►
From YouTube: Jupyter Community Call - August 31, 2021
Description
Recording from the Jupyter Community Call in August 2021.
The notes from this call can be found here: https://jupyter.readthedocs.io/en/latest/community/community-call-notes/2021-august.html
Read more about these calls on Discourse: https://discourse.jupyter.org/t/jupyter-community-calls/668
A
Bite
from
nothing
cool
federic
did
you
have
you
had
a
good
experience
with
the
program?
Also.
B
A
Indeed,
indeed,
but
you've
been
getting
a
lot
done
this
these
past
couple
months,
so
it's
pretty
impressive.
I
look
forward
to
hearing
some
of
your
feedback
on
the
program
later
on
we're
all
interested
in
these
newcomer
programs
and
how
to
bring
people
into
the
community
all
right.
Let's
are
there
any
other.
Did
anybody
else
like
to
introduce
themselves
say
hi?
If
not,
we
can
continue
along.
A
Alrighty,
it
looks
like
the
security
crew
and
the
hpc
crew
is
about
to
tell
us
some
things
roland.
You
want
to
step
up
and
let
us
know
what
you
all
have
going
on.
C
Sure
I
just
have
a
couple
short
announcements
that
I
wanted
to
share
with
everybody.
Both
of
them
relate
to
just
upcoming
meetings
and
trying
to
get
the
word
out
about
these
meetings.
The
first
one
is
that
there
is
a
monthly
jupiter
hub
and
high
performance
computing
meeting
that
we
hold.
I've
got
the
link
to
the
agenda
there
in
the
minutes
for
this
meeting
and
there's
also
a
link
there
to
the
to
the
project
community
calendar.
C
C
We
have
bi-weekly
meetings
around
jupiter
security
topics,
not
for
discussing
specific
vulnerabilities,
but
around
talking
about
general
issues
around
securing
jupiter
deployments.
Those
are
bi-weekly
on
fridays
at
9
00
a.m,
and
the
next
call
is
going
to
be
on
the
10th.
So
this
friday
we're
off.
But
next
friday,
we're
back
on
and
the
topic
of
that
meeting
is
going
to
include
one
of
the
topics
for
that
meeting
is
going
to
be
the
trusted
ci
engagement
that
we
have
with
trusted.
C
Ci
they've
done
an
initial
review
of
all
of
the
security
documentation
across
all
of
the
jupiter
subprojects,
and
so
we're
going
to
go
over
that
at
that
meeting,
and
I
think
that
should
be
very
interesting.
A
very
interesting
topic.
D
Very
cool,
are
we
gonna
have
formal
cve
processes
in
the
future?
We
don't
have
them
now.
E
Rick
yeah
we're
trying
to
cover
two
things.
One
is
like
what
the
hell
the
trusted
ci
is
what's
outward
facing
to
the
users
and
deployers,
and
then
we
just
kind
of
want
to
start
this
conversation
like
hey.
How
do
we
manage
credentials
to
the
github
orgs,
who
manages
the
ssl
search
for
jupiter
oregon?
E
A
Very
cool
I'm
excited
for
that
work.
I
know
matthias
from
ipython
has
been
pretty
aggressive,
with
filling
out
some
of
these
things
too.
So
hopefully
he
gets
in
touch
with
you
all
soon.
Also.
E
A
E
A
Right
right
I
mean
I've,
it's
been
like
the
last
five
or
seven
years.
I
mean.
Isn't
the
blip
in
marvel
like
a
real
thing
or
is
it?
Is
that
what
we
just
lived
through,
I'm
not
sure
anyway.
Are
there
any
more
announcements,
so
rick
and
roland
said
their
piece?
Does
anybody
have
any
recent
releases
any
celebrations,
any
kind
words
they'd
like
to
say
before
we
get
into
the
agenda
items.
F
Every
once
in
a
while,
it's
like
a
reverse
werewolf.
So
yes,
I
I
normally
have
a
beard
down
to
here
and
hair
down
to
here,
but
I
don't
today
so
and
I
won't
for
a
while.
I
guess
there's
been
we're
starting
the
ramp
up
for
python
3.10.
F
A
lot
of
our
old
friends
like
noes
are
not
coming
along
for
the
ride
and
so
be
prepared
for
a
little
bit
of
rocky
stuff.
We're
trying
to
get
some
releases
out.
