►
From YouTube: JupyterLab Team Meeting - April 27, 2022
Description
A meeting to share and discuss features, ideas, issues, and pull requests in JupyterLab and other Jupyter frontends. This meeting is open to anyone and everyone.
Join future calls via the Jupyter community calendar: https://docs.jupyter.org/en/latest/co...
Notes for upcoming meetings can be found on the agenda: https://hackmd.io/Y7fBMQPSQ1C08SDGI-fwtg
Past notes can be found on the JupyterLab team compass: https://github.com/jupyterlab/team-compass/issues?q=is%3Aissue+label%3A%22Dev+Meeting+Minutes%22
A
Okay,
welcome
to
the
april
27th
jupiter
lab
weekly
call
please
find
in
the
chat
a
link
to
today's
agenda.
A
A
If
what
you
want
to
discuss
will
take
longer
than
five
minutes,
then
please
add
it
at
the
bottom.
In
the
additional
discussion
section
and
with
that,
why
don't
we
get
started?
We
have
15
people
on
the
call
right
now
it
might
cramp
up
a
bit,
but
that's
pretty
good.
The
first
person
on
the
agenda
today
is
luciano.
B
Hey,
hey,
I'm
gonna
be
quick,
so
I
just
want
to
bring
the
attention.
So
we
from
the
library
community
has
we
have
created
a
proposal
to
incorporate
lyra
into
a
projector
project,
so
just
feel
free
to
take
a
look
at
the
pr
provide
comments,
feedbacks
or
or
anything
and
we're
just
getting
started
with
the
process.
B
A
Cool
thanks
luciano.
I
I
have
a
question
sure.
What
is
the
end
state
for
this
jeff?
Would
it
be
the
elira
repos
ending
up
in
a
separate
jupiter
sub
project,
or
would
it
all
end
up
in
the
jupiter
lapse
of
project.
B
So
that's
a
good
question.
I
think
in
the
it's
a
question
that
a
few
folks
usually
ask
right
away.
So
we
are
trying
to
target
just
make
the
lyra
ai
organization
part
of
jupiter,
so
they
will
be
a
separate
organization
and
and-
and
the
reason
is
mostly
because.
B
The
extensions
that
we
provide
sort
of
like
not
necessarily
work
very
independently.
We
have
like
metadata
services
and
a
couple
of
other
things.
I
think
if,
if
we
were
very
kind
of
like
independently,
I
think
that
the
jupiter
lab
could
be
an
option,
and-
and
that's
why
I
I
didn't
go
into
that-
that
direction
so
for
now,
we
just
want
to
make
lyra
the
way
it
is
today
kind
of
like
coming
to
jupiter.
B
Any
other
thing
I
think
could
be
subject
to
further
discussions
later
on
when
we
see
like
more
integration
more
community
anyway.
Having
said
that,
I
think
we
work
very
closely
with
the
lab
team
and
we
contribute
a
lot
of
those
things
into
the
base.
So.
B
I
don't
see
anything
changing
drastically.
It's
just
kind
of
like
officializing,
something
that
we
internally,
as
the
library
community,
have
been
doing
for
a
while,
just
making
sure
yeah.
We
already
follow
most
of
the
things
from
from
from
jupiter
and
everything
so.
A
A
Go
ahead,
read
more
okay,
yeah
sure,
so
the
name
of
elira
is
the
proposal
that
everything
remains
the
same,
but
it
just
becomes
a
jupiter
subproject,
that's
called
a
lyra.
The
way
say:
zeus
did,
although
actually
zeus
changed
his
name
to
jupiter
zeus,
I
guess,
or
what
are
you
imagining
here,
because
I
know
you
have
packages
published
under
the
scope
at
elyra
and
you
have
an
org
with
some
repositories
and
stuff.
So
what
is
your
thinking
on
that
aspect?.
B
I
think,
to
a
certain
extent
we
are,
we
are
we're
flexible.
I
think
I
was
modeling
this
on
on
voila,
which
they
they
I
believe
they
they
didn't
change
anything.
