►
From YouTube: SunPy Developer Discussion
Description
PRs!
C
B
B
E
E
F
D
B
B
B
F
F
E
Yeah,
I'm
a
I
wanted
to
say
a
bit
about
this,
the
PRS
it
kind
of
so
this
PR
I
feel
like
it's
sort
of
our
fault
right
I
mean
Dan
did
as
much
as
she
could
and
at
some
point
like
it's
our
responsibility,
her
right
and
I
would
suggest
that
if
the
PR
is
outstanding
for
too
long,
then
we
should
kind
of
lower
our
are
sort
of
expectations
or
and
accept
it
such
that
it
becomes.
It
forces
us
to
be
our
responsibility
to
fix
it.
Does
that
make
sense,
I.
E
Means
that,
like
we
haven't,
been
doing
our
job,
you
know
if,
if
the
developer
has
been
responsive
and
has
been
doing,
you
know
making
the
changes
that
we've
been
suggesting.
You
know
if
we
keep
asking
for
changes
eventually,
it
kind
of
turns
it.
This
like
evil
cycle,
where
you're
just
asking
them
to
make
it
more
and
more
perfect,
and
it
will
never
reach
that
goal
and
I
think
it's
unfair
to
developers
that
put
in
PR
is
that
we
wake
make
them
wait
too
long.
E
D
B
E
B
E
F
E
I
mean
it's
easier
because
people
have
limited
amount
of
time,
it's
easier
for
people
to
look
at
a
smaller
PR
and
accept
it.
You
know
smaller
one,
smaller
one,
when
it's
one
huge
one.
You
know
people
just
you
know
hard
they're,
not
willing
to
give
it
a
huge
chunk
of
their
time.
You
know
I
mean
right
so.
E
E
F
B
F
Well,
since
Steve
brought
that
up,
let
me
just
pull
that
up
for
myself
here
so
the
late.
The
goes
one
that's
still
outstanding,
and
this
is
this
Michael
record.
This
is
another
person's
in
last
year
and
had
a
lot
of
helpful
comments
and
over
the
months,
and
they
have
all,
as
far
as
I
can
tell
them
as
far
as
nowhere
have
been
implemented
and
the
most
reasons
of
which
was
just
about
three
weeks
ago,
which
is
about
the
the
Sun
pipe
quantity
and
hands.
F
It
also
now
all
passes
Travis
and
it
is
the
complete
it's
the
second
part
of
essentially
the
goes
ssw
software,
the
facilities
or
the
things
that
you
can
do
and
goes
as
W
would
go.
This
is
the
second
half
of
that
capability,
and
so
once
this
is
merged,
you'll
be
able
to
do
all
the
all
the
goes
nervousness
and
sun
pie
and
that
you
can
do
in
SS
w.
So
that's
what
this
is.
As
far
as
I'm
aware,
it's
good
to
go
up.
Artist
reverses
unless
they're
more
comments.
Of
course,
I.
B
B
F
B
F
Weren't,
actually
adding
okay,
so
so
far,
the
previous
goes.
Pr
allowed
us
to
get
temperature
and
emission
measure
from
goes
observations.
This
is
now
allows
us
to
get
the
radiative
losses,
both
across
all
wavelengths
of
the
goes
admitting
plasma
and
also
literally
just
sort
of
adding
up
the
integrating
the
the
flux
in
the
goes
past
bands.
F
So
the
second
one
is
is,
is
a
is
almost
essentially
sort
of
integration,
but
it's
it's
something
that
is
just
part
of
the
go
sort
of
package
of
software
analysis,
package,
inevitable
and
the
second
actually
use
a
chianti
to
get
the
energy
radiated
across
all
wavelengths.
So
it's
not
it's!
It's
a
lot
more
than
just
a
simple
integration.
So
this
completes
the
goes
analysis.
Functionality
for
sun
fire
makes
it
equivalent
to
that
better.
Existing
as
a
double.
B
F
D
D
B
B
E
F
E
F
Well,
let
me
see
hang
on,
I
think
I've
already.
I
have
that
capabilities
pub.
This
didn't
I.
We
had
a
function
up
in
visit,
you
till
death
or
something
check,
download
file
which
basically
checks.