So
this
has
impacted
traitlets.
This,
which
no
longer
requires
a
bunch
of
things.
This
is
impacted,
ipi
kernel,
it's
going
to
impact
ipi
widgets,
so
you
know
if
you
didn't
lock
down
your
environments
like
two
weeks
ago.
F
Well,
you've
got
to
kind
of
have
to
watch
out
for
what's
going
to
be
happening
in
the
next
couple
weeks,
but
if
anybody
recalls
how
the
3.9
rollout
went,
it
was
not
super
smooth
and
so
we're
trying
to
do
a
little
bit
better.
This
time
and
we've
got
more
ci
running
against
more
nightlys
and
rc's
and
stuff
like
that.
So,
but
there
will
be
releases
and
they
will
break
your
stuff.
So
you
know
don't
say
I
didn't
warn
you.
A
A
So
if
anybody,
if
things
break
blame
dick
and
let's
see
we
got
some
updates
from
mike,
you
want
to
tell
us
a
little
bit
how
translation's
going
mike.
I
Yes,
just
very
quickly,
I
want
to
mention
that
troop
the
lab
translation
strings
are
now
all
in
crowding,
which
is
the
system
that
we
are
using
to
collaborate
on
translations,
and
we
would
be
very
happy
if
you
can
contribute
or
spread
the
word
around
many
languages
are
mostly
ready
to
go.
I
A
How
cool,
how
cool
that's
very
exciting,
to
hear
what
languages
do
we
have
good
coverage
on
at
this
point?
Are
there
quite
a
few.
I
Certainly
chinese
is
fully
completed
and
french
is
fully
completed.
As
steve
just
said,
in
the
chat-
and
I
guess
polish
has
pretty
good
coverage,
there
are
some
languages
which
have
good
coverage,
but
the
translations
are
not
confirmed
because
we
have
to
change
context
of
some
strings.
Those
were
unmarked
and
needs
just
someone
to
press
a
tick
and
confirm
whether
this
is
still
a
good
translation.
A
A
J
This
I
hate
how
like
the
zoom
screen,
sharing
toolbar
overlaps
with
my
chrome
tab,
okay
anyways,
I
was
supposed
to
have
you
know,
fancy
presentation.
This
was
for
the
fellowship,
but
I
guess
for
this
call
I
can
be
a
little
more
casual.
The
title
of
the
presentation
was
getting
good
at
jupiter
labbing.
This
is
me,
this
is
my
partner
who
is
not
here
at
the
moment
he
helped
out.
J
His
name
is
cena,
and
this
is
frederick,
of
course,
past
nine
weeks
we
spent
contributing
like
features
and
tackling
major
issues
in
the
jupiter
lab
git
extension.
So,
although
it
wasn't
the
you
know,
the
whole
editor
code
base.
At
least
we
got
a
very
vital
extension.
That's
used
by
a
lot
of
people
in
a
lot
of
teams,
so
yeah
frederick
was
very
vital.
He
show
he
taught
tell
me
a
lot
of
stuff
about.
You
know
how
open
source
works
and
how
jupiter
lab
itself
works.
J
J
So,
in
the
past
nine
weeks,
my
if
this
could
load
yeah,
we
contributed
to
five
pull
requests.
Two
of
them
were
just
bug
fixes
to
get
our
hands.
You
know
dirty
on
the
code
base
and
the
other
three
which
I'll
be
demoing
today
are
features
so
without
further
ado.
This
is.
J
One
git
commit
amend
so
what
this
does
it
just
plops
your
stage
changes
onto
the
you
know
the
last
committed
commit
before
in
jupiter
lab.
You
didn't
have
a
way
to
do
this,
so
you
had
to
do
it
to
the
terminal
and
if
you
weren't,
terminal,
savvy
or
just
didn't
know
about
get
well,
this
would
be
frustrating
and
hard.
So
now
we
bring
the
ability
into
jupiter
lab
by
adding
a
a
button
that
says,
amend
next
one
I
want
to
talk
about
before.
J
I
show
this
demo
for
these
two
is
the
git
history
for
a
selected
file.
So
in
jupyter
lab
you
before,
you
would
only
see
the
get
history
for
the
entire
repository
and,
if
you're
someone
who's
like
just
a
data
analyst
working
in
a
large
code
base
that
only
worries
about
one
notebook.
