►
From YouTube: SimPEG meeting Nov 14
Description
Bane Sullivan (Colorado School of Mines) shares his work on PVGeo, an open-source framework for visualizing geoscientific data and models in VTK powered platforms like ParaView.
The PVGeo Python package contains VTK powered tools for data visualization in geophysics which are wrapped for direct use within the application ParaView by Kitware. These tools are tailored to data visualization in the geosciences with a heavy focus on structured data sets like 2D or 3D time-varying grids. As an effort to bring interoperability between SimPEG projects and PVGeo, Bane will be demoing his work in SimPEG/discretize MR#114 to create an interface between discretize meshes and VTK data objects.
C
D
C
A
C
C
D
D
C
D
You
know
we're
good
to
go
with
it.
Then
I
want
to
focus
a
bit
on
that.
They
interface
and
discretize
just
run
you
through
how
you
could
use
the
API
of
requests,
maybe
talk
about
where,
where
to
go
from
there
and
didn't
focus
on
what
that,
for
that
base
allows
us
to
do
in
discretizing
TV
to
you
and
sort
of
how
we
do.
D
How
do
we
prepare
a
video
and
digitizing
a
Python
environment,
and
not
only
can
you
do
that
also
just
literally
TK
algorithms
and
then
I'll
finish
it
up
with
just
in
general
examples
of
how
we
use
video
and
what
that
looks
like,
and
maybe
some
features
and
EEG.
Oh,
that
might
be
of
interest
to
you
all
working
with
this
impact
projects
and,
of
course,
feel
free
to
just
interrupt
me
at
any
time
and
ask
questions
there.
D
Keep
it
casual
the
first
PBG.
Oh,
it's,
it's
just
a
Python
package
for
creating
like
really
compelling
and
integrative
visualizations,
both
in
3-d
and
4-d.
So
we
do
this
all
in
in
care
of
you
right
now
and
the
whole
libraries
built
upon
dt
k,
which
is
incredibly
scalable
and
well
maintained,
and
so
the
whole
purpose
of
why
we
chose
everything
off
BTK
is
we
didn't
really
want
to
reinvent
the
wheel?
D
Btk
and
pair
of
you
are
already
who's,
really
robust
scientific,
visualization
libraries,
and
so
we
might
as
well
harness
on
this
common
visualization
tools
or
Geoscience
data,
because
our
Geoscience
data
isn't
really
special
thanks.
You
know
it's
foxholes
its
points,
that's
all
about
so
a
role
that
we're
trying
to
have
PPG
of
though
it
is
just
be
this
link
between
all
the
Geoscience
and
this
and
this
VT
can
a
pair
of
you
realm
of
data
visualization,
and
so
you
allows
us
for
things
like
our
verdict
models.
D
Our
geophysical
data,
like
the
observed
data
in
space
and
just
other
geo
scientific
information
that
is
spatially
referenced.
We
have
all
these
separate
entities
in
geo,
so
tiring
go
from
software
platform,
a
software
platform
you
them
and
what
we're
trying
to
impede.
You
is
just
find
a
way
to
bring
all
of
it
together
in
one
rendering
environment
that
is
geo
referenced,
so
maybe
do
takes
all
of
that
and
it
types
of
innings
it
all
into
data
structures
into
air
view,
and
and
then
it
allows
us
to
take
all
of
our
geosciences
into
new
domains.
Such.
C
D
K
KS,
which
is
sort
of
a
javascript
library
for
web
visualization
of
BTK
data,
objects,
the
new
structures
it
allows
us
to
take
all
of
our
integrated
scenes.
So
as
soon
as
we
have
something
impaired
you
visualized
pair
of
you,
it
allows
us
to
take
all
that
into
virtual
reality,
which
we've
been
doing
a
lot
of
work
with
here
at
school
minds
and
then
just
other
platforms
at
Caribou
web
institute
processing
with
really
catalyst
all
types
of
high
performance,
algorithms,.