If
the
file
is
in
a
given
location
and
the
default
is
the
sunfire
data
download
directory
and
if
it's
there,
it
doesn't
download
it.
The
downside
of
that,
though,
is
if
there
is
a
war.
Up-To-Date
version
of
the
file
that
function
in
itself
will
not
know
that
it
will
just.
E
B
E
F
These
ones,
not
not
the
chianti
ones
at
the
moment,
the
we
did
discuss
this
and
and
it
would
involve
using
chianti
pie
if
we
wanted
to
do
it
completely
separately,
and
there
were
a
number
of
issues
with
had
such
as
licensing
such
as
I
mean
it
didn't
sound
like
it
was
impossible.
That's
not
like
they
were
numb.
It
was
a
big
project
to
make
happen,
so
that
would
be
not
a
short-term
solution.
F
B
F
E
E
B
B
On
putting
in
an
application
sort
of
under
the
Python
in
astronomy
type
umbrella,
astra
by
son,
pajas
requires
affiliated
projects,
maybe
whitey
anyone
else.
Who's
interested
that
does
astronomy,
related
stuff
could
jump
on
the
train.
Basically
yeah
so
basically
create
another.
You
have
our
own
umbrella,
which
is
not
the
SF.
Basically.
B
B
E
E
I'll
I'll
email
him.
That's
what
about
it.
B
And
I
think
the
moment
our
current
running
project
ideas
is
this:
three
months
air
WTF
missed
three
months:
wiki
page
I,
don't
know
whether
we've
got
a
specific
g-shock
one
up
and
running
yet
nope.
Don't
think
we
do
that's
fine.
You
should
probably
stay
in
fremont
project
ideas
for
a
while,
anyway,.
C
B
F
E
B
F
Sure
we'll
sell
me
the
link
to
that
wiki
because
I
haven't
actually
good
use
it
before
it's
in
the
tip.
We'll
just
tell
me
where
to
find
a
date
in
the
group
chat.
Oh
it's
in
the
group,
chen
rui,
em
I,
a
spurt.
F
E
F
B
Gulf
PR
from
Oban
that
what
I
was
just
about
to
suggest
after
that,
an
epic
detour
we
just
went
on
I
mean
this
needs
a
good
m
I
need
to
read
through
it
and
review
it
properly,
but
sunglass
someone
else,
probably
as
well
yeah.
This
needs
a
thorough
review.
I
mean
it
does
look
like
it's
mostly
there,
but
it
still
does
look
like
it
neater,
probably
going
over
yeah
the.
E
One
thing
I
would
say:
is
that
I,
but
not
to
the
initial
like
curve,
so
what
I
see
you're
doing?
Is
you
take
the
goes
like
her
with
an
input?
You
do
the
calculations
and
you
go
in
and
you
add
in
some
new
time
series.
So
that
goes
like
her.
I
would
say
you
just
I'll
put
a
new
light
curve.
It
doesn't
know.
B
E
F
F
E
B
B
B
That
to
me
that
makes
perfect
sense.
You
have
you
have
an
object
that
you
pass
into
the
function
and
then
you
get
a
new
object
out.
Yes,
you've
got
cut,
yes,
you've
copied
the
sum
of
the
data
over
again,
but
then,
if
you
want
to
say
past
that
light
curve
into
plot
or
something
all
the
day
is
in
one
place,
I
wouldn't
want
to
have
some
columns
at
the
same
time,
series
in
one
object
in
some
columns
with
the
same
time
period.
Series
in
another
object,
ya.
E
B
B
B
E
B
E
E
Goes
like
curve
that
you
have
this
luminosity
data
in
there
like
nobody
can
go
up
and
look
up.
The
goes
like
curve
and
be
like
what
the
hell
is.
This
luminosity
is
doing
in
here
all
right.
Somebody
was
handed
a
random
good
liquor.
What
happens?
Have
those
luminosity
data
they'd
have
to
go
off?
Look
up
the
goes
like
curve
documentation
and
be
like
what
the
hell
is.
This
luminosity
data
in
here,
where.
B
D
B
B
Want
if
you
have
a
mouth
on
you
if
you've
read
in
the
Rose,
go
raw,
goes
data
file
you're
going
to
have
you're
not
going
to
have
this
information
unless
you
explicitly
call
this
function
on
it,
that's
right,
yeah!
