
and
I'll
probably
want
some
RTF
lib.
Let's
get
some
pandas
in
here,
and
maybe
some
matplotlib
just
in
case
and
then
the
last
one
that
we
wanted
with
schema.org.
F
G
Right,
fine
and
delivery,
so
we
go
and
take
a
gist,
it's
being
upload
connections,
kind
of
stinky
right
now,
unfortunately,
all
right,
so
we
get
to
oh.
This
is
not
my
binder.
It's
my
binder
org
I
always
do
this.
So
we
go
over
to
my
binder
and
one
of
the
things
that
binder
didn't
have
when
it
first
came
out
was
the
ability
to
use
just
read
those,
and
now
it
does
so
if
I
drop
this
just
in
here
and
then
I
do
gist
and
then
I
hit
launch
and
we
go
and
watch
everything
happened.
G
G
G
The
schema.org
thing
is,
and
then
we
get
the
ammo
so
well.
This
is
going
on.
Yes,
so
if
we
go
and
look
at
my
personal
gist
here,
this
will
be
worth
it
when
it
actually
happens.
Okay,
great
now,
it's
been
defined,
but
it's
and
it
was
put
into
our
build
right,
yep.
What
so
now
it's
running
and
generally,
when
I
do
that
on
the
action,
but
by
choosing
hip
things
get
into
your
things,
build
a
little
bit
quicker
and
sometimes
what
we
might
do
is
we
might
just
spin
up
a
binder.
G
Well,
it's
an
idea.
Sometimes
we
might
intend
to
go
back
and
make
changes
so
for
mister,
be
yours
on
there,
so
that
I
can
go
back
and
reuse
them
later
on
and
I
can
also
share
them
with
other
people
cool,
so
we're
pushing
the
image
it's
really
hard
to
keep
talking
through
a
build.
This
was
NPA
might
have
to
talk
like
ten
times
longer.
What
night
and
there's
back
on
submission
with
doing
this
I
know
that
you're
using
you
just
made
a
commit
to
telemetry
that
uses
identity
right
yeah.
A
Yeah
so
we're
using
it
to
kind
of
explore
validating
schemas
in
a
much
cleaner
way.
It
allows
you
to
define
those
things
like
schemas
in
Python
directly,
so
that
makes
it
helpful
for
both
auto-generating
JSON,
schemas
or
PMO
files
from
that,
but
also
for
testing
the
data
that
you're
bringing
in
that
are
just
that
are
defined
by
some
schema
or
follow
some
schema,
and
so
we've
been
exploring
pedantic
for
telemetry.
A
A
C
So
what
I
wanted
to
talk
about,
there's
a
few
different
extensions
that
we've
been
building
under
the
name
rich
context,
and
so
one
of
these
is
a
data
Explorer
for
Jupiter
lab,
and
the
idea
is
that
that
you
can
often
find
datasets
and
lots
of
different
places
in
Jupiter
lab
like
here.
There's
one
in
a
notebook
that
we
output
I
have
a
s
you
file
on
your
you
know,
file
system.
C
You
might
have
a
sequel
database
that
you
want
to
view
data
from
and
oftentimes,
if
you're
writing
extension
in
Jupiter
lab
to
view
data,
you
want
to
view,
as,
as
you
know,
as
many
different
types
of
data
as
you
can
from
whatever
sources
that
are
available
to
you.
You
don't
really
care
if
this
table
came
from
a
notebook
or
from
a
plot
it
so
I
was
a
punch
line.
C
So
the
data
Explorer
is
is
a
new
you,
a
new
plug
you
can
use,
and
so
here
we
see
we
have
this
notebook
and
if
we
look
at
our
file
system
here
we
can
see
we
have
this
notebook,
this
table
notebook,
and
we
can
look.
Oh
this,
fourth
cell.
Here
we
see
that
this,
this
panda's
data
frame
is
actually
output
as
data
inside
the
notebook.
C
So
we
looked
at
the
notebook
we'd
see
this
is
JSON
in
the
output
because
we
said-
and
one
thing
that
can
view
this
JSON
JSON
is
the
interact
data
Explorer.
So
we
can
click
this
button
and
obviously
we
have
this
interactive
kind
of
Explorer.
For
this
data
set
here
example
of
how
this
interact.
Data
Explorer
doesn't
really
care
where
this
data
came
from.
