►
From YouTube: SunPy Developer Discussion
Description
After a week off it's back...
* rotate() issues
* 0.6.0 - 0.8.0 roadmap
* PR swatting
B
B
B
D
A
C
C
A
A
A
A
A
A
A
A
A
A
A
A
A
Tell
you
how
it
is.
One
is
a
lot
21
I,
a
town
hall,
the
Braves.
If
that
one
HR
example,
I,
have
breakdowns
on
the
fly
to
rotate
managing.
A
A
C
F
A
C
A
If
you
free
scaled
the
petals
of
their
image,
because
when
you've
effectively
done
the
reefs
tail
first
you've
moved
your
box,
you
not
only
have
you
changed
a
seat
up,
but
you've
also
scaled
your
thing
further
away
the
pics.
With
hordes
of
that
comment.
No,
why
is
my
internet?
What
is
going
on?
Oh
I,
don't
know
I
can't
remember.
I
may
be.
C
C
D
D
A
C
A
A
Is
an
easy
thing:
sticker
and
float
for
sample
further
float
conversion
before
we
parted
aside
the
image,
but
I
really
dislike
it
really
works
for
integers
are
the
point
of
why
that
academic
app-
and
I
did
also,
of
course
in
the
process
doing-
is
discover
another
fantastic
bug
in
some
pipe,
which
is
that
and
sub
map
doesn't
work
with.
The
image
is
rotated
because
our
WCS
and
pixel,
the
world
transformation
doesn't
take
into
account
and
rotate
matrix.
So
if
you
take
an
AAA
roll
over
image
and
triton
sub
map
that
you
get
garbage.
D
A
A
Trying
to
fix
that
up
when
I
get
some
time:
okay,
yeah
fighting
that
anyway
pack,
you
could
have
a
look
over
it
and
see
if
you
can
spot
your
clearing
error,
yeah.
A
A
E
G
A
E
E
D
C
E
Well,
as
far
as
I
room
that
I'd
make
again
I
just
made
the
connection.
Yes,
friends,
remember
what
I
read
with
with
when
we
open
the
page
right,
or
at
least
when
I,
we
first
discuss
the
saying
and
someone
bring
up,
cuddling
and
I
think
it
was
because
a
Steve
and
awesome
myself.
We
had
an
affair
play
with
that
and
we
were
trying
to
change
the
coding
exchange
the
race
array,
but
so
when
you
were
having
a
list
of
arrays
like
XY,
if
you
want
to
use
it,
you
have
to
transfer
them.
C
C
H
C
Should
be
the
other
way?
Basically,
we
use
the
issues
and
then
we
pull
issues
into
this
to
say:
okay,
this
issue
will
be
treated
by
this
room
man,
but
by
this
release,
rather,
the
gift
of
issues
would
be
our
wish
list
and,
as
we
identify
items
that
are
you
know
on
the
horizon,
or
you
know,
we
have
people
that
are
going
to
be
implemented
and
then
we
can
move
them
to
the
roadmap.
C
I
A
C
C
A
So
our
look
at
seven
release
is
going
to
be
another
long
gaze.
Up
from
these,
like
you.
D
A
A
A
A
C
C
I
need
to
get
a
fit
lines,
fit
spectral
components
and
presumably
also
pass
that
through
a
response
matrix
for
converting
photons,
two
counts,
rejected
response.
Matrices,
so
I
mean
that
that's
what
I
need
to
do
kind
of
my
kind
of
science
in
sun
pie,
but
again,
whether
that
belongs
in
course
on
pies.
It
is
questionable
well.
C
C
Usually
you
want
that
you
know
we
talked
about
in
the
way
we've
handled
on
sun
fight
where
we
try
to
get
back
to
the
physical
unit,
or
you
know
we
use
data,
that's
already
calibrated,
but
when
you
do,
you
know
sophisticate
a
spectral
fitting.