D
D
D
Just
has
some
you
know
overview
since
that's
a
few
quick
examples
about
how
to
use
or
what
can
be
created
with
video,
well
I
hope
my
streams
not
too
low
for
you
all,
but
we
have
these
these
interactive
visualizations
and
these
are
all
made
using
the
VT
kgs.
Travellin,
that's
cool
yeah.
So
as
soon
as
we
take
our
visualizations
in
kit
review,
which
we
you
know,
we
may
need
using
the.
C
D
Just
proxies
a
Dropbox
URL
for
the
data
files,
that's
sweet,
yeah
BTS
is
still
in
its
early
stages
of
developments
with
you
know
they
create
these
these
fairly
simple
renderings
or
we
don't
really
have
being
right
included
system
or
anything.
You
know
publication
quality,
but
it's
a
good
way
to
way
we
use
it.
Is
we
just
we
make
our
renderings
real
quick
and
we
Pina
non
slack
over
to
someone
else
in
a
research
group
like
hey.
Does
this
look
weird
to
you
or
what's
going
on
here,
check
this
out?
D
That's
just
a
good
way
to
share
data
fast
yeah
and
the
website.
We
have
links
to
do
the
Vimeo
channel,
which
is
some
general
videos
of
how
do
you
use
our
readers
there's
so
I
just
off,
we
have
a
slack
workspace,
please
feel
free
to
join
it.
We
try
to
keep
it
on
DVD,
do
users
or
just
anyone
using
care
of
you
in
the
realm
of
Geoscience.
Just
you
know
a
place
to
ask
questions
and
get
feedback
person
polarizations.
D
D
And
so
the
this
webpage
runs
you
through
how
you
could
use
in
Python,
environment
and
then
really
what's
important,
is
how
we
use
them
in
interview
explicitly
so
like
this
one
right
here
we
just
open
up
the
UVC
octree
format,
goodne
from
sim
peg.
You
will
help
me
in
whom,
at
this
octree
reader
and
we
just
file
open
just
like
that,
choose
the
off
tree
reader
and
I
cut
out
a
little
bit
of
lag.
It
took
dilute
the
file,
but
it
just
constructs
a
BTK
data
object
in
here
and
you're
going
immediately.
D
The
moment
we're
really
doing
all
at
post-processing
sort
of
analysis
and
trying
to
keep
it
all
visualization,
because
that's
what
we've
been
able
to
build
up
instantly
as
far
as
like
really
trying
to
do
any
type
of
true
to
physical
processing,
the
Bloor
anything
like
an
inversion.
We
just
don't
have
that
implemented
because
we'd
need
this
to
write
interfaces
back
channel
very
like
stimpack,
rather
than
trying
to
build
in
like
two
processing
algorithms
and
key,
be
cool
yeah.
D
And
that's
that's
really.
What
I
want
is
I
want
to
find
a
way
that
we
can
link
pair
of
you
and
PTO.
So
that
way,
you're
interviewing
you
and
have
your
data
sources
in
the
render
environment
and
then,
through
the
filters,
many
go
and
apply
a
filter
that
funds,
a
quick
algorithm
on
your
model
space
to
do
some
sort
of
processing
and
falls
back
to
yonder
or
synthetic.
And
then
the
results
come
back
to
pair
of
you
rather
than
trying
to
implement
all
that
in
pgo
or
imperial
I.
D
D
D
I'm
looking
forward
to
it
a
ton
just
get
the
work
out
there
and
meet
tons
other
people
and
figure
out
some
other
processing,
libraries
or
visualization
libraries-
how
we
can
link
it
all
together,
it's
going
to
be
exciting
yeah,
so
the
other
big
takeaway
we
push
a
lot
here.
Is
this
virtual
reality
here
view?
Has
some
native
VR
functionality
where
you're?
Just
in
pair
of
you,
you
had
a
training
environment,
you
just
click
a
button
that
says
send
to
VR,
and
you
you
put
on
your
VR
headset
and
it's
there.