You
really
don't
care
about
everyone
else's
history
and
just
wants
to
come
in
history
for
that
one
file.
J
So
we
bring
that
into
jupiter
lab
by
simply
right,
clicking
onto
a
context
menu
and
it's
nice
and
pretty
so
I'll
demo
that
now
first
it's
this
one.
So
this
is
actually
the
code
base
for
jupiter
lab.
Get
because
I
don't
have
any
other
code
bases
right
now
with
a
lot
of
commit
history.
So
let's
start
with
this:
the
single
file
history.
So
if
we
go
to
the
get
tab
and
then
you
click
history,
nothing
new
here,
it
shows
the
entire
history
for
the
repository,
fine
and
dandy.
J
And
press
save
now
we
go
back
to
the
git
panel.
We
see
so
now,
if
I
right
click
and
then
press
history.
I
could
also
do
this
through
just
the
regular
file
browser
get
history.
J
Now
we
see
this
nice
pretty
div.
That
says.
Oh,
this
is
the
file
history
for
just
the
readme,
and
it
shows
only
the
commits
for
this
file.
Something
too
hard.
You
press
on
it
it'll
show
the
diff
as
normal,
but
this
node,
this
uncommitted
changes
is
also
new,
so
there
wasn't
really
a
way
to
know
hey.
This
is
the
stuff
you've
just
typed
in
jupiter
lab.
So
this
is
also
new.
J
J
And
you
press
commit,
so
that's
fine,
but
now
let
me
revert
these
changes
now
staged
now,
when
you
press
these
three
dots,
there's
the
regular
create
a
new
commit,
but
you
can
also
amend
one
so
when
you
press
that
the
commit
message
in
the
commit
description
gets
disabled
and
then
we
press
commit,
but
that
wasn't
supposed
to
be
there.
Okay
anyways,
it
should
commit
normally
and
then
it
would
be
appended
to
this.
You
know
this
brand
new
commit
moving
along
to
the
very
last
one.
I
think
this
try.
A
J
Yeah,
this
worked
in
my
regular
demo.
I
don't
know.
A
A
B
A
J
J
J
Are
very
opinionated
on
that
statement,
so
I
don't
want
to.
You
know,
move
on
with
that,
but
this
is
the
last
one
I
want
to
talk
about,
and
this
was
the
most
important
contribution
I
made.
It
brings
merge
conflict,
preview
and
resolution
into
the
editor.
So
you
know
if
you're
a
developer-
and
you
know
like
about
version
control,
you
know
that
merge
conflicts
are
like
one
of
the
biggest
pain
in
the
asses.
J
J
So
when,
like
data,
analysts
or
data
scientists
see
a
merge
conflict,
it's
like
five
times
the
hell
that
a
developer
would
face
when
they
get
a
merge
conflict.
So
what
this
feature
brings
is
you'll
get
a
new
conflicted
tab
which
I
will
show
after
that
says
which
files
are
conflicted
and
you
can
preview
and
resolve
notebooks
and
text
files
within
the
editor.
J
So
without
having
to
leave
you
the
editor
or
just
enter
a
terminal,
so
this
is
a
example
repository
I
made
for
this
demo
and
we
have
a
conflicted
file,
conflicted
text
file
and
a
conflicted
notebook.
So
if
I
were
to
you
know
diff
the
text
file,
we
now
get
a
triple
diff,
which
is
brand
new
to
the
editor.
So
we
can,
you
know,
do
whatever
we
can
revert
this
change
and
you
know
edit
stuff,
whatever
we
want
and
just
to
save
the
file
as
resolved.
J
You
just
press
mark
as
resolved,
and
now
it's
in
your
staging
area.
That's
wanted.
So
now,
let's
look
at
the
notebook
which
is
more
used.
I
gotta
be
honest.
I
don't
have
much
experience
with
jupiter
notebooks,
so
I'm
not
going
to
like
modify
too
much,
but
what
I
do
know
is
that
if
I
press
mark
as
resolved
it's
going
to
work,
fine,
so
there's
you
can
see,
there's
different
outputs
have
changed.
The
cell
has
changed.
J
So
if
I
were
to
press
mark
as
resolved,
this
would
hop
into
our
staging
area,
and
then
I
could
do
you
know
whatever
I
want
and
then
commit
and
that's
about
it.
The
last
slide
was
really
for
everyone
else.
