►
From YouTube: OpenShift Commons Briefing #134: Deploying Dash on OpenShift - Chelsea Douglas(Plotly)
Description
Dash is an Open Source Python library for creating reactive, web-based applications. No JavaScript required. Released in June of 2017, Dash is for those who use Python for data analysis, data exploration, visualization, modeling, instrument control, and reporting. Chelsea will provide an overview of the framework, and Anudha will demo a sample Dash application she has deployed on OpenShift.
A
A
Hello,
everybody
and
welcome
again
to
yet
another
OpenShift
Commons
briefing
this
time,
we're
going
to
talk
about
something
that
I'm
really
excited
about
the
folks
in
Red
Hat.
It
means
I
have
done
some
amazing
work
around
using
a
product
called
and
a
company
built
by
a
company
called
Lockley,
and
we
have
their
developer
advocate
Chelsey
with
us
today,
as
well
as
Randy
Ackerman
and
Anita
Gupta
from
Red
Hat,
who
have
been
using
it
internally
and
I'm
gonna
stop
there
and
let
them
introduce
themselves
and
they're.
A
B
I've
read
that
and
we've
been
partnered
with
Chelsea
on
this
demo
and
also
for
some
internal
applications
that
we've
been
building
out
on
open
ship
and
I'll.
Just
give
a
little
bit
of
a
background
on
what
brought
us
to
using
on
open
shift
and
then
Chelsea
will
give
a
little
bit
more
of
a
in-depth
on
what
is
and
annuda
will
put
all
that
together
and
show
how
we
can
launch
applications
using
on
open
ship.
B
So,
just
a
little
bit
of
a
context
on
what
problem
we
were
trying
to
solve
was
we
were
looking
for
a
scalable
way
to
deploy
pre
data
intensive
applications
to
all
of
our
sales
and
Finance
associates
at
Red
Hat.
Previously
we
were
hosting
our
our
own
internal,
are
shining
server
and
having
some
struggles
kind
of
kicking
up
with
the
and
on
that,
so
some
of
the
key
things
that
we
were
considering
was
one
is
we
wanted
to
be
able
to
host
applications?
That
was
something
that
you.
B
That's
really
not
our
skillset
to
wanted
to
stay
as
much
as
possible
using
Python
and
that's
typically
what
we
build
all
of
our
analysis
and
models
in
and
be
great
if
we
could
kind
of
take
it
the
last
mile
and
deliver
it
to
our
users,
just
all
staying
in
the
Python
ecosystem.
And
thirdly,
we
wanted,
to
you
know
as
simple
as
possible
from
a
development.
In
you
know,
we
wanted
to
have
something
to
look
nice
out
of
the
box
and
with
very
little
customization
needed,
and
we
certainly
think
solve
these
problems.
C
So
it's
built
on
top
of
flask,
Flatley,
j/s
and
react.js,
and
it's
ideal
for
any
anyone
working
typically
in
Python.
So
basically
it's
just
a
tool
where
you
can
create
your
web
application
all
while
working
in
Python.
So
in
order
to
get
started,
you
have
all
the
packages
on
pip.
So
it's
pretty
easy
to
install
you
just
have
to
and
stall
HTML
components,
core
components
and
renderer
and
all
of
those
are
easily
installed
with
the
pip
install
commands.
C
Great
so
here
we'll
see
the
app
top
hi,
and
so
this
is
just
a
Python
file
where
I've
imported
a
few
Python
packages
here.
So
in
addition
to
the
packages
of
I'm
using
pandas
and
then
date
time
as
well,
first,
we
can
initialize
the
app
and
then
also
set
the
server
variable
and
then
here
since
we're
building
a
web
application
will
want
to
use
a
CSS
file,
and
our
team
here
has
created
this
very
basic,
yet
CSS
stylesheet,
which
we're
hosting
on
code
pen
and
I've
added
that
to
my
app
with
this
app
dot.
C
So
setting
up
my
app
here,
I'm
first
creating
the
layout,
which
will
be
an
overall
HTML
div
and
then
filling
in
those
children
with
different
sections.
That
I
want
the
different
sections
that
I
want
in
the
app.
So
first
I
mean
setting
up
a
header
here
with
a
different
background,
color
and
then
I'm
dividing
the
rest
of
the
app
there
into
some
columns,
which
we'll
see
in
one
moment.