You
always
want
to
fit
to
the
raw
data,
and
so
you
need
to
forward
fit
through
your
response,
and
so
that
should
be
carried
inside
as
a
Python
object
of
some
type,
and
you
can
envision
a
fair
amount
of
metadata
associated
with
response
measures
as
well.
C
A
C
Yeah
the
challenge
that
I
see
for
that
and
the
way
I
just
described
it
here
is
even
if
spectral
fitting
or
an
affiliated
package
in
some
sense
you
you
I,
want
access
back
to
the
lower
level
stuff,
and
so
some
pie
at
some
level
would
need
to
still
make
that
available.
Even
if
that's
not,
as
you
know,
the
way
that
most
it's
not
the
pretty
way
to
look
at
the
data
or
something
like
that,
you
know.
C
I
mean
if
we
could
do
you
know,
I,
don't
know
how
much
has
been
discussed
and
a
broader,
but
certainly
even
I
have
cut
to
talk
about.
You
know
if
you
it's
just
always
just
or
just
the
raw
as
data
or
mostly
raw
data,
and
then
do
you
know
non-destructive
manipulation
of
the
data
and
store
that
history.
A
H
A
A
A
C
Yeah,
so
this
part
of
what
I'm
taking
goes
beyond
the
scope
of
some
pie,
almost
certainly
in
the
sense
that
you
know
you
want
to
have
petrol
components
and
the
spectral
components
that
generally
people
worry
about
in
solar
are
not
necessarily
the
same
as
an
astrophysics.
When
we're
talking
about
fitting
flares-
and
they
are
not
just
gaussians,
for
example,
so
and
then
the
extra
complication
is,
you
can't
often
fick
just
treat
gaussians
to
data,
that's
an
account
active.
C
They
don't
actually
look
like
gaussians,
and
so
the
whole
I
mean
I
haven't
looked
at
what
a
surprise
doing
like
I
said
I'm
sure,
especially
the
higher
energy
after
people
would
be
having
lots
of
the
same
concerns
as
I
would
in
terms
of
how
you'd
actually
do
data
analysis
or
at
same
interests
today,
I'm
not
concerned.
So
no
doubt
there's
motion
on
that
side,
but
yeah
I'm
not
entirely
sure
exactly
with
I
would
want
to
a
course
on
PI
to
make
that
make
sure
that's
enabled.
E
It
should
we
hear
we
make
something
up
a
document
with
the
workflow,
so
you
like,
you
could
expose
the
x-ray
workflow
like
how
what
exactly
what
kind
of
fittings
on
you
know.
You
can
run
off.
I
snapshot
of
the
steps
and
I
will
do
the
same
for
the
UV,
and
then
we
can
decide
from
there
like
stick
out
with
which
part
can
be
done
by
occupy,
which
first
can
be
done
by
other
packages
and
intend
to
get
a
better
view.
How
the
loop,
which
things
will
fill
most
of
solar
people,
need.
E
Know,
yeah
and
and
Kenneth
I
still
dunk
it
where
the
line
separates
between
what
we
do
next
to
fire,
what
we
can
do
with
sci-fi,
for
example,
so
and
I.
I
I'm
trying
to
do
that
anyway,
though,
to
discuss
it
with
Mateo,
see
what
what.
E
So
maybe
we
could
just
look
that
an
aside
and
then
and
then
put
them
yes,
the
jhino
parson,
together,
like
the
expected
par
now
in
San
pie,
is
not
out
there
with
30
seconds
yesterday,
number
of
counts
versus
a
a
wavelength
or
frequency,
or
even
you
could
be
in
time
frequency
in
the
other
axis.
So
so
it's
not
the
same
thing
that
we
were
like
having
an
x-ray
or
in
UV.
B
B
I
agree
with
you,
the
spectral
announced
this
would
be
spectral
fitting
will
be
very
important,
that's
something
that
we
haven't
well,
we
thought
about
obviously,
but
we
need
happens
about
fully.