It
was
just
to
show
that
hey,
open
source
is
really
cool
and
if
you're,
a
large
company
that
hasn't
delved
into
open
source
sponsorship
and
stuff
like
that,
now
is
your
chance
to
open
source
your
heart
out
so
yeah.
That's
basically
my
presentation.
A
B
J
Well
so,
thanks
to
nbdime
like
it
resolves
as
much
as
it
can
on
its
own
and
for
the
example
I
had
like
it
was
able
to
resolve.
You
know
automatically
without
me
having
to
make
any
changes.
Suppose
it
was
an
even
larger
notebook
with
you
know
something
that
needs
to
be
manually
resolved
and
I
press
mark
as
resolve
by
accident.
You
get
a
pop-up
saying,
hey,
you
still
have
conflicts
and
then
the
user
would
have
to
manually
resolve
them.
It
was
just
because
my
example
was
very
simple
that
it
worked
out.
J
Fine,
but
yeah.
B
Okay,
I
understand
thanks
and
for
for
the
people
that
are
looking
for
a
release
luciano,
I
see
you,
so
the
release
has
not
been
done
yet.
For
that
last
feature,
the
other
has
been
released,
but
not
that
one,
because
we
need
a
bit
more
testing,
especially
for
hkz
cases
where
a
file
was
deleted
in
one
branch
and
not
in
the
other
one,
those
kind
of
things.
So
hopefully
we
can.
We
can
do
that
quickly
and
we
can
release
it.
B
In
fact,
I'm
really
interested
by
this,
because
I
write
notebook
for
scientists
here
at
ucrf
and
when
I
make
some
update
and
usually
they
make
their
own
modification
on
their
end
and
then
it's
sometimes
pain
to
you
know,
put
everything
together
and
resolve
the
conflict.
So
that
would
be
very
helpful
to
have
this
feature.
A
Cool
if
this
feature
is
still
in
progress,
I'm
wondering
if
we
could
use
like
ours
and
theirs
for
these
merge
conflicts.
But
aside
from
that,
I'm
excited
that
everybody
else
is
excited
to
land,
to
see
this,
and
thank
you
for
sharing
it's
now,
then
you
said
there
is
another
person
that
helps
you
with
this
is:
were
they
part
of
the
fellowship?
Also.
J
They
were,
they
just
had
to
leave.
You
know
early
due
to
external
reasons,
so
I'm
the
only.
A
So
so
so
you
had
like
not
only
management
mentoring
support,
but
you
had
some
coding
team
support.
Also,
while
you
were
doing
this.
A
Neat,
my
last
question
is
frederick:
how
did
you
all
get
involved
in?
How
did
you
wind
up
getting
involved
in
this
program.
M
B
B
Oh
go
ahead.
The
their
pattern
is
that
companies
pay
them
to
to
finance
students
to
contribute
to
open
source
project,
and
so
one
of
the
company
was
interested
in
that
particular
extension.
And
so
they
took
the
burden
to
contact
maintainer
and
to
be
sure
that
some
people
were
ready
to
help
on
the
open
source
front.
A
Alrighty,
let's
move
on
and
mike,
let's
hear
about
the
citation
manager,
you
know
you're
looking
for
feedback
last
time
or
a
couple
meetings
ago.
Let's
hear
where
you
are.
H
A
I
Okay,
okay,
so
I
will
dive
right
into
the
demonstration
of
citation
manager,
maybe
first
I'll
be
showing
one
thing
which
is
site
to
see.
So
this
is
basically
the
predecessor
and
why
I
have
this
idea
to
create
a
citation
module
for
triple
lab.
So
site2c
was
a
way
to
add
citations
for
book
10
notebook,
and
it
was
very
good
software,
a
very
good
extension.
However,
it
wasn't
really
developed
recently.
No
was
it
supported
to
triple
up
so.
I
So
it
basically
works
in
multiples.
So
I
I
just
switch
this
back
down
and
that's
these
two
patterns,
so
one
is
to
add
a
citation
and
you
can
see
it's
a
lot
of
various
journals.
Various
articles
from
from
journal
articles
and
other
items
which
come
from
the
zotero
citation
manager
and
that's
that's
it
so
basically
renders-
and
the
next
thing
that
you
can
do
is
add
typography.