Exactly
so,
you're
always
going
to
see
where
this
data
comes
to
these
columns
come
from.
They
can
repair.
E
Because
you
could
be,
somebody
could
then
go
off
and
write
a
new
function
that
takes
in
one
of
these
goes
like
curves,
but
they
expect
it
to
have
the
luminosity
data
and
it'll
look
like
to
user.
Oh
I
can
pass.
It
goes
like
curve,
except
no,
you
cannot
pass.
It
goes
like
that.
You
have
to
pass
it.
A
special
type
of
goes
like
curve
that
has
had
this
other
things
done.
It's
fine.
B
E
B
E
Why
do
we
have
these
like
curve?
Six
at
all,
right,
like
you're
saying
we
can
extend
the
art
based
objects
in
any
way
that
we
want,
and
you
know
people
just
have
to
get
used
to
it
like
what,
if
you
could
take
the
map
object
and
all
of
a
sudden
you
edit
it-
and
you
add
this
like
new
new
function
to
it
right
and
you're,
saying:
well,
that's!
Okay!
Well,
if
you've
returned,
if
I
take.
B
E
B
E
B
B
E
Are
black
box
because
you
know
people
will
not
are
going
to
use
this
function
like
sure,
nothing's,
technically
black
box,
because
they
could
go
check
on
what
this
function
is
doing.
But
you
know
we
shouldn't,
require
users
to
go
check
and
to
see
what
functions
have
been
run
on
particular
objects.
They
don't
know
how
to
use
them
in
the
future.
The
user
has
to
run
the
function.
No,
no,
no,
it
could
be
in
some
script.
It
could
be
some
other
function
calling
this
function.
You
don't
know
what
right
it
could
be
completely.
We
can
notification.
B
E
E
So
if
you
wanted
to
make
it
goes
object
by
default,
contain
luminosity,
then
we
can
go
and
add
the
script
there
and
it
does
it
there
and
that's
fine
what
this
should
return,
though,
if
we
do
not
want
to
include
there,
it
should
have
nothing
we're
related
to
the
goes
object.
It
should
be
a
new
light
curve
object,
not
a
goes
object.
E
E
B
Doesn't
in
the
metadata
because
he's
still
holding
the
original
goes
data
cuz?
It's
why
it
does
not
need
to
hold
the
original
go
thinner!
That's
what
I'm
are
there
again?
Nobody
shows
because
I
the
data
that
is
it's
taken
exactly
the
same
time,
I'm
from
exactly
this
is
like
calc,
it's
a
calculated
product
from
that
data.
That's
at
that
time
with
that
metadata
properties
with
all
of
the
information
that
comes
with
the
goes
object
the
luminosity
data.
B
Yes,
it
might
be
calculated
and
extra,
but
it
still
is
dependent
on
all
the
metadata
and
header
information
and
the
time
stamps
and
everything
that
is
the
goes
or
I
curve
object.
You
couldn't
have
the
luminosity
object
on
another
light
curve.
It
wouldn't
make
any
sense.
It
is
inherently
goes.
Yes,
it's
calculate,
but
it
is
still
a
property
of
a
goes
like
her,
because
it
can
only
be
calculated.
Trauma
goes
like
a
doom.
C
Anybody
edged
up
in
here
for
a
second
could
quite
some
time.
No
me
it's
my
understand
correctly
Steve.
The
issue
you
have
really
here
is
that
it's
consistently
right.
So
your
problem
is
that
you'll
have
some
goes
like
a
robeks
to
have
these
attributes
and
some
that
don't
and
there's
he
may
end
up
with
situations
a
priori
where
you
know
you
don't
really
know
whether
they're
there
or
not,
and
that
seems
to
be
like
a
design
problem.
C
C
That
calculates
some
sort
of
derived
quantity
and
attaches
that
to
your
map,
your
HMI
sub
map
in
the
same
way-
and
you
don't
know
necessarily
when
you're
writing
a
script,
whether
it's
going
to
have
that
property
when
you
run
the
script,
so
you
could
actually
do
this
for
for
any
kind
of
map
or
any
kind
of
light
curve
if
you
wanted
to,
but
maybe
we
don't
want
that
kind
of
design.
Is
that
kind
of
what
you're
getting
at
yeah.