D
So
that's
that's
sort
of
my
overview
of
PD
Geo.
What
we
can
do
with
it,
why
we're
doing
it,
and
all
that
and
I
want
to
move
on
to
talking
about
the
the
pull
request
I
made
and
discretize
to
get
the
vtk
interface
between
SEO
types,
mesh
objects
in
TK
data
objects
and
sort
of
just
talk
about
how
you
can
use
that
in
your
find
on
environment.
B
Just
before
we
kind
of
get
down
to
like
detailed
stuff,
how-how-how
is
supposed
to
use,
because
it's
more
like
this,
it's
more
like
that.
If
you
have
some
sort
of
input
file
after
that,
basically
we're
using
pair
of
you
or
you
have
more
interactivity
like
using
some
on
script.
And
then,
if
you
run
the
function,
something
pops
up
or
what's
the
sort
of
scope
of
PBG.
D
Yes,
so
at
the
moment,
it's
really
all
about
you
have
your
data.
You
have
a
process
to
have
your
working
directory
of
sort
of
everything,
you're
running
and
you're,
your
processing
software,
whether
that's
you
know
synthetic
some
commercial
software
and
you
have
your
intermediate
files,
your
working
files.
The
whole
thing
we're
focusing
right
now
is:
let's
make
TV
geo
work
for
your
working
data
formats.
That
way
you
don't
have
to
convert
or
make
output
files.
D
You
just
you
run
your
your
your
your
software,
your
urgence,
all
that
in
your
working
directory,
you
have
your
project
and
you
don't
do
anything
extra
really.
Besides,
maybe
saving
out
intermediate
results
at
pjo,
King
and
see
it
can
look
at
and
immediately
render
and
care
of
you
and
that's
what
we're
gonna
focus
in
on
right
now,
but
there
is
definitely
a
way
where,
if
you
use
curve
use
Python
interpreter
where
you
can,
you
can
link
up
your
virtual
environments
in
a
way
that
you
can
use
anything
in
a
normal,
Python
environment.
D
You
might
such
as
impact
on
for
your
processing
and
then
use
tbo
and
all
of
the
rendering
functionality
of
interview
the
meter
processing
and
then
make
a
call
to
visualize
that
data
structure
immediately
I'm
gonna
have
a
little
window
pop-up.
So
there
are
ways
to
do
that
right
now.
We
can't
we
don't
have
a
stable
way
of
doing
that
in
jupiter
notebooks,
but
that's
definitely
what,
if
the
next
big
item
on
our
on
our
agenda,
it's
you
know.
E
B
Let's
suppose
you
got
quite
kind
of
different
data
set.
You've
got
a
service
data
for
whole
data
and
probably
some
voxel
models
in
3d
yeah,
yes,
and
space
k.
If
like
I
need
some
sort
of
python
array
and
then
at
somehow
using
pb0
I'm
outputting
those
file
formats,
like
paraview
format
using
PVG,
oh
and
I'm
I'm
reading
in
that,
is
that
what's
gonna
happen,
if
I
want
to
generate
that
image
with
the
PPG?
Oh
that
right.
D
No
I
mean
you
could
do
it
that
way,
but
that
would
be
a
bit
cumbersome.
The
sort
of
what
we
want
you
to
do
is
just
if
you
have
you
know
if
you
have
a
working
data
format
like
the
UVC
mesh
and
model
formats.
Okay,
I
have
final
readers
than
Keaney
geo
right
now,
but
read
that
directly
I'm.
So
there
won't
be
any
data
conversion
but
yeah
if
you're
in
a
Python
environment
right
now
and
you're
running
your
inversion,
and
you
have
sort
of
your
inverted
model
as
a
Empire
array
in
your
live
environment.
C
D
Will
have
to
save
out
that
a
model
array
to
to
a
file
format,
whether
that's
you
know
your
UBC
mission,
motto,
format
or
just
anything
else.