I
I
The
next
thing
which
it
allows
you
to
do
is
to
change
the
citation
style
and
you
can
just
start
typing
and
see
lots
of
styles
which
are
available
and
the
preview
of
how
any
given
style
would
look
like
oops.
So
I
clicked
on
one
and
now
the
citations
are
just
numbers
and
the
bibliography
changed
as
well.
I
Another
thing
that
I
think
is
cool
is
this,
which
allows
me
to
check
what
applications
I
have
added
to
my
library
without
leaving
jupiter
lab.
So
first
it
displays
the
citations
which
are
already
in
this
document
and
they
are
sorted
by
how
many
times
an
item
is
cited,
but
I
can
also
search
for
publications
which
are
not
currently
cited
and
maybe
open.
Then
wow,
let's
see
if.
I
No,
this
is
only
a
collection
from
my
my
collection,
so
basically
this
is
it.
So
here
is
a
list
of
items
that
I
added
from
my
browser.
I
Through
iron
software
connection,
it
was
synchronized
with
the
software
server,
so
this
is
a
proprietary
but
really
open
source
service,
and
it
allows
me
to
now,
after
synchronizing
this
button,
to
explore
everything
and
also
go
into
more
details,
see
how
how
it
looks
and
then
read
that
yeah
so
moving
on
to
how
it
works.
I
I
Storing
the
citations
metadata
in
the
notebook,
because
if
you
will
share
your
notebook
with
a
collaborator-
and
they
don't
have
any
specific
item
in
their
collection,
they
wouldn't
be
able
to,
for
example,
change
the
citation
style
and
we
render
it
or
could
corrupt
your
collection
of
references.
If
this
information
is
not
attached
to
the
notebook
and
further
because.
I
The
citations
references
are
attached
to
the
cells
in
the
cell
metadata,
you
can
move
cells
around
move
to
another
notebook
and
those
will
still
be
attached
to
the
cell
and
the
extension
will
be
able
to
recover
them
and
match
to
the
appropriate
item.
I
I
I
Story
reference
and
if
you
have
a
citation
there
and
linking
to
a
toy,
then
it
will
be
automatically
imported
when
you
open
this
next
time
and
what's
underneath
what's
driving
this,
so
this
is
basically
using
citation
styles
language,
defining
how
the
various
citation
styles
work
and
it's
shipped
from
the
typical
app
citation
data
package
and
thanks
to
nick
for
suggesting
this
idea
to
distribute
those
separately
which
are
basically
x
ml
files,
but
with
a
different
extension
which
define
how
the
citations
should
behave.
I
So
there
is
several
thousand
of
those
and
then
for
fetching
citations.
Currently,
it's
integrated
with
etc
web
copy.
So
it's
just
sending
plain
http
request
servers,
but
in
theory
this
accession
is
structured
in
a
way
that
any
other
manager
which
offers
a
rest
up
could
be
connected
to
that
and
what
you
need
to.
I
To
connect
this
only
providing
okay,
this
video
expired.
So
I
cannot
show
you
that's
really
well,
because
you
need
to
provide
suitable
apple
key,
which
can
be
read
only
and
generated
after
logging
in
on
their
website,
and
there
are
some
other
settings
like
default
style
or
output
format.
I
So
this
is
basically
it.
The
next
steps
is
really
defining
the
scope.
What
features
are
still
missing
and
what
would
be
useful
to
implement?
So
I
can
structure
the
code
to
its
final
form,
interpolating
any
major,
refactorings
and
then
also
adding
support
to
jupiter
with
deep
deck
and
possibly
extending
the
support
from
some
sort
of
specific
things
like
labors
and
collections
and
hopefully
making
a
release
before
the
end
of
this
year.
A
Very
nice
share
mike.
H
A
All
right,
well
one!
I
had
one
question
for
you
mike:
what
is
it
that
you
study
there's
this
like
weird
software
demo,
mixed
amongst
cancer
and
cigarettes?
So
what
what
are
you
working
on.
I
A
Neat,
it's
surprising,
you
said
that
you're
a
student-
and
it
still
surprises
me
because
you
get
so
much
work
done.
I
have
a
hard
time
remembering
you're
still
in
your
studies.
Good
luck
and
thank
you
for
sharing
that.
Have
you
shared
a
link
to
the
repo
in
the
heck,
md
or
zoom
chat?
A
N
H
N
Okay,
so
just
a
bit
of
context,
so
this
is
ipad
kernel,
so
we
all
know
it.