I
think
we
will
change
kind
of
like
the
aspects
of
like
making
sure
we
make
sure
that
we
are
jupiter.
It's
going
to
become
yeah,
abductor,
liar
and
stuff,
but,
like
I
think,
for
packages
either.
B
A
Okay,
cool:
if
there's
no
other
questions
right
now,
we
can
move
on
to
the
next
agenda
item.
B
I
I
can't
I
can
bring
up
one
issue
that
I
know
that
I
have
a
question,
but
I
think
we
are.
We
address
it
on
the
proposal.
The
only
difference
that
I
see
is
that
when
we
started
we
started
with
the
apache
licensing,
the
goal
is
getting
the
approval.
We
we're
gonna
re-license
and
bst
and
everything
only
jupiter.
So
I
think
that's
the
only
only
item
I
thought
could
get
some
questioning
or
something
like
that
but
like
we
are
aware
of
that,
and
we
have
plans
for
that.
A
Great,
thank
you
luciano
next
on
the
agenda
today
is
carlos.
D
D
D
D
The
problem
is,
there
is
already
an
schema
in
jupiter
lab
the
documents
the
they
have,
the
bl
in
one
place
and
the
model
in
another
place
the
model
it
representing
the
schema,
for
example,
the
document
model
or
the
notebook
model,
so
yeah.
There
is
a
lot
of
duplication
in
that
site
and
I
don't
think,
is
the
right
way
to
go.
A
Cool
thanks
carlos,
so
I
would
like
to
learn
more
about
this
part
of
jupiter
lab
4,
because
it's
one
of
the
key
features.
So
I
think
I'm
going
to
spend
some
time
reading
and
reviewing
this,
but
it's
a
large
pr
with
a
lot
of
work
that
you
put
into
it.
So
hopefully
other
people
on
this
call
and
in
the
community
will
also
take
a
look,
because
this
is
this
is
the
low-level
work
that
underlies
all
of
the
real-time
collaboration
features
that
we
want
to
ship.
A
So
it
will
benefit
from
lots
of
eyes
on
it.
The
fact
that
the
apis
are
changing,
but
that
we
no
longer
need
to
have
two
in-memory
models
in
order
to
be
backward
compatible
since
we're
doing
a
major
version.
Bump
is
really
good,
but
it
also
means
it's
our
last
opportunity
to
weigh
in
on
these
public
apis
before
we
commit
to
them
for
at
least
a
year.
But
given
our
history
longer.
A
Pr,
okay,
jeremy
you're
next
on
the
agenda.
E
Yeah,
just
to
click
a
heads
up
that
there
is
a
new
prerelease
for.
E
F
A
I
was
talking
to
someone
who
is
a
geneticist
by
training
and
she
uses
r
for
everything
and
she
said
she
might
be
willing
to
try
jupiter
if
she
can
just
go
to
a
webpage
and
download
something
and
double
click
on
it,
so
I
am
really
glad
to
have
this
ammunition.
Thank
you.
A
A
Okay,
we
have
an
empty
bullet
point
and
then,
after
that,
empty
bullet
point
we
have
frederick.
Is
he
on
the
call
right
now.
A
Okay,
so
why
don't
we
skip
past
him
and
no
other
bullet
points?
Oh
my
is
there
anything
anyone
wants
to
talk
about
while
we
wait
for
fred
or
we
can
come
up
with
a
funny
thing
to
say
to
him
when
he
arrives,
I
don't
know.
No
really
is
there
anything
that
you
would
have
added
to
the
agenda
but
thought
better
of
it,
but
now
we're
thinking
better
of
thinking
better.
C
Let's
say
if
we're
just
killing
time,
I
know
I've
been
banging
my
head
against
restorer,
getting
restore
working
on
my
code
viewer
and
I
was
gonna,
try
to
figure
out
who
to
reach
out
to
who
knows
the
intricacies
of
the
restorer,
because
it's
not
expected
to.
A
Did
you
read
the
essay?
I
wrote
about
state
restoration
alex,
probably
not
so
here's
here's
one
of
my
favorite
tidbits
about
the
jupiter
lab
code
base.