If
it's
a
binary
file,
we
have
ways
to
read
and
just
ran
them
back
random
formatted
binary
files
and
make
a
data
structure
out
of
it,
so
we're
trying
to
make
it
as
universal
as
possible.
Okay,.
D
Cool
and
I'm
gonna
I'm
gonna
step
step
into
that
a
bit
more
after
I
talk
about
this
discretize,
the
ek,
mixin
and
I'm
gonna
talk
about
sort
of
now
that
we
have
this
vtk,
mixin
and
and
in
discretize.
What
that
enables
TV
geo
to
do
so.
I'll
go
I'll
circle
back
to
that
in
a
minute.
So,
overall,
this
this
whole
vtk
pull
request.
D
Otherwise,
this
this
function
is
just
simply
not
available
and
it
would
be
available
on
any
mesh
object
that
you
have
so
tensors
trees
and
the
curvilinear
matches
are
just
called
two
vtk.
It
handles
understanding
which
mesh
object
you
have
and
then,
which
data
structure
we
need
to
make,
and
so
just
an
example.
Here,
I
call
mesh
TV
a
on
my
tensor
mansion.
It
makes
a
BTK
rectilinear
grid.
D
The
two
are
pretty
much
synonymous
so
that
the
vtk
rectilinear
grid.
It
really
is
just
exactly
how
you
all
define
your
tensor
mesh
just
enough.
Etk
data
objects,
but
I
also
added
in
this
cool
request.
The
ability
to
define
your
reference
fame.
So
here
I
just
decide
to
rotate
my
reference
frame
a
little
bit
just
redefine
my
x-axis
direction,
my
y-axis
direction
and
same
with
a
Zed.
Keep
it
positive
or
keep
it
the
same
and
I
just
validate
it
make
sure
they're
all
whatever,
where
it
is
no.
D
Though
oh
yeah
and
so
I
rotated,
my
my
tensor
mesh
now
and
so
everything
on
the
back
inside
of
this
critize
right
now
remains
the
same.
We
just
have
this
reference
frame.
We
can
use
if
we
need
it
and
the
two
vtk
functions
use
it,
and
so
now,
when
I
call
mesh
to
vtk
on
my
tensor
mesh
object,
it
sees
hey.
That's
that's
a
rotated
mesh,
let's
not
use
a
vtq
rectilinear
grid,
because
that
doesn't
make
sense
in
BTK.
Let's
use
a
vtk
structured
grid,
and
this
would
be
the
difference
between
the
two.
D
C
D
It
all
it
would
be
is
just
a
little
bit
more
matrix
rotation.
It
wouldn't
be
hard
to
implement.
So
yes,
now
we
have.
This
really
simple.
Api
for
training
is
BTK
data
objects,
so
we
could
use
this
in
PvP.
You
know
we
could
use
it
in
other
gtk
powered
Python
environments,
and
so,
if
you
have
you
know
your
BTK
rendering
scripts
for
this
a
standard,
Python
environment,
then
all
you
would
have
to
do.
It
just
bothers
to
be
deeply
and
pass
that
data
object
on,
and
so
now
I
want
to
talk
about.
B
D
C
D
Then
we
can
use
a
plug-in
I've
made
in
PVG,
so
it
calls
dis,
critize
and
it
says:
okay,
I'm
gonna
have
a
file
name,
I
want
to
use,
let's
use
dis,
critize
all
their
file
IO
for
this
format,
because
it's
their
format,
I
loaded
in
as
their
data
object.
So
district
ID
is.
This
is
a
tensor
mesh
object
when
it's
loaded,
but
PTO
is
sort
of
indifferent
about
what
that
data
object
is,
and
then
we
just
pass
that
object
onto
this.
D
Get
vdk
data
object,
function
and
TV
geo
just
handles
the
rest,
and
so
this
passes
on
my
I
I
discretized
mesh
as
a
vtk
data
object,
which
enables
me
to
use
pair
of
you
to
just
a
file
open,
and
then
you
just
navigate
to
where
you
have
your
UVC
mesh
format
in
it.