C
C
So
down
here
when
I'm
setting
up
the
mark
or
the
slider
I'm
going
to
be
using
all
of
the
elements
in
that
year's
column,
that
I
have
there
so
now.
You'll
notice
that
in
both
my
GCC
drop
down
and
my
GCC
slider
I've
set
up
an
ID.
So
we're
going
to
use
these
AI
IDs
below
in
the
callback
function
that
I've
written
so
before
we
get
to
the
callback
function.
We'll
just
quickly
check
here,
this
last
tip
that
I've
set
up
is
composed
of
a
DCC
graph
component,
and
this
has
the
ID
of
my
graph.
C
So
what
I'm
trying
to
do
here
is
use
the
drop
down
component
and
the
slider
component
to
update
the
graph.
So
we
can
do
this
with
the
app
callback
where
I'm
using
those
IDs
that
I've
set
above
and
then
saying
what
I
want
to
update
so
in
the
graph
I
want
to
update
the
figure,
so
the
output
is
going
to
be
my
updated
graph
and
the
input
will
be
both
the
drop
down
and
slider
values.
C
So
once
those
are
updated
here,
we
just
have
some
code
to
create
the
graph.
So
if
you
have
used
Watley
before
it
will
be
super
familiar
with
this
code,
it's
very
similar
to
creating
graphs
with
our
Python
API,
where
I'm
just
defining
the
type
of
chart
that
I
want
to
create,
as
well
as
x
and
y
values,
and
then
some
stylistic
settings
as
well,
including
some
layout
style
here
great.
So
now
that
we
kind
of
see
what
the
app
is
set
up
to
do.
We
can
run
this.
C
C
Already
have
it
open
here
so
here
we
see
that
editor
with
a
different
background
that
I've
created,
and
we
can
also
take
a
look
zoom
out
a
little
bit
at
the
drop
down
as
well
as
the
slider
and
the
graph
that
I'm
trying
to
update
so
you'll
notice
that
the
graphs
are
flatly
graph.
So
they
have
all
of
our
interactive
functionality
built
in
so
on.
C
C
And
it
will
get
the
Americas
in
there
too
great.
So
as
I've
added
the
options
you
can
see.
If
then
added
to
my
chart
down
here-
and
we
do
have
this
one
outlier
over
here-.
D
C
Maybe
you
will
want
to
first
take
a
look,
removing
Asia
and
so
clicking
on
the
chart
will
add
and
clicking
on
the
legend
will
add
and
remove
those
traces
from
the
charts.
So
it's
kind
of
another
option
do
using
the
continents
drop
down
here,
so
would
also
have
the
option
of
just
removing
Asia
that
way
as
well,
and
then
we're
also
welcome
to
click
and
zoom
click
and
drag
to
zoom
in
a
bit
more
there.
C
So
with
this
example
and
many
of
the
dash
examples,
it's
a
really
good
way
to
kind
of
quickly
put
an
application
together
where
you
can
kind
of
explore
your
data
set
further,
whether
you're
looking
at
changes
over
time
or
comparing
multiple
traces
by
adding
different
things
via
drop-down.
It's
a
really
good
way
to
quickly
get
up
and
running
with
an
exploratory
tool
great.
C
So
I
went
and
went
through
our
app
dot
pie
already,
but
we
also
have
the
proc
file,
which
will
be
very
relevant
when
we
do
the
deploy.
We
were
just
setting
up
the
unicorn
definition
there
and
then
adding
the
app
server
and
then
also
the
readme
that
we're
using
with
the
new
push
to
gate
and
then
finally,
our
requirements,
x-square
I've,
listed
out
the
Python
packages
that
I'm
using
here.
A
A
C
I'm,
sorry
about
that,
so
should
I
explain
those
yeah.
C
C
C
Text
file
here
that
includes
all
of
the
packages,
as
well
as
the
other
packages
that
I've
used
in
my
my
application,
so
just
forgot
to
add
pandas
thing
so
we're
adding
pandas
in
and
then
I
have
a
readme
dot
and
which
I'll
use
when
pushing
to
github
and
then
the
proc
file
as
well,
which
will
be
used
during
the
deployment
process
which
we'll
discuss
next.
C
C
C
A
D
How
to
deploy
a
application
to
open
ship?