All
it
cares
about
is
it
can
handle
a
certain
type
of
data,
which
is
this
this
event,
this
data
resource
on
mime
type
I'm.
C
Some
other
examples
here
we
can
actually,
if
we
have
a
link
to
data
like
to
a
CSV
file,
can
also
view
that,
like
in
a
normal
grid,
widget
and
so
the
ideas
we
have,
and
if
we
have
a
like
a
file
on
disk
here,
a
CSV
file,
we
can
also
do
that
except
didn't
save
it
yet.
So
the
ideas
we
have
a
uniform
way
of
talking
about
where
data
is
located
using
URLs
and
the
type
of
data
that
we
want
using
mime
types,
and
then
we
can
define
converters
that
move
between
brookmire.
It's
the.
F
C
Set
so
one
example
of
this
is
Max
has
been
working
on
hdf5
Viewer.
So
if
we
have
here,
we
have
some
hdf5
data
sets
and
he's
implemented,
an
extension
that
hooks
into
the
data
registry
and
allows
you
to
browse
hdf5
data
sets
and
then
look
at
their
grids,
and
so
he
so
he
also
implemented
this
with
a
custom.
C
Comments
manager
that
you
see
over
here
so
when
you
double
click
this
file,
it
opens
up
this
contents
manager
and
you
can
view
the
data
this
way
so
there's
similar
ways
of
tax
in
a
problem,
but
it's
a
bit
cleaner
with
the
data
registry,
because
you
don't
have
to
the
this.
The
current
contents
management
is
pretty
tied
to
the
idea
of
having
file
and
kind
of
file
based
operations,
and
things
like
that.
Where's,
the
data
registry
doesn't
really
care
about
what
your
data
is.
L
C
Nice,
that's
cool,
so
yeah,
so
the
idea
here
is
that
we
can
kind
of
more
easily
if
you're,
writing
an
extension
that
if
you
use
data
you
can
more
easily
plug
it
into
different
sources
by
using
this
view
yeah.
So
that's
the
first
thing
and
then
the
other
one
that
builds
on
that
is
this
link
data
browser,
and
so
this
is
something
we've
been
exploring
and
so
that,
if
you
have
say
a
data
set
on
on
disk
or
in
a
notebook,
you
want
information
about
that.
You
can
so
here
we
see.
C
We
have
this
publication
and
this
publication
is
lots
of
like
metadata
about
it.
Right
have
different
creators,
people
created
it,
we
have
its
publisher
and
if
we
click
on
one
of
these
people,
we
can
see
information
about
them
as
well.
So
this
is
yes
to
the
publication,
and
the
idea
is
that
this
would
be
tied
to
the
data
set.
So
as
you're
browsing,
a
data
set
you'll
see
metadata
about
that
data
set
as
you
browse.
C
So
this
link
data
browser
uses
under
the
covers
just
json-ld
format,
to
describe
your
metadata
I'm,
an
agnostic
to
what
actually,
what
keys
or
what
you
know
what
what
types
you're
talking
about.
But
it
just
allows
you
to
specify
metadata
about
resources
that
you
care
about
juju
lab
yeah,
so
it's
as
you
can
tell
they.
The
UI
for
both
of
these
is
still
pretty
rough,
and
so
we're
still
in
progress
there.
But
if
you
have
any
places,
you
think
that
this
might
be
useful.
C
A
I
C
No,
it
doesn't,
it
doesn't
disappear,
it's
tied
to
the
notebook
model.
Actually,
so
it
basically
went
yeah,
so
it
when
you
view
a
notebook,
it
basically
says:
hey.
Do
we
already
have
a
notebook
model
for
this?
If
so,
use
that
one
you
know
otherwise
open
it.
Just
like
this,
it's
the
same
logic
when
you
open
a
notebook,
it
says:
do
I
already
have
a
model,
if
not
then
create
a
new
one
for
this
notebook.
C
C
C
C
I
C
C
Let's
find
that
here
it
is
yeah.
This
is
it
here.
So
we
see
we
have
this
data
resource
and
we
see
these
are
the
dick
it's
for
it,
and
so
we
see
one
of
them.
Is
this
mind
data?
You
see
that
this
is
the
actual
data
yeah,
so
they
and
all
of
these
these
objects
are
actually
a
reserved
rebels
so
which
are
like
you
know,
they're,
like
signals
and,
and
you
fill
up,
but
they
have,
they
have
a
way
to
say
like
if
you
unsubscribe.