Okay
and
I
just
collect,
let's
use
that
UVC
tensor
mesh
reader
and
I
can
also
choose
model
files
if
I
need
to.
So
we
have
that
Oh
pop.
E
B
Got
a
mesh
and
model
as
a
file
mm-hmm
then
you're,
basically,
loading
up
and
in
pair
of
view,
that's
what's
that.
Why
do
I
need
you
I,
wasn't
sure
what's
actually
happening
at
like
I
said
when
you're
doing
messed
up
vtk?
Is
that
actually
writing
the
files
or
it's
simply
you're,
generating
the
object?
Yeah
sure
we're
like
soon
as
I
get
the
end
of
the
herbs.
You're
writing
some
sort
of
files
right.
D
So
I
already
had
those
files
saved
on
my
device
so
right
here
when
I
call
mesh
to
vtk
I,
create
that
vtk
data
object
in
my
rendering
environment.
So
if
I
just
assign
that
object
and
I
can
make
calls
on
it
like
it's
a
live
object
in
my
environment,
and
so
this
is
a
data
structure
with
all
of
its
all
this
stuff
going
on
or
we
could
write
out
those
vtk
data
formats
by
mesh
right,
TG,
right
vtk,
which
is
also
implemented
in
that
pull
request.
D
But
I
just
don't
really
see
a
need
to
use
that
I
think
it's
better
to
use
your
UVC
for
now
you're
working
processing
formats,
rather
than
these
static
copies
of
a
vtk
format,
because
then
you
get
sort
of
messy
with
having
random
versions
or
static
versions
of
data
that
isn't
up
to
date
with
your
processing,
routine
I'm.
So
jumpin.
C
B
B
B
D
You
definitely
right,
like
these
VTA
files.
I
put
these
in
here
just
because,
just
because
they're
easy,
and
just
because
we
should
have
them,
but
you
definitely
don't
need
to
use
them
and
I
wouldn't
recommend
it.
You
should
you
should
keep
your
data
in
your
data
format,
for
the
things
that
are
most
convenient
to
you
and
to
your
processing.
It.
B
D
D
So
I
put
this
example
together
from
a
data
set
neg
website
we're
just
already
had
some
mesh
expertise
meshes
and
all
that
and
a
topography
surface,
and
so
I
wanted
to
run
through
this
example,
where
I
just
I
load
in
my
my
mesh,
you
can
create
this
in
your
Python
environment.
However,
you
want
to
make
your
mesh
and
then
we
have.
This
live
mesh
object
from
different
eyes.
There
was
also
a
process
modeling
just
because
they
give
us
a
little
bit
of
visualization
to
see.
What's
going
on
and
I
want
to
run
through
this
example.
D
Okay,
now
I
have
my
mesh
have
saved
to
poverty
file
as
well,
which
is
this
file
right
here,
my
working
environment.
So,
let's
just
say
all
I
have
is
my
model
space
which
maybe
I'm
tweaking
and
trying
to
design
to
make
it
better.
I'm
gonna
have
my
typography
surface.
So
what
can
you
do
in
a
in
a
Python
environment
to
play
around
with
that?
What
I've
done
is
I've
created
this
little
example,
where
we
read
in
our
topography
file,
which
is
maze
and
XYZ
values,
and
in
pvt.
D
D
Here,
right
here
to
just
read
in
that
topography,
naver
I
have
my
BTT
poly
data
and
then,
let's
say
I
want
to
use
that
to
poverty.
Information
which
is
right
here
above
my
service
and
I,
have
my
model
space
which
may
be
that
model
space
Billy's
a
bugger
fee.
Let's,
let's
then
create
a
way
like
a
let's
create
an
active,
so
array
of
whether
or
not-
and
you
can
even
spell
and
that
mesh
above
or
below
the
typography
surface.
We
can
hit
that
right
back
out
into
Python
or
into
numpy
arrays
use.