If
you
search
the
entire
code
base
for
the
word
subtle,
it
appears
only
once-
and
I
do
this
occasionally
to
see
if
anyone
else
has
used
the
word
saddle
and
they
haven't-
and
the
word
subtle
appears
in
a
giant
comment
that
I
wrote
that
says
the
life
cycle
for
state
restoration
is
subtle,
the
sequence
of
events
is
and
then
it
lists
out
seven
points.
A
Oh,
I
have
read
that
okay,
so
yeah
you
can
talk
to
me
about
it.
I
will
every
single
time
I
have
to
relearn
what
exactly
it
is
that
it
does
sadly,
but
I'll
be
happy
to
bang
my
head
against
the
wall
with
you,
yeah.
C
C
Yeah
I'll
re-read
that
after
the
meeting-
and
maybe
maybe
that
will
because
I
think
I
read
that,
like
really
early
in
my
in
my
study
before,
I
discovered
a
lot
of
the
edge
cases
that
I've
found
since,
like
the
fact
that
the
restoration
code
is
running
and
then
it's
deleting
more
itself
here
and
I
found
the
line
where
it
I
found
the
line
where
the
node
deletes
itself
to
because
it
literally
goes
through
and
says.
Oh,
is
this
widget
in
both
these
lists?
If
it's
not
set
parent
to
null.
H
I
I
G
I
H
No,
no
it's
just.
I
was
still
on
the
icmd
from
the
previous
meeting
on
notebook.
So,
okay
about
my
points,
let's
start
with
the
easy
one.
So
I
did
the
release
of
the
beta
version
for
the
3.4.0.
H
H
H
I'm
coming
at
the
middle
point,
there's
been
a
question
about
from
jason
and
brian
about
the
output
teaming,
because,
typically
vega
is
able
to
have
a
dark
mode
or
a
light
mode,
but
it
needs
to
re-render
the
graph
when
the
team
is
changing
and
the
question
is:
should
we
change
the
render
mine
interface
to
have
some
kind
either
to
have
a
signal
passed
to
their
that
interface?
For
saying
to
the
to
the
widget
that
the
team
must
change
or
do
we
need
to
to
do
something
else?
H
I
I've
put
some
points
there,
because
the
question
is
like
for
node
or
under
my
interface
doesn't
depend
explicitly
on
luminos
signals.
So
if
we
want
to
extend
the
interface,
that
means
we
add
a
new
dependency
and
that's
probably
more
a
question
about.
H
What's
the
scope,
do
we
want
that
that
those
interface
to
to
to
be
as
independent
as
possible,
or
is
it
okay
to
to
depend
on
more
part
of
lumino
things?
H
That's
mainly
my
my
question
on
that
that
specific
issue,
so
as
neither
jason
or
bryan
are
there,
I
think
the
best
is
to
for
people
interested
to
to
look
at
the
issue
and
please
comment
there,
and
that
brings
me
to
the
last
points,
but
the
first
one
from
my
list,
so
we
had
last
friday
a
specific
discussion
about
the
usage
of
a
toolkit
within
jupiter
lab.
The
conclusion
of
that
specific
meeting
was
that
people
were
in
favor
of
adopting
a
toolkit
based
on
the
specific
library,
fast
design
from
microsoft
and
to
to
move
forward.
A
A
So
we
can
tag
jupiter,
lab
council
and
ask
everyone
to
either
to
probably
vote
with
the
emojis,
I
think
is
a
good
plan
just
a
thumbs
up
and
a
thumbs
down
emoji
to
officially
record
a
vote
unless,
unless
on
this
call,
we
just
arrive
at
a
consensus
and
decide
cool
we're
doing
it,
but
because
we
want
to
make
sure
people
who
can't
come
to
the
call
have
a
chance
to
weigh
in
and
because
votes
nominally
are
supposed
to
take
a
week
and
because
I
don't
think
we
have
even
half
of
the
jupiter
lab
council
on
this
call
right
now,
it's
probably
best
to
do
it
with
a
vote
instead
of
trying
to
gauge.