B
Well,
I
would
say
that
something
that
we
don't
have
yet
and
we
will
need
to
have
a
future
as
more
instrument,
specific
tools
that
again,
of
course,
that
leads
to
the
same
question
of
whether
those
should
be
course,
some
pie
or
something
affiliated
or
elsewhere,
or
what
so
that
we
can
process,
maybe
process
raw
data
or
just
hyper
beam
use
tools
for
all
the
individual
instruments
that
we
have
in
it.
Kela
physics,
observateur,
bassist,
reuters.
Oh,
we
don't
have
that
many
of
those
yet.
A
G
C
C
Go
ahead
of
it
well
put
adds
a
part
of
the
quotes
or
question
there's,
but
when
you're
talking
about
an
x-ray
and
gamma-ray
world,
it's
not
something
that
you
could
get
calibrated
data
doesn't
necessarily
mean
that
you
can
bring
it
back
out
to
photon
fluxes
as
a
unit,
because
you
can't,
unlike
other
energy
regimes,
you
can't
just
invert
your
response.
You
can't
say
you
had
this,
you
know
this
many
counts,
and
that
means
you
had
this
much
flux
in
that
energy
bin
or
something
like
that.
C
You
really
have
to
it's
a
it's
a
wonder
its
most.
It
can
be
a
very
ugly
one
directional
thing,
so
you
can't
bring
it
back
to
a
high
level
where
you
can
say.
Okay,
this
is
a
foot.
This
is
the
photon
spectrum
from
reci.
This
is
the
photon
spectrum
from
Fermi
and
then
just
just
compare
them
just
at
a
high
level
way.
C
Normally,
when
we
look
at
something
like
rest
data
or
Fermi
data
or
most
of
the
time,
we
look
at
something
like
that
in
the
x-ray
gamma
ray
regime,
you
show
it
in
an
instrument,
specific
counts
way
and
that
by
you
you,
you
work
with
the
data.
You
analyze
the
data
where,
when
you
fit
it,
you
take
a
photon
model.
You
fold
it
through
the
same
thing.
You
put
it
into
count
space,
and
then
you
compare
that
with
your
data.
C
So
in
some
sense
the
calibrated
data
that
you
look
at
when
you're
an
x-ray
and
gamma-ray
or
regimes
still
has
units
of
count,
they're
good
chance,
the
counts.
You
really
believe
they're
calibrated
counts,
but
they're
still
count,
there's
still
a
lot
of
physical
unit,
and
so
when
some
pie
wants
to
plot
a
ressie,
gamma
ray
spectrum,
it's
going
to
not
be
in
physical
units
and
whether
we're
I
guess
happy
with
that
or
how
we
want
to
treat
that
I.
A
C
A
E
A
The
problem
I
have
with
that
work
in
general.
Not
specifically
what
he'd
be
doing
them
in
general
is
the
problem.
Is
that
while
it
scratches
the
ice
quite
nicely,
it
is
so
not
scalable,
it
hurts.
So
we
have
no
no
I
been
prevented
scalable,
but
we
have
owned
a
3d
spatial
data
here,
but
it's
all
time
to
use.
So
we
need
a
4d,
keep
object.
A
H
A
A
E
A
Outfits
see
set
up
already
is
like
that.
It
assumes
every
HD
you
inside
of
its
violence,
separate
yeah.
So
if
you
implement
it,
a
spectral
cube
factory
that
takes
in
a
fit
file,
it
would
spew
out
a
list
of
tubes,
exactly
which
is
the
crimper
yeah,
and
what
I
would
for?
What
I'm
just
was
saying
is,
but
yes
for
a
3d
spatial
cute.
A
D
E
Yes,
to
be
able
to
play
with
them
and
see
how
you
doing
over
stuff
and
the
phone
is
like
on
my
side-
I
thought.
Yes,
look,
I
understand.
Yes,
this
picture
of
us
Ace
or
CVS
or
summer,
a
picture.
The
parameters
that
the
happy
snaps
in
your
structure.