We
can
write
some
statements
like
this
one.
I
attain
the
value
to
a
and
I
display
the
value
of
a
and
it's
it's
expected.
It's
there's
one
in
it
and
then
I
changed
a
to
two
and
here
it
still
says
a
equals
one
right,
because
it's
surgical,
it
depends
on
the
order
in
which
the
cells
are
executed.
N
So
if
I
re-execute
a
now,
it
prints
two,
and
so
we
are
very
used
to
this
behavior,
but
this
is
not
really
good
for
reproducibility,
because
now
it
depends
on
the
execution
order
of
the
of
the
notebook,
and
this
is
because
I
mean
of
the
semantics
of
this
assignment
operator.
It's
basically
in
python,
just
setting
a
value
to
setting
a
key
value
pair
in
a
dictionary.
N
The
the
assignment
operator
in
python
is
not
very
interesting.
It's
not
it's
not
like
in
c
plus
plus,
for
example,
where
you
can
overload
the
assignment
operator
and
do
something
different
to
the
variable
that
already
holds
this
value
in
python.
We
don't
have
this
possibility,
and
I
see
it's
more
like
a
possibility
to
to
upgrade
this
this
equal,
this
assignment
operator
and
the
best
place
to
do
it
is
in
a
kernel
because
kernels
usually
provide.
N
More
features
than
raw
python,
so
that's
what
I
did
in
in
a
knl
recently,
so
I
launched
this
notebook
with
a
kernel
react.
So
that's
a
special
mode
of
a
kernel.
It's
not
the
default
mode,
because
I
mean
it's
still
very
experimental
and
I
I'm
kind
of
exploring
possibilities
with
it
still.
So,
let's
try
and
see
what
we
can
do
with
that.
N
So
this
thing
should
fail
right
because
b
is
not
defined
yet,
but
in
a
kernel
it
it's
completely
authorized.
Basically,
now
this
assignment
operator
means
that
a
is
b
plus
one,
so
a
a
is
completely
defined.
The
result
of
a
is
not
defined.
So
if
I
show
a
now
well,
it
shows
known
there's
nothing
in
it
because
b
is
not
defined.
So,
let's
look
at
b
same
it's
known
now.
N
If
I
assign
a
value
to
b,
then
b
is
updated
and
a
is
also
updated
because
a
always
is
b
plus
one,
so
that's
reactivity
at
least
the
beginning
of
it,
and
you
can
try
it
out
yeah.
That's
all.
I
have
to
share.
D
A
Any
questions
for
for
david.
N
Yeah,
it
recognized
that
this
thing
didn't
exist
so
actually
b
is
not
when
we
define
b
as
one
here,
it's
not
an
ink
something
more
than
a
mint.
It's
like
an
int
wrapped
into
another
object.
This
thing
is
probably,
I
think,
a
bug.
I
should
probably
also
check
that
c
is
not
defined
at
all,
and
actually
I
don't
know
if
it
shouldn't
be
the
right
behavior,
because
it's
not
even
assigned
yet
so.
N
N
But
yeah
it's
kind
of
a
a
path
towards
observable
notebooks,
because
I
mean
it's
a
very
typical
feature
of
of
observable
notebooks
that
users
seem
to
like.
So,
let's
see
if
we
can
have
something
similar
in
jupiter.
N
Than
so
for
this
to
work,
languages
must
have
something
similar
to
ast
transformation,
because
I
use
that
a
lot,
especially
to.
N
N
Well,
it
yeah
it's
it's
it's
pattern,
only!
It's
not
titanium
unless
you
have
support
for
ast
transformation
or
something
similar
and
just
execution
cool.
A
This
is
pidgey
time.
I
see
you
know
one
day,
one
day
I'll
get
to
release
this.
So
philosophically
the
philosophical
question
here,
so
I've
got
a
pidgey
kernel
actually
before
I
demo
pidgey.
Let
me
demo
jupiter
lab
markup
because,
like
I'm
absolutely
in
love
with
this
jupiter
is
it
jupiter
lab
markup
for
jupiter
markup
nick.
A
All
right,
so,
let's
get
a
mermaid
diagram
here
and
if
you
have
this
chart
install
if
you
have
this
extension
installed,
you
can
go
over
into
your
markdown
cell,
and
not
only
can
you
put
citations,
but
now
you
can
put
little
diagrams
and
there's
some
other
fun
stuff
you
can
put
in
here,
but
now
you
render
it
and
you
get
these
flowcharts.