A
I
don't
know
if
some
of
you
were
on
the
last
call.
We
tried
on
a
much
simpler
issue
with
only
seven
people
on
a
call
and
we
failed
to
arrive
at
a
definitive
answer.
So
oh
isabella
says
github.
Discussions
has
polls
yeah.
I
think
in
this
particular
case
an
issue
is
really
good
because
it
really
is
an
issue
that
you
know
might
show
up
in
a
search
in
the
future
and
like
it's
a
good
thing
to
point
to,
and
it's
not
one
that
is
like
a
sensitive
subject
or
anything
like
that.
A
It's
just
a
question
of:
do
we
pick
a
technical
direction
that
we're
happy
to
go
with,
so
there
does
remain
the
larger
question
of
votes
that
you
might
not
want
your
name
attached
to,
but
you
might
want
to
vote
on
and
we
don't
really
have
a
great
answer
for
that.
Yet,
thus
far
we've
been
conducting
those
in
a
mailing
list,
but
they've
only
ever
been
about
voting
for
new
council
members.
They
haven't
been
about
anything
else.
A
Yeah
so
so
yeah,
I
suggest
opening
up
a
team
compass
issue,
but
I
wanted
to
ask
about
the
previous
point
you
were
talking
about.
Unless
someone
wants
to
talk
about
this
one
right
now.
Are
there
any
comments
on
the
fast
elements
issue.
A
A
I
know
you
proposed
one
ish.
One
way
of
doing
it
is
with
the
signal.
Is
there?
Is
there
a
javascript
only
way
of
doing
it
like
adding
an
optional
callback,
or
something
like
that
or
like
what
are
the
things
that
you've
considered
here.
H
Way,
people
needs
to
know
about
it.
I
don't
know
I
mean
the
other
possibility,
I'm
I
I
don't
know
what
should
be
the
other
possibility,
actually,
the
only
one
I
have
fought,
but
I
am
very
not
happy
with
that
would
be
to
say
on
every
widget
that
is
rendering
an
output.
We
have
a.
We
have
an
attribute
and
instead
of
passing
a
signal
we
will
automatically
set
that
attribute,
but
I'm
not
a
big
fan
of
of
doing
that.
Neither.
A
So
it
sounds
like
you
could
get
away
without
modifying
rendermine
interfaces
at
all.
You
could
open
a
pr
to
the
vega
plug-in
that
simply
is
using
a
mutation
observer
on
the
body
tag
and
response
to
this.
Without
touching
render
mime
interfaces
at
all
right.
A
It
is
part
of
the
api
because
we
put
it
in
our
theme
style
sheets.
Don't
we.
A
A
All
right,
so
that
is
the
end
of
our
agenda,
and
today
we
don't
appear
to
have
additional
discussion
items.
Is
there
anything
anyone
would
like
to
announce
or
talk
about
that?
They
have
held
off
on
thus
far.
G
Yeah
it's
a
kind
of
silly
one,
but
I
was
just
looking
into.
I
know
that
words
no
stop
simple
mode
in
jupiter
lab
a
lot
of
the
decisions
for
how
that
was
made
right
exist
across
several
issues
was
kind
of
what
I
was
seeing.
I
wanted
to
check
like.
Is
there
a
place
like
those
all
just
exist
in
other
issues,
and
I
just
kind
of
have
to
sort
through
them
right?
There
was
no
like
place
that
people
made
decisions.
I
guess
I'm
just
trying
to.
G
I
want
some
info
on
like
why
some
decisions
were
made
because
I
like
to
understand
before
like
proposing
other
things,
but
I
wasn't
sure
if
there
was
like
a
good
place
to
go
or
a
person
that
was
like.
Oh,
I
spent
a
ton
of
time
thinking
about
simple
mode.
You
should
talk
to
me
that
sort
of
thing
it's
kind
of
a
weird
question,
but
open
call,
simple
mode
enthusiasts.
A
G
Jason
might
in
the
chat,
saying
yeah,
okay,
cool,
thank
you.