D
So
that's
what
I
do
here?
I
just
called
ki
BPO
in
this
Chris.
We
listen,
extract,
typography,
filter
and
I.
Take
that
mesh
to
V
okay,
but
get
that
BTK
object
of
your
mesh
from
disparate
eyes
and
I
passed
that
topography
file.
That
departed
follows
which
straight
from
file
and
I
just
run
this
algorithm.
So
now,
I
have
another
VT
data
object.
Here.
D
And
so
here
we
see
there's
my
topography
line
top-down
view
of
the
valley,
and
so
it's
just
a
quick
way
to
run
any
you
know
any
TV
geo
or
even
a
vtk
algorithm,
get
the
results
over
PVG
toes
of
BTK
interface
and
use
it
again
in
your
district
ice
code,
and
so
this
is
a
really
simple
example
and
I
think
you
all
have
some
function
like
instant
bag
or
discretized
of
doing
this
kind
of
thing.
But
I
don't
want
you
to
focus
so
much
human.
You
know
it's
only
used
for
extracting
fog
review.
D
The
whole
point
is
now
that
we
have
this
mesh
to
VT
today,
method.
We
can
interface
the
discretized
code
with
the
ppto
code
so
that
they
can
talk
to
each
other
and
share
the
results.
This
active
topography
array,
or
maybe
something
else,
and
we
can
use
it
in
discretized,
so
this
doesn't
only
hold
for
PBT.
Oh
I
was
like.
D
We
could
also
use
this
for
vtk
algorithms,
so
just
using
the
VTA
Python
suite,
here's
a
really
simple
filter
on
the
cell,
size,
filter
and
and
so
as
soon
as
I
have
my
my
BTK
algorithm
I
just
said
in
my
input
date
object.
Has
the
mesh
etk
object,
I'm
straight
from
discretize
update
my
algorithm
to
the
outputs.
You
see,
video
on
the
interface
and
they're,
sharing
the
results.
C
D
E
D
C
D
D
B
D
D
So
I
think
that's
yeah
most
exciting
thing.
Here
we
can
bring
together
all
these
libraries
and
just
sort
of
the
wrap
everything
up
there
on
the
website.
There
is
tons
of
examples
which
I'm
trying
to
fill
out
as
best
I
can
but
I
just
put
a
few
here.
That
might
be
interesting
to
you
all
for
sort
of
the
data
types
and
structures
you're
working
with
so
in
P
video
I
just
made
this
UVC
suite
has
file
I/o
for
the
tensor
meshes
tree
meshes
as
well
as
for
model
data.
D
That
might
be
time
varying,
and
this
allows
you
just
you
know.
If
you
have
all
of
your
UVC
files
already
saved
out,
you
just
get
a
little
bit
impaired
of
you,
you
can
make
these
really
integrative
visualizations.
You
know,
along
with
other
other
information,
so
maybe
you
have
other
geospatial
information
given
to
you
in
some
other
data
format,
PV
yukio
or
a
pair
of
you
can
read
that
in
and
just
put
it
in
your
environment
right
alongside
all
the
over
working
anymore.
That's
so.
B
D
D
So
if
you
go
to
the
application
contents,
wherever
you
download
pair
of
you,
there's
a
little
executable
called
PV
Python,
it
allows
you
to
use
all
of
the
rendering
functionality
and
pair
of
you
in
a
Python
environment.
So
I
can
set.
The
color
map
like
and
I
can
say,
hey
apply
a
slice
here
and
just
automate
these
functions,
and
you
know
you
could
use
pgo
directly
in
that
Python
environment
to
automate
any
of
the
filters
we
have
for
creating
different
types
of
slices
and
all
types
of
stuff.
C
So
let's
keep
in
touch
on
that
and
just
like
keep
that
on
your
radar
that
I
will
be
there
and
excited
to
make
this
happen.