A
A
If
I
switch
over
to
pidgey-
and
I
run
this
again-
then
this
actually
renders,
as
it
doesn't
run
as
code.
So
if
I
tab
in
now
it's
code
and
it
says,
b
doesn't
exist.
A
So
if
I
was
to
go
and
put
a
here
and
b
is
these,
are
I
basically
I'm
accepting
markdown
the
source
here?
Oh
boy,
what
a
bad
demo
already
okay!
So
I
get
an
error
that
says
b
is
undefined.
So
now,
if
I
go
in
and
put
b
in
here-
and
then
I
run
this
so
it
says,
b's
three
b
equals
ten
import
pandas
and
then
we
can
say
something
like
b
equals
pandas.util.test.
A
And
the
update
did
not
happen,
but
something's
busted
here.
Nevertheless,
the
idea
is:
oh
html
min
sorry,
pip
install
html
min.
If
you
notice
every
time
I
have
to
tab
these
in
just
so
it
like
runs
as
a
code
thing
and
yeah.
We
have
to
use
html
min
here
just
because
the
thing
gets
wonky
and
then,
if
I
put
it
transpose
here
and
now,
if
I'm
over
here
and
I
keep
updating,
b
that'll
keep
changing,
then
I
can
say
something
like
b
dot
plot
dot
figure.
I
think
this
will
work.
A
Okay,
so
now
that's
there
and
then,
when
I
go
and
run
this
that
updates
and
changes.
So
what
I'd
love
to
talk
to
you
about
david
is
being
able
to
use
this
async
kernel
with
async
ginger
templates
and
see
if
we
can
meet
in
the
middle
somewhere.
A
Opinion
right,
so
what
I'm
thinking
is.
Is
that,
like,
or
at
least
what
I'm
feeling
is,
if
there's
an
output
and
it's
a
default
output
that
just
shows
up?
But
if
you
have
like
these
like
interactive
templates
or
ui,
that
somebody
might
have
made,
that's
a
really
nice
place
where
I
could
like
totally
see
like
immediately
using
some
of
the
async
stuff,
but
making
the
balance
between
the
synchronous
parts
and
the
asynchronous
parts,
like
that's
a
really
hard
discussion
to
have
and
I'd
love
to
get
that
get
that
conversation
going
too.
N
A
A
All
right
does
anybody
have
any
questions
or
comments.
I
suggest
people
try
out
jupiter
lab
markup,
it's
pretty
fun,
there's
mermaid's
a
cool
feature,
and
you
know
if
we
had
these
would
be
fun
to
have
during
our
rtc
calls.
A
We
could
put
diagrams
and
stuff
in
our
notes
and
hackmd
would
be
okay
with
it
too
excellent.
Well,
any
we've
got
five
minutes
left
any
less
minute
comments
or
questions
or
things
you
all
want
to
wind
down
the
time
with.
A
Nope
all
right.
Well,
I
hope
that
all
you
jovians
have
an
absolutely
delightful
week
and
we
will
see
you
again
in
a
month
so
get
your
demos
ready
and
your
release
is
ready
and
let's
keep
bringing
some
fun
demonstrations.
H
Yeah
and
if
you'd
like
to
get
feedback
ever
on
these
calls,
if
we
mess
something
up,
you
can
do
that
with
the
link
there,
and
the
next
call
will
be
on
september
28th.
I
actually
already
have
an
agenda
for
that
too,
which
this
time
will
be
updated
in
the
jupiter
community
calendar,
but
here
it
is
in
case
you're,
so
excited
and
inspired
by
today.
Those
are
just
the
last
two
things
I
want
to
add.
A
Yeah
and
if
anybody
else
wants
to
host,
I
had
a
really
fun
time
doing
this
and
isabella
was
super
supportive
in
doing
it.
So
if
you
all
are
looking
to
become
more
involved
in
your
open
source,
communities
and
running
good
events
and
stuff
is
a
great
place
to
try
it
out
so
contact
isabella
and
let
us
know
if
you
wanna
wanna,
help.
J
H
These
are
the
lurkers
very
nice
tony.
You
can
stop
recording
whenever
I
don't
think
I
can
over.
Oh,
I
guess
I
can
actually.