No,
you
don't
have
to
unmute
either.
I
was
just
it's
more
of
an
open
question
than
that.
I'm
having
a
problem.
I
just
I
was
like
watching
some
of
the
issues,
but
then
I
was
going
back
and
like
oh
there's,
a
bunch
of
issues.
I've
never
seen,
and
I
don't
know
if
I'm
just
out
of
touch.
So
thank
you.
A
So
I
think
some
of
the
discussions
you
might
find
will
be
obviated
by
the
new
direction
we
took
with
notebook
7,
because
for
a
really
long
time
we
thought
that
if
only
sample
mode
were
simple
enough,
that
that
would
be
sufficient
for
users
of
the
classic
notebook.
A
And
now
I
think,
simple
mode
is
a
little
bit
more
like
the
mode
in
an
ide
that
lets
you
focus
on
something
some
ids
call
it
zen
mode,
jason
says
focus
mode
exactly,
and
so
it
might
well
be
that,
even
if
you
find
a
lot
of
context
that
that
context
is
superseded
because
it
was
operating
under
the
assumption
that
jupiter
lab
would
be
the
primary
front
end
that
project
jupiter
releases,
but
that's
no
longer
the
assumption.
G
Yeah,
I
recall
that
discussion.
The
thing
that
I
was
looking
at
it
for
was,
I
know,
simple
mode-
is
used
for
the
like
mobile
view,
basically
or
like
to
basically
to
handle
all
the
responsiveness
from
my
understanding
like
I
was
playing
with
it
more,
and
that
seems
to
be
the
case,
so
I
yeah
that
that's
more.
Why
I'm
looking
into
it,
because
I'm
worried
about
the
responsiveness
for
some
accessibility
stuff,
but
I
don't
have
any
like
big
understandings
or
critiques
yet
I
just
would
rather
like
have
an
informed
understanding.
A
Yeah
yeah
and
if
you
run
into
something
that
one
of
us
wrote
some
time
ago
and
it's
ambiguous,
you
know
feel
free
to
try
to
get
us
to
resurrect
that
context.
But
I
can't
promise
anything.
G
H
A
H
I
can
start
by
that,
if
you
want
just
speaking
in
the
middle
to
let
me
the
time
start
the
demo,
so
we
have
been
working
on
on
a
new
extension
for
jupiter
app,
that's
a
search
and
replace
across
files.
It's
heavily
inspired
by
the
the
extension
you
get
in
vs
code
for
searching
across
files
version.
H
V1
was
just
released
like
a
couple
of
like
I
don't
know
if
I
can
even
say
a
couple
of
hours
ago,
so
very
fresh,
it's
not
yet
on
on
conda,
because
I
still
need
to
to
create
the
the
recipe.
But
it's
gonna
come
so
it's
using
it's
using
tools.
That's
called
rip
crap
behind
the
scene.
For
those
that
don't
know,
rip
crap
rip
crap
is
like
grep
but
very
efficient.
H
H
H
The
file
will
always
be
open
with
the
text
editor
because,
as
the
search
is
done
on
a
plain
text,
even
if
it's
the
json
or
notebook
file
that
will
be
open
with
the
file
editor,
because
that's
the
only
type
of
file
we
can
deal
with
within
the
extensions,
so
that
could
be
seen
as
a
limitation.
But
that's
due
to
the
technology
used
quick
question
on
that.
Frederick.
C
So
why
is
the
limitation
on?
How
is
there
a
limitation
on
needing
to
open
in
the
file
editor
and
it
can't
be
opened?
I
guess,
through
the
document
manager,
where
it
like
picks
the
best
factory.
H
So
it's
because
the
positioning
we
are
getting
from
the
underlying
tools
is
the
positioning
within
the
row
within
the
plain
text
file
version.
So
we
don't
know
what's
the
correct
position
of
the
like,
if
it's
a
notebook,
we
don't
know
which
cells,
if
it's
an
output
or
if
it's
an
input,
we
don't
have
context
of
of
anything.
We
just
know
that
it's
line
52
at
position.