I
haven't
used
them
all,
so
I,
don't
really
know
or
hot
researchers
could
compare
Damon
and
guys,
not
okay,
and
I
did
a
feeling,
I'm
completely,
I
really
mean
with
you
in
the
sense
of
the
puck.
E
A
Yeah,
I
don't
know
I
mean
the
thing
is
we
need
to
come
up
with
a
way
of
dealing
with
that
in
the
generalized
fashion
before
expired,
what
they
out
of
control,
because
I
think
we're
at
the
point,
unfortunately
reasoning
to
be
ultimately
where
we
have
people
wanting
to
use
sapphire
for
ice
data
or
40
data,
cubes
or
different,
all
manner
of
good
things,
but
to
scratch
each
one
of
those
issues
we're
going
to
create
a
new
command,
a
fight
and
you're
moving
rapidly
LRA.
The
problem
is
the
upstream
at
reply.
Tools
are
really
there.
A
E
No,
we
could.
We
could
write
that
in
our
guidelines-
and
here
I
say
we
are
going
to
do
that
if
we're
doing
something,
it's
for
testing
purposes
or
something
like
that,
you
know
what
yes
I
mean
in
that
pull
request
like
what
he's
doing.
We
can
just
hold
you
there
and
just
make
anything
work
in
a
better
way.
A
Yeah,
it's
unconditional
time.
Hopefully.
D
E
I'm
going
to
then
then
wrote
an
email
about
what
Oh,
what
we
think
where
our
limits
of
some
pies
regarding
his
coat
and
I.
Why
I
squeaking
see
if
someone
answer,
but
no
one
else?
Yes,
I'm
going
to
put
my
comments.
There
I
think
that
Dan
took
a
pic
topic,
but
the
same
is
that
whether
his
coat
or
his
calibration
algorithm,
it
should
be
part
of
some
pie
or
not
I,
see
points
in
favor
and
against
it
in
favor
is
that
they
haven't
been
any
code
down
in
the
30
years
against
it.
E
E
E
Now
you
mention
it
and
but
then
the
area
see
Chad
we
were
having
like
I
was
telling
him
why?
Let's
go
the
white
shouldn't,
look
good
and
in
a
way
like
if
no
one
has
ever
done,
that
the
early
I've
only
up
only
just
afraid,
on
the
sense
that
people
will
say
that
that's
the
official
thing
from
ghost
or
whatever,
where
it
is
just
hop
in
our
stomach
now
on
the
other
side
and
eat,
but
I'm
gonna
say
like
if
he
has
like
the
code
there
and
gets
you
know,
updated
and
he'd
never
change
it.
E
Maybe
no
one
will
change
it.
Now,
he's
only
probably
Murphy
only
he
doesn't
what
he
wants
to
make
it
easier
for
the
user
that
his
point,
so
he
didn't
make
a
fair
user
installed
two
different
packages.
It's
it's
a
problem,
hey
I,
think
it
might
be,
but
pity
fitting
is
coming
better
and
better.
So
I
would
not
make
that
as
a
problem.
E
B
Right
I
mean
I,
think
I,
know,
I
know
where
it's
coming
from,
in
the
sense
that
as
a
user,
you
know
you
kind
of
want
to
like
install
one
thing
and
then
like,
since
this
is
great
I've
got
all
the
tools
I
need
so,
but
on
the
other
hand,
you
could
set
up
something
to
meet
package
installation
very
easy.
That's.
A
B
A
B
I'm
another
Chris
on
up
front,
though,
if
we,
if
we
start
taking
stuff
out
of
some
pie,
let's
say
put
it
in
an
affiliated
packet,
we're
gonna
have
problems
with
testing
to
learn.
We
there
and
change
something
in
sunfire.
You
might
break
facility
package
and
need
your
testing
show
up.
Any
of
that.
You
know,
I,
don't
know
how
that
would
work
well,.
B
B
A
B
A
A
A
A
B
B
You're
right,
I
guess
what
we
would
maybe
we
could
maybe
have
them
as
I'm.
Just
thinking
about
the
future.