E
C
B
C
C
B
To
worry
about
their
conceptually
attached
to
the
like
of
object,
like
all
the
map
objects,
they
are
conceptually
part
of
that
thing.
Uh-Huh
like
also
I
there
is.
There
is
a
slight
difference
in
my
mind
between
doing
this
to
a
map
and
doing
it
to
a
like
of,
like
I,
understand
that
they
are
comparable,
but
if
we
do
it
with
one,
we
should
probably
allow
it
to.
The
other
is
fine,
but,
to
my
mind,
adding
columns
to
tabular
data
and
adding
a
tribute
to
an
object
or
two
different
things.
C
E
Because
I'm
worried
from
the
users
perspective,
you
know
I
if
they're
handed
it
goes,
my
curb
they
should
be
able
to
go
back
and
see
what
what
things
are
right
if
they
see
if
their
hand,
it
goes
like
curve
object
with
this
luminosity
data
in
it
yeah
right.
They
don't
necessarily
know
what
that
is
or
where
they
came
from.
They'd
go
to
the
goes
like
curve
documentation
and
it
wouldn't
be
there
and
somehow
they'd
have
to
figure
out.
Oh,
this
was
calculated.
B
B
Anyway,
but
I
think
basically,
this
becomes
like
at
this
point
to
the
discussion
between
function
and
methods
becomes
a
API
design
decisions.
The
whole
of
the
library
which
is
do
we
want
a
functional
interface
where
we
write
functions
that
take
our
data
types
as
arguments
and
return
new
data
types,
which
will
be
in
some
way
modified
and
probably
the
same
type
as
the
one
that
was
past.
B
It
like
this
and
like
things
like
AI,
a
prep
for
instance,
rather
than
we
could
have
had
a
IA
preppers
and
a
method
on
an
AI,
a
map
object
which
I
out
of
hand
rejected,
because
I
want
the
feet
to
my
mind
in
map
and
also
in
like
oh,
maybe
not
so
much
at
the
moment.
But
when
we
refactor
it
thee,
there
should
be
no
difference
in
api
between
the
base
data
type
object,
be
it
map
or
light
curve
and
all
of
the
subclasses,
because
I
feel
about
it
more
confusing
to
the
user.
B
E
Oh
yeah,
I,
guess
I'm,
just
not
convinced
by
your
arguments
to
earth
that
I
mean
yes
this.
This
information
is
highly
related
to
the
goes
data,
but
just
because
it's
so
highly
related
does
not
mean
it
needs
to
be
packaged
together
with
it,
I
mean
you
could
extend
that
argument
to
like
any
kind
of
drive
quality.
Does
that
mean
it
will
stay
with
the
thing
that
generated
it?
E
I
mean
that
that's
not
really
feasible
right,
I
mean
you
could
think
of
you
know
maps
the
same
thing
right
like
if
I
generate
some
physical
value
from
the
pixels.
Does
that
mean
I
can
extend
the
map
and
add
a
new.
You
know
dot
data
feel
to
it
make
now
all
of
a
sudden.
My
data
is
multi-dimensional
and.
B
The
deranged
so
making
that
the
making
data
multi-dimensional
is
strictly
prohibited
by
virtue
of
the
fact
that
map
is
inherently
two
dimensional
data
type
right,
but
you're,
but
you're
increasing
the
dimensions
of
this
data
but
curve
is
like
her
visa
tabular
data
type
and
different
light.
Curves
have
different
number
of
columns,
and
that
is
there
is
no
restriction
on
that
in
the
same
way
that
there
is
in
that
map
by
design
is
a
two
dimensional
data
type.
Where
is
a
light
curve,
is
a
tabular
data
type,
but.
B
But
this
is
my
point.
My
point
is
that
the
guy
likes
to
understand
your
comparisons.
You're
making
to
map
and
I
do
to
some
extent
agree
that
there
is
a
valid
comparison
you
have
to.
You
have
to
appreciate
subtle
tea
in
the
difference
between
map
and
light
curve,
in
that
all
we're
doing
here
is
adding
columns
to
a
tabular
data
type
where
those
columns
are
like
logically
belong
in
that
table.
B
Having
that
data
in
that
table
makes
perfect
sense,
because
it
is
the
luminosity
the
x-ray
luminosity
at
that
time
and
attributed
to
both
dick
or
already
existing
columns
in
that
data
type.