Yeah
yeah
and
then
I
want
to
just
get
your
thoughts
and
sort
of
trajectory
on
the
Python
to
seven
to
Python
three
migration,
with
pair
of
you
and
cuz,
I'm
just
sort
of
on
the
out
of
the
loop
on
how
that's
happening
and
what
the
timelines
are
for.
That.
D
C
D
And
I
don't
have
totally
time
off
the
top
of
my
head
or
we
in
care
using
a
drop
to
seven
thing.
Work
I
think
it's
scheduled
for
their
next
major
release,
which
would
be
in
the
summer
I.
Think.
Okay,
that's
that's
the
timeframe
on
it
in
mind,
but
I
could
double-check
and
try
to
keep
up
on
earth,
but
I
know
like
with
an
entire
community,
there's
a
whole
bunch
of
people
there
begging
for
it.
D
B
So
just
like
a
kind
of
writing
the
files
and
like
if
I'm
gonna
have
a
Python
there
a
and
write
the
file
and
then
read
it
of
you
like
using
Python
through
there's,
no
problem
right
and
then
there's
different
types
of
data
like
a
surface
data
or
point
data
and
like
for
whole
data.
Yeah
like
the
PB
geo,
has
all
functionality
that
I
can
write
out
as
a
vtk
file
format.
Is
that
correct,
yeah.
D
Yeah
so
yeah.
So
if
you
have
all
of
your,
you
know
your
borehole
data
playing
like
the
open
mining
format,
I've
been
working
a
bit
and
it's
not
on
a
master
branch
but
like
you
can
read
in
the
open
when
I
get
your
whole
data
visualized
in
here,
and
you
can
either
save
it
back
to
the
omf
omf
format
or
you
could
save
it
to
a
vtk
format
or
something
else.
C
D
Yeah
I
don't
have
much
questions,
I,
guess
the
main
thing
I
don't
think
about
the
other
day.
Is
you
know
whether
or
not
we
should
rotate
around
the
origin
and
that's
pretty
simple
when
we
do
those
rotation
things,
the
thing
that
I
tried
the
play
around
with
was
getting
the
rotation
code
within
the
grid
code.
So
when
you
call
grid
nodes,
whether
or
not
it
rotates
Internet
I've
had
some
trouble
with
that.
I
didn't
know
where
everything
was
making
calls
to
all.
D
B
C
D
The
bad
stuff
and
I
ran
into
a
ton
of
validation
issues:
okay,
that's
just
which
I
put
a
little
issue
on
discretize,
but
it
there
is
something
back
in
the
I
think
in
the
C++
code
for
the
tree,
men
should
I
just
got
lost
in
life,
though
I
think
it
looks
good,
just
definitely
check
it
out,
whoever's
interested
and
leave
some
comments
and
some
feedback.
If
it
doesn't
look
right.
D
C
C
D
Now
I've
put
together
like
a
really
lightweight
small
edges
like
off
to
the
side
for
reading
Landsat,
imagery
so
I,
this
Python
package,
where
we
can
just
read
in
Landsat
imagery
straight
off
earth
Explorer,
which
is
the
us's
geologic
surveys,
data
portal
yeah
that
data
and
then
this
Python
package,
I
made
just
reason.
Those
formats
creates
a
nice
object
metadata
with
your
own
set
of
imagery
and
image,
and
then
I
create
a
little
interface
for
kami
geo.
Just
make
calls
back
to
that
interface,
so
I
haven't
really
done
this
with
anyone.
D
C
D
C
But
the
goal
is
to
try
and
figure
out
like
if
we
have
a
giant
body
like
a
aju
that
is
supportive
of
open
source
like
what
should
it
be
doing
and
so
yeah
I
I,
think
it'll
be
an
interesting
conversation
and
hopefully
well
I
know
they'll
be
quite
enough
number
of
other
sort
of
developers
that
are
fake
in
the
open,
Geoscience,
Python
communities
so
I'll
be
sure
to
we'll
make
some
connections
and
do
all
that
then.
So
we
good
yeah.