It's
a
person
who's
responsible
for
keeping
in
touch
with
affiliate
package
managers
so,
like,
let's
say,
calm
a
communications
person
for
lack
of
a
better
term.
Like
do
you
know
who
the
package
manager
of
tips?
Is
you
know
what
Stan
Ryan
you
like?
You
know
that
some
fire
is
being
updated.
You
say:
well,
look
you
keep
in
touch
with
them,
saying
we're
making
these
changes.
B
A
Yeah
I
agree,
but
I
also
get
I
mean
short
in
the
future,
with
lots
of
the
most
packages,
maybe
most
of
it's
from
pipeline
stuff.
Something
like
that
is
that
a
good
idea,
but
I
still
feel
the
burden
of
responsibility
is
the
other
way
around
in
it's
up
to
the
affiliation
package,
people
to
keep
up
to
date
with,
what's
going
on
in
holy
like
and
like,
in
only
waits,
they
should
be
called
developers
of
makeup
from
Minnesota
corners,
but
certainly
active
develop.
B
A
B
A
B
I
mean
I,
actually,
don't
think
one
of
his
points
was
that
in
SS
w
you
get
everything
you
want.
First
time
Ryan
but
I
actually
think
that's
not
it's
not
really
accurate,
because
when
you
install
ssw
okay,
you
get
to
choose
your
packages
plate
you
there
may
find
that
you
want
some
other
ones
later
and
then
you
have
to
go
and
install
those
separately.
Just
like
you
would
in
some.
D
B
So
it's
really
the
same
view.
It's
very
similar.
Well
that
happens
to
me
a
lot
where
I
realize
there's
Isis
w
package.
They
don't
have-
and
you
know,
okay,
it's
easy
to
install
the
next
one,
but
it
would
be
easy
and
some
pie
too,
but
I
mean
in
Python
too
so
I'm,
not
sure
I
agree
that
that's
a
different,
no.
A
Anyway,
I
think
it's
an
interesting
question,
because
in
all
about
discussions
about
my
affiliated
packages
and
own
personal
stuff
that
we've
had
out
at
this
point,
we
have
been
considering
considering
my
primary
instrument
pipeline
code
and
it
can
its
root
specific
calibration
code
or
instrument
specific
this
that
the
other.
Where
is
what
dance
will
and
watch
crop
top
in
other
places
in
the
last
few
months,
is
physical,
physical
analysis
code
yeah,
and
we
haven't
really
discussed
that
very
much.
A
What
we're
going
to
do
with
physical
knows
its
code,
and
where
do
we
draw
the
line
where
it
becomes
it's
a
thing,
as
opposed
to
being
part
of
something
which
is
really
a
like
it?
What
I'm?
Currently
at
least
it's
very
data
focused
library,
is
for
gay
data.
Munging
data
yeah,
giving
it
the
users
access
to
data
in
a
standard,
bios,
uniform,
Anna.
B
B
A
E
E
Yeah,
no,
no
well,
lady
L
either
like
that,
because
this
calibration
even
hasn't
been
gone
at
all,
like
those
people
never
come
across
to
do
it
in
that
way.
I
don't
like
there's
another
way
to
doing
it,
but
it's
in
has
I
mean
no,
and
that's
the
stance
point
like
to
say
it's
not
a
like
if
it
won't
change
in
that
respect
and
yeah,
we'll
just
change
the
number
of
the
new
satellite,
whatever
dreams
are
like
it
up
there.
In
the
respect
I
mean
it's
a
collaboration
thing
that
we
could.
E
E
Supporting
it
or
of
maintaining
a
maybe
doesn't
he
says
well,
I
won't
need
any
maintenance
I
doubt
that
from
a
nice
soft
work,
but
I
mean
it
under
some
high
in
my
open
the
doors
to
other
people
to
bring
all
the
coats
and
it
will
open
the
door
to
say.
Oh,
why
now
my
tote
that
detects
whatever
or
it
just
calibrate
this
in
a
fancy
way?