To
my
mind,
having
these
extra
columns
is
just
like
in
columns
to
this
is
perfectly
packed
in
doing
it
in
pandas
data
frames,
doing
it
of
a
trip,
I
tables,
adding
columns
to
tabular
data
is
a
perfectly
logical
thing
to
do.
E
Except
that
you
don't
I
mean
the
lowes
guy
goes
like
her
provides
you
with
a
number
of
you,
know,
methods
a
and
it
will
not
be
able
to
use
like
plotting
it'll.
Not
it
won't
plot
this
depth,
because
you
know
the
ghost
light
curve
cannot
depend
on
it.
This
being
there
at
all
right,
like
the
goes.
My
curve
object
is
not
going
to
provide
you
any
useful
capabilities
for
these
columns.
E
B
That
is
an
inherent
usefulness
and
it
is
very
useful
from
from
a
user
perspective,
at
least
in
my
mind,
it
makes
I
would
Martin.
Oh
yeah
I
do
see
your
point
in
there
isn't
doing
anything,
but
I
suppose
you
could
write
a
specialized
subclass
of
goes
like
curve
specifically
to
do
it.
I'm
just
not
register
it
with
the
upstream
thing,
and
then
you
could
what,
if
I
plot
to
it?
Oh
my.
E
B
E
F
Of
this,
the
other
thing
to
be
said,
though,
is
that,
for
instance,
if
you're
looking
at
the
radiative
losses
as
opposed
to
the
x-ray
luminosity,
there
is
a
logical
sequence
of
the
of
the
calculation
as
well.
So
you
start
off
with
your
flux,
values.
You
then
have
to
calculate
a
temperature
in
their
mission
measure
and
then
from
including
the
temperature
memory
calculation.
You
then
get
to
the
goes
rated
losses
so
having
it
packaged
together,
like
that
means
that
it's
like
it's
rigidly
all
related.
F
F
E
I
mean
that
I
I
see
that
argument
as
well,
but
I
do
think
one.
We
need
to
come
up
with
some
sort
of
framework
for
how
we're
going
to
deal
with
this
with
like
curves
in
general
and
then
to
I,
actually
would
argue
that
this
is
so
fundamental
to
goes,
data
that
we
probably
should
just
put
it
in
that
goes
like
goes
object.
E
E
I'm
saying
if
we
are
gonna
accept
that
this
is
something
that
can
be
done.
I
think
we
need
to
write
up
some,
something
that
you
know
explains
when
and
how
this
is
going
to
be
acceptable,
because
this
can
get
quickly
at
go
out
of
hand
and
then
and
then
the
other
thing
we
don't
have
to
talk
about.
I'm.
B
E
B
Have
it
would
have
all
the
metadata
all
of
the
all
the
original
data
and
the
new
data?
The
only
difference
is,
it
would
be
a
lighter
object,
crowd
and
a
ghosts
like
Earth
object,
and
therefore
you
wouldn't
have
the
peak
method
specified.
This
is
made
specific
to
goes
basically
which,
as
Stephen
pointed
out
when
you
add
this
data
to
it,
is
made
of
relevant
by
virtue
of
the
fact
that
it's
not
going
to
plot
the
new
data
for
you
anyway.
D
B
B
E
B
E
B
A
B
Could
the
reason
we
keep
it
as
a
different
type
is
because
it
makes
it
the
reason
we
do.
That
is
because
that
was
how
it
was
written
at
the
beginning,
blame
Keith,
Nura
know
the
reason.
The
reason
it
makes
sense
is
because
it
makes
sense
to
put
the
aia
specific
data,
mangling
and
property
values
and
stuff
in
an
aia.
Object
like
having
the
mapping
what
we
map
specifically
there's.
The
subclasses
map
fits
keyword,
head
of
values
to
property
names,
and
the
subclass
is
specified
that
mapping
for
each
instrument.
B
B
White
curves
and
everything
and
then
well,
that's
not
that's.
My
point
is
why
we
have
specific
ones
to
do
the
data
mangling
we
we
we
should
maintain
the
API
in
that
we
do
practically
have
maps
and
like
curve,
because
with
the
factory
class,
that
is
how
it
appears
to
the
user
right.