So
there
are
a
few
things
to
take
care
of
in
the
plan
and
deploying
the
on
open
ship
like
the
first
thing
is
that
the
app
should
be
created
and
run
through
a
file
called
this
G
dot
path
and
within
this
3.5
we
should
have
variable
application
pointing
to
this
last
server,
and
we
should,
since
we
will
be
using
unicorn
as
a
Vinci
server.
So
the
unicorn
configurations
should
be
added
to
a
file
we'll
call
it
country,
dot
path
and
then
an
environment.
D
D
D
D
Okay,
so
now
we
need
an
environment
variable
up
conflict
on
open
ships
to
point
to
this
considered
PI
file.
For
that
we'll
have
to
create
a
file
named
environment
within
our
dot
s,
2
is
folder,
and
these
projects
to
directories
so
first
we'll
make
a
folder
named
tortoise
to
us
and
then
we'll
create
a
file
environment
within
this
tortoise
to
a
folder.
D
D
A
A
A
Wondering
if
we
can
course
the
plotly
folks
into
posting
some
of
that
in
their
their
documentation
to
so
that
other
people
who
are
using
Watley
and
openshift
have
something
similar
so
someday.
Hopefully
we
can
get
that
dub
page
there
and
move
that
into
someplace
that
will
get
maintained
for
the
long
run,
but
annuda
is
not
maintaining
it
forever,
but
this
has
been
really
good.
It
was
pretty
pretty
simple
process
to
do
that:
I
loved
how
it
was
telling
you
to
update
your
pip
install
come
through
the
new
version
there.
Well,
we
were
waiting.
D
A
A
Awesome
all
right-
well,
this
I
have
to
say
is
really
amazing,
because
you
know
how
many,
how
many
of
us
have
tried
to
create
graphical
interfaces
and
applications
and
started
from
scratch.
I,
don't
know
me,
I've
been
coding
for
so
many
years,
and
I
am
a
devout
Python
ista
though
I
certainly
love.
This
and
I
know
that
it's
something
that
I
definitely
will
be
using,
and
it's
been
a
wonderful
tool
for
the
Red
Hat
insights
team
and
in
the
finance
group
over
the
past
few
years.
A
So
that's
that's
one
and
we
would
definitely
love
to
see
things
that
are
so
easily
deployed
on
on
openshift
with
a
few
minor
tweaks
and
the
use
of
source
to
image.
I
think
is
a
really
good
way
of
making
sure
that,
as
new
releases
of
plotly
and
new
updates
to
your
application,
go
then
you're
kind
of
safe,
because
it's
easy
to
upgrade
the
images
and
without
affecting
the
application
itself.
So
Randy
I'm
wondering
if
you
want
to
add
more
words
in
there.
B
It's
been
a
huge
success
being
able
to
scale
these
applications
in
a
way
that
really
lets
open
shift
do
all
the
head.
It's
opened
up
a
lot
more
of
those
resources
that
we
were
using
other
platforms
on
now.
We
can
do
more
cool
analytics
on
those
openshift
hosts,
all
of
that
and
Wadley
itself
as
a
framework
I
think
it's
just
a
terrific
in
its
ease
of
use.
It
really
has
a
nice
learning
curve
that
I
think
most
most
people
can
get
into
and
a
couple
days
to
get
up
and
running.
A
Yeah,
thank
you
both
and
Anita.
Thank
you
for
staying
up,
late,
I
know
you're
in
the
far
far
flung
part
of
Red
Hat
out
there
in
India
and
Pune,
but
we
really
appreciate
you
staying
up
and
doing
this
for
us
and
we
look
forward
to
also
showcasing
other
tools
at
Randy
and
his
team
are
using
on
in
different
aspects
of
AI
and
ml
and
data
analytics.
A
So
look
for
more
of
those
to
come
as
I
coerce
them
into
presenting
the
other
pieces
of
their
toolkit
because
it's
as
we
say,
we
do
eat
our
own
dog
food
and
we
treat
our
dogs
very
well.
Hopefully,
we'll
get
some
more
time.
So,
thanks
again
everybody-
and
we
really
appreciate
you
taking
the
time
to
listen
to
this-
and
if
you
have
any
questions,
I
will
put
Chelsea
and
her
resources
and
how
to
get
to
those
links.
And
then
it
is
links
into
the
YouTube
video.