Why
cannot
be
here
as
well
so
that.
A
H
H
E
B
E
E
A
B
As
a
few
and
that
their
non
unique
as
well
as
the
problem,
the
answers
you
get
non-unique
so
I
I
saw
agree
with
what
the
beep
is
saying
like
there's
miles
poor
answers
you
can
get
out
and
you
don't
want.
You
won't
have
a
situation
where,
like
people
think
that
something's
official,
because
it's
in
some
pie
like
this,
is
the
trick.
This
is
the
correct
answer,
because
it's
in
some
pie
we're
in
reality
everyone's
got
their
own
method
and
getting
lots
of
different
results.
E
Well,
I
think
the
affinity
packages
need
also
like
that
kind
of
control
as
well.
You
know
and
some
kind
of
firm,
as
you
say,
we
are
not
going
to
reinvent
the
wheel.
So
if
something
is
done
in
one
or
it's
something
is
done
in
some
pie
and
definitely
package
to
use
on
its
opponents
on
high
and
then
build
on
that
it's
not
yesterday.
So
let
me
feel
the
whole
reading
fire
pits
file
and
all
the
stuff
you
know
so,
but
as
it
should
have
sent
on
a
controller
or
classification.
E
E
Don't
see
any
such
problem,
my
only
problem
is
like
us.
Can
we
say
when
we
start
to
when
the
people
sucked
assume?
Oh
it's
in
some
pie?
Is
it
true,
actually
nothing?
I
dl
div
Alicia.
Oh
this
patrolling
give
me
that
or
this
program
which
I
don't
know,
is
the
en
calculation
and
then,
but
you
know
who
did
leave
perhaps
what
it
starts
with.
You
know,
people
kind
of
done
don't
go
into
it.
G
E
That's
right,
but
it
will
make
them
an
extra
effort
to
get
to
that
place.
You
know
they
will
let
you
have
the
documentation
by
themselves,
so
the
people
don't
want
to
do
anything
because
I
mean
it
doesn't
matter
if
it's
an
athlete
package
or
yes,
I
responsively
anywhere,
but
he
still
have
to
go
into
that
promise.
You
know
the
duck
thing,
oh,
not
for
first,
if
people
will
be
able
to
be
wrong.
E
A
A
Because
I
mean
in
the
old
time
again
come
back
needing
money
and
build
one,
because
we
could
maintain
binary
bill
versions
of
all
the
like
accepted
Adam
affiliated
packages
on
bid
star,
which
means
that
if
you
were
using
anaconda,
you
could
just
install
them,
stray
off
no
problem
and
it
should
just
work.
We
have
set
compile
packages
that
we
work
very
making
against
at
all
times
and
it
just
by
haters,
so
I
reckon
that
anaconda
is
actually
the
solution
to
those
problems.
A
A
Mean
these
mines
not
the
answer
to
all
the
problems,
because
there
are
fundamental
limits,
climbing
it's
it
statically
linked
within
itself.
So
it
does
engine
system
moment
which
has
its
advantages
but
also
I've
run
into
problems
with
hpc
on
doing
my
research
work,
conduct
I,
couldn't
because
it
didn't
use
the
clusters,
NPI
library,
for
instance.
So
as
well
and
I
said
it's
the
solution
to
all
the
problems
that
I
really
mean
it's
just
a
solution
to
log
from
any
need
for
a
normal
end
user.
On
the
on
your
laptop
cutting.
What
review?
B
A
Is
the
guys
are
the
liberators
until
we
come
to
us
I?
Don't
the
problem
is
I,
don't
think
we
can't
honestly
soul
feels
like
it's
we're
going
to
have
to
find
a
lesser
roof.
Lesser
of
two
evils
come
to
an
answer,
but
any
dealer
document
we
draw
up
specifying
what
belongs
in
the
library
doesn't
will
be
challenged
every
time.
I
minus
dis,
be
along
in
the
library,
for
this
reason
is
reason
enough.
People
yeah.