They're,
not
calling
something's
pacific
for
a
specific
class.
They
just
called
the
factory
and
getting
back
an
object
right.
A
Right
but
I'm
just
saying
if
it
was
clear
you
know
when
you
look
at
the
tight
that
says
AI
a
map
rather
than
just
map,
there's
clear
that
every
type
of
map
was
just
a
nap.
Then
this
notion
of
having
the
common
API
is
pretty
much
locked
in
I
mean,
like
you,
can't
add
a
bet.
It's
just
a
a
bath
because
you
don't
get
AI
a
nap.
So
just
get
that
yeah
sure
and.
B
I,
don't
like
you
agree
with
what
you're
saying
I
mean
that
would
be
a
valid
approach,
but
it
would
cause
like
there
would
be.
There
are
issues
we
do
like
sure.
These
are
real
harder
and
we
have
a
basically
the
whole
point
of
subclasses.
Is
that
you
you?
The
whole
point
of
the
object
orientated
programming
is
that
you
have?
Yes,
you
make
a
further
newsmaker
object,
more
specific
to
handle
a
thing,
and
it
can
act
like
it's.
This
subclass
should
be
a
drop-in
replacement
for
the
super
class
I.
You.
E
Know
I
actually
think
that's
a
kind
of
an
outdated
view,
object,
oriented,
programming,
cuz
now,
at
least
at
least
on
the
objective-c
side.
So
you
know
I
I,
don't
speak
with
a
lot
of
experience
here,
but
there
people
are
very
often
now
like
mixing
classes,
so
you
can
have
two
super
classes.
If
you
wanted
to,
for
example,
pieces.
E
But
it
doesn't
mean
that
we
have
to
tie
ourselves
to
that.
I
mean
the
way
I
think
of
it
a
superclass.
So
the
way
I
always
think
about
it
is,
you
know,
super
class
is
what
is
common
to
all
of
the
objects
right.
So
you,
if
you
have
all
the
child
objects,
the
superclass
is
what's
common
to
all
of
them,
but
each
of
the
children
can
have
their
hope.
Xq
thing.
Second,
they.
B
Can
and
that
isn't
isn't
a
yes
in
having
an
extra
methods
on
the
Sun
I,
a
map
wouldn't
break
the
0
/
principle,
because
it's
not
that
the
subclass
shouldn't
the
the
subclassed,
the
superclass,
doesn't
have
to
be
a
drop-in
replacement
for
the
subclass.
It's
the
other
way
around
this.
The
subclass
has
to
be
a
drop-in
replacement
for
the
super
class.
E
E
A
Think
our
current
use
of
subclasses
these
it
hard
to
understand
why
we
can't
do
what
Steve
is
talking
about
if,
instead,
we
were
returning,
not
all
these
crazy
subclasses,
but
really
just
the
base
class,
then,
like
I,
said
it
locks
in
this
notion
that
it
really
is
fundamentally
the
same.
A
single
API
prevent
it
yeah.
A
A
B
First,
read
like
what
he's
I
like
the
idea
that
greater
basically
I
would
my
opinion
is
that
we
should
the
library
should
have
a
more
functional
interface,
but
I
am
willing
to
just
see
like
willing
to
listen
to
counter
arguments
to
this,
because
it's
just
my
opinion
but
I
certainly
disagrees
with
me.
Then
fine
I,
certainly.
A
Like
functional
interface
is
more
for
stuff,
where
it
can
be
where
people
might
want
to
try
sightly
different
takes
on
stuff.
It's
hard
to
modify
methods
have
been
objects.
If
there's
a
calculation
done
a
way
that
you
don't
like
or
calculation
won't
do
really.
My
treat
the
calculation
plumber
to
modify
a
class
method
rather
is
calling
a
different
function.
So,
if
there's
a
function,
interface,
it's
easier
to
just
like
take
the
existing
function,
modify
yeah.
A
F
Bringing
it
back
just
as
you
leave
to
the
specifics
of
what
started
this
whole
conversation.
What
what
about
coming
out
of
this
this
meeting?
What
are
we
saying
about?
What
we're
going
to
do
with
this
going
here?
Are
we
gonna
modify
we're
gonna?
Let
it
go
as
it
is.
Are
we
going
to
put
on
the
lung
finger
again
for
a
while?