Five,
that's
all
okay.
So
it's.
C
A
limitation
more
on
notebooks,
then
it's
like
a
json
or
or
like
other
other
types
of,
I
guess
yeah.
I
guess.
Okay.
H
For
the
user,
I
agree
with
you
that
would
be
very
surprising
for
the
notebook
because
they
may
expect
to
have
the
notebook
view
when
they
will
deal
with
that
for
the
json
by
default.
The
json
is
using
the
json
viewer.
That's
this
treeview
thing,
so
I'm
basically
never
letting
the
doc
manager
open
the
preferred
factory,
I'm
forcing
it
to
use
the
the
file
editors,
because
that's
the
only
one
I
can
use
for
saying.
Okay,
that's
the
match.
H
C
I
think
that
should
be
something
at
least
customizable
or
extendable,
because
I
know
I'm
thinking
of
my
my
big
use
case
in
like
a
lyra,
for
instance,
where
we
have
a
python
editor
and
our
editor
and
like
job
and
like
json
editors,
where
we
have
editors
where
they
would
have
those
lines
correct,
because
they
extend
the
file
editor
and
so
like.
We
would
want
to
see
those
open
in
those
extended
editors.
C
So
having
the
ability
to
say.
Oh
yeah,
no
use
this
or
something
along
those
lines
would
be
helpful.
C
A
Yeah,
you
have
a
couple
questions
in
the
chat
as
well
from
jason,
asking
what,
if
it's
a
binary
file
and
from
william
asking,
do
you
support
git
graph
for
doing
search
but
restricted
to
only
files
in
a
git
repo
and
there's
some
more
context
in
the
chat
as
well.
H
For
now
I
did
not
add
customization
for
saying:
don't
ignore
what
is
in
the
git
ignore,
for
example,
so
I'm
using
really
the
the
most
basic
option
of
for
crap,
saying
that
okay,
binaural
files
are
extruded,
git
ignore
file
are
taken
into
account
and
what
you
get
in
addition
is
more
like
you
can
tune
by.
You
can
add
filters
and
you
can
like
you,
can
do
in
vs
code.
H
If
that's
answer
the
question
and
of
course,
of
course
the
tools
has
lots
of
options.
I
I
I
did
not
for
no
have
a
full
mapping
of
or
options.
The
idea
is
really
to
have
a
first
version
and
then
extend
depending
on
the
need
of
the
user,
like
we
are
usually
doing.
A
If
you
wanted
to
support
jupiter
notebooks
as
first
class,
so
that,
instead
of
highlighting
every
text
occurrence,
ui
only
every
input
occurrence-
and
you
can
say
what
cell
it's
in
jump
to
that
cell-
those
sorts
of
things,
what
what
would
it
take
to
get
from
here
to
there,
because
it
seems
very
difficult
to
do
that.
But
I
have
to
imagine
you've
given
it
some.
What
do
you
think.
H
So,
for
now
the
there
is
a
couple
of
limitations,
it's
for
instance.
So
if
you
have
a
file
open,
it's
not
taken
into
account,
so
it's
always
searching
in
the
server
side.
So,
for
example,
if
here
I'm
doing
some
modification,
so
what
I've
done
for
now
is
that
if
you
have
matches
in
a
file
and
the
file
is
dirty
you'll
get
a
warning
saying:
okay,
what
you
are
seeing
as
a
result
of
the
search
may
not
be
accurate,
because
what
it's
in
the
current
editor
is
not
it's
not
saved.
H
So
that's
not
the
version
it's
on
the
back,
so
that's
mean
that
the
next
level
of
the
this
extension
would
be
to
say.
Okay
for
files
that
are
closed,
I
can
use
the
back
end,
but
for
files
that
are
open,
I
probably
need
to
use
something
directly
in
the
front
end
to
have
access
to
the
current
model
and
stuff
like
that,
and
that's
mean,
if
I
do
that,
I
may
be
able
to
say:
oh
I'm
I
probably
can
use
what
the
search
look
in
the
document.
Search.