Basically,
you
have
a
clean
up.
C
I
A
A
B
G
A
G
You
mind
if
I
finish
up
real
quick,
yeah
yeah
you're
back
Tony,
oh
cool!
Thank
you!
Sorry,
okay!
This
is
where
we
get
to
the
play,
sometimes
sometimes
playing
there's
best
here
some
days
than
other
days.
Okay
made
this
binder
and
one
of
the
things
fun
to
do
when
you
get
your
boss
is
if
you've
got
friends.
If
you
have
friends
so
here,
we've
got
this
little
hash.
Can
you
guys
see
the
status
bar
I?
G
Don't
know
if
that's
a
thing,
that's
visible,
but
anyway,
if
you
take
one
of
these
little
hashes
and
then
you
ask
or
you
look
for
the
token
and
then
that
hash
you'll
get
a
token
back
so
now,
if
I
share
that
in
the
chat
screen.
Okay,
so
I'm
gonna
share
that
in
the
chat
now,
so
everybody
will
would
be
working
off
of
the
same
running,
running
binder,
which
I
think
is
kind
of
funny,
and
then
what
we
can
do
is
we
can
sort
of
go
through
the
binder
together
and
talk
about
what
it
does.
G
I
know:
I
stopped
sharing
my
screen.
Let
me
share
that
again,
and
so
you
know
you
guys
can
do
malicious
things
and
please
choose
to
if
you
do
desire,
but
we're
going
to
go
through
Zach's
code,
so
I'm
running
it
now
and
all
your
kernels
are
updating
at
the
same
time
that
this
is
and
we
have
a
we
have
some
errors
happening
here:
app
ID
antics
not
defined
okay.
G
So
that
looks
like
something
like
this
all
right,
so
Zach's
notebook
does
not
restart
and
run
all,
which
is
a
good
goal
for
a
little
tiny
ideas,
but
basically
by
using
these
Pied
antic
based
models.
Here
we
can
start
to
define
Jason
in
pure
Python,
which
is
super
cool
and
the
back
is
doing
is
zach
is
doing
this
specifically
for
schema.org,
which
is
a
semantic
place.
This
one
doesn't
I,
don't
know
what
it's
doing
well,
I
will
keep
going.
You
know
whenever
somebody
sends
me
one
of
these
things.
It's
like.
G
G
Well,
well,
you
know
that's
a
good
example
of
a
test
and
a
little
bit
informally
hey.
Where
are
we
supposed
to
get
that
error
here?
Is
that
it
was
at
the
validation
error
right
here.
Let's
just
say
that
this
were
it.
What
we
might
want
to
say
here
is,
we
might
want
to
say
import,
JSON
schema
and
then
accept
Jason
scheme
or
something
like
that.
I
put
an
assertion,
or
you
know
some
kind
of
you
could
use
a
test
in
here.
You
can
actually
import.
G
Putting
two
things
together
is
a
really
reasonable
place
to
start
up
in
a
small
notebook
and
we're
not
trying
to
write
object-oriented
programming,
we're
trying
to
get
use
high-level
packages
and
put
them
together
and
create
neat
things.
So,
thanks
for
sharing
Zack
and
letting
me
critique
this
a
little
bit,
thank
you
Tony.
Any.
H
G
You
are
so
yeah
I
mean
I,
I,
don't
know
it's
nice
I
I,
don't
know
is
Tim
on
the
fall
or
Chris.
This
is
not
a
good
security
practice.
I
just
think
it's
funny
and
please
don't
ever
put
pure
detail,
but
basically,
if
you
search
for
token
in
the
status
bar
you'll
get
bunch
of
finders
that
do
would
have
opened
up
and
then,
if
you
find
the
specific
hash
to
the
machine
that
you
opened
up
originally,
then
you
can
share
this
token.
G
A
Awesome,
thank
you,
Tony,
we'll
with
that.
We
can
end
this
call
and
let
anybody
has
anything
they
want
to
share
last
minute:
go
ahead
and
type
in
if
not
I'll
go
ahead
and
stop
the
recording,
we'll
leave.
The
call
out
feel
free
to
chat
and
hang
out
on
the
call
as
long
as
you
like
and
then
yeah
with
that,
we'll
have
our
next
community
call
in
a
month.
So
all
right
thanks,
everybody
I'm
gonna,
stop
the
recording.