H
Extension,
that's
in
jupyter
core
is
to
say:
okay,
I
may
be
able
to
say
okay
for
certain
type
of
document.
I
could
use
another
way
then
asking
the
the
back
end
to
do
the
search
for
me.
It
should
be
slower,
of
course,
but
it
if
it
brings
more
more
insight
for
user
and
if
it's
behind
a
configuration
flag,
it
may
make
sense.
H
H
A
So
it
could
be
that
you
use
rip
rep
to
find
the
subset
of
files
to
find
the
subset
of
notebooks
that
potentially
contain
the
word
you
want
and
then
having
some
sort
of
intermediate
format
may
be
where
inputs
are
stored
in
some
sort
of
temporary
file,
where
it's
one
line
per
input
or
something
like
that,
and
then
you
map
that
or
yeah.
It
seems
really
hard.
A
It
seems
really
hard,
but
it
also
seems
like
what
what
I
would
want
from
a
search
in
jupiter
lab
is
to
treat
notebooks
like
their
notebooks
and
they're
like
like
it
understands
that
they're
notebooks,
in
a
way
that
another
editor,
maybe
wouldn't
jason,
has
an
ambitious
question.
Can
we
ship
this
in
jayla
4.
H
C
H
Thank
you
for
taking
the
hat
off
nick
so
because
we
discussed
it
a
bit
on
tuesday
one
one.
One
usage
also
is,
of
course,
for
people
that
are
not
using
a
file
system
for
the
content,
and
that's
the
that
bring
the
same
kind
of
question
is
okay.
We
need
another
tool
for
doing
the
search,
so
that
definitely
is
possible
because
we
could
simply
reroute
the
request
to
something
else
or
do
like
the
content
manager
to
be
able
to
change
the
the
class
that's
doing
the
search
on
the
back
end
to
say.
H
Oh,
I
provide
the
custom
one
because
I
have
a
custom
content
manager,
for
example,
so
that
that
that's
that's
totally
possible
to
to
solve
that
trouble.
Let's
see
a
solution,
for
example,
also
for
a
jupiter
lit
could
be
like
you
could
preload
some
some
index
of
your
of
europe
of
your
folder.
If
it's
not
too
big
that
that
could
make
the
trick
too
so
yeah,
there
are
technical
solutions.
H
Yeah,
I
freely
agree
jason,
it's
just
I
don't
know
if
it's
a
a
road
we
want
to
take
like
to
have
features
that
are
different
depending
if
you
installed
the
software
from
conda
package
or
from
pip
package.
A
H
I
could
have
put
something
else
so
and
when
you
do
replace
you,
you
get
like
vs
code.
You
can
replace
everything
across
every
file.
If
you
do
that,
you
are
you're
prompt
with
a
confirmation,
because
you
cannot
undo
it
and
you
can
ask
for
not
seeing
the
the
dialogue
next
time,
but
still
to
get
some
kind
of
protection.
H
We
are
calling
the
creation
of
a
checkpoint
for
every
file
before
doing
replacement.
So
there
are
some
kind
of
backup,
let's
say,
and
otherwise,
if
you
are
replacing
a
single
it
in
a
file,
so
I
may
be
able
to
do
it
again.
But
basically,
if
the
file
is
closed,
it's
going
to
be
opened
and
the
change
will
be
in
the
editor
so
that
so
this
this
has
changed
so
that
you
can
eat
the
undo
and
the
replace
will
be
undone.
H
H
H
A
Yeah
yeah,
it
seems
as
though
like
what's
the
general
answer
to
how
do
you
get
meaningful
search
results
for
notebooks
like
notebooks,
are
one
kind
of
special
file,
but
you
know
there
might
be
other
special
thoughts
that
we
want
to
do
special
sort
of
searching
on,
and
I
think
the
answer
for
that
probably
has
to
be
something
on
the
server
side
of
this
but
yeah.
I
don't
know
I'm
just
thinking
out
loud.
H
A
Right
we
have
four
minutes
left,
so
I'm
gonna
stop
the
recording
in
case
there's
something
that
somebody
would
like
to
announce.