►
From YouTube: SimPEG meeting Jan 15, 2020
Description
SimPEG meeting on January 15, 2020
A
B
B
A
B
A
B
B
B
B
C
Sure
so,
I've
just
been
kind
of
going
through
and
working
with
Doug
and
going
through
the
notebooks
for
the
EOPS
I.
Just
boarded
project
went
through
some
stuff,
I'm
sure
you
guys
are
working
on
that
and
then
I've
been
trying
to
clean
up
and
tackle
some
of
these
bugs
in
the
simulation
branch.
That
I
don't
think
anyone
else
is
looking
at
just
yet
so
like
some
of
the
things
that
was
causing
the
flow
simu
flow
test
to
fail,
I'm
a
few
other
things
and
then
trying
to
get
a
start
on
updating
some
of
the
documentation.
D
A
That
so
there
wasn't
a
pull
request
automatically
generated.
You
just
need
to
couple
check
and
merge
it.
This
is
just
the
first
time
I've
done
that,
so
we
haven't
updated
it
yet
I'm
your
adding
Joe
throughout
to
all
of
our
Conda
Forge,
all
of
our
kind
of
Forge
packages
and
Tebow.
Actually,
since
you
have
the
pie
pie
right,
would
you
be
interested
in
joining
that
as
well
sure.
A
I
mean,
and
so
there's
a
PR
generated,
and
we
can
walk
through
this
with
the
simpad
one,
maybe
I'll
just
ping,
both
you
and
Joe,
and
we
can
just
take
a
look
at
this
in
peg
one,
maybe
even
just
right
after
the
meeting
and
just
kind
of
quickly
go
through
that
yeah,
because
then
it's
basically
like
once
you're
added,
you
have
rights
to
merge
the
flow
requests
and
the
only
thing
you
should
have
to
do,
or
the
only
thing
the
only
manual
thing
you
should
have
to
do
is
update.
If
there's
any
dependency
changes.
Okay,.
E
E
Yeah,
but
mostly
I,
guess
what
I've
been
working
on
I
think
you
saw
that
I
made
the
2d
forward
simulation
available
for
the
the
Myanmar
notebooks
and
one
I
think
I'm,
just
following
what
you
have
been
working
on,
it
looks
like
john
has
been
continuing
to
to
look
at
those
tests
to
make
sure
that
travis
passes.
So
that's
that's
wonderful.
It
looks
like
you
guys
are
working
on
all
the
stuff
that
I
was
hoping
you
would
be
working
on.
E
So
you'll
notice
in
that
that
that
notebook
that
I
provided
that
I
actually
compute
the
data
as
a
voltage
then
use
the
survey
to
go,
get
the
geometric
factor
and
convert
it
into
a
parent
resistivity
before
I,
plot
and
I.
Do
that
after
I
run
the
simulation.
So
there's
some
kind
of
bug
in
there,
where
you
can't
directly
define
the
receivers
to
measure
apparent
resistivity.
So
that's
something
that
I
think
needs
to
be
fixed
but
that
that's
more
or
less
it
yeah.
A
G
E
A
B
H
E
K
J
A
F
F
I
A
I
I
I
If
you're
trying
to
put
in
elevation
I,
don't
know
what
the
plan
is
for
using
that
data,
but
I
will
also
put
up
I
have
to
extract
from
the
from
the
report
a
few
things
for
you,
so
that
you
can
see
what
the
interpretation
was
plus.
There's
some
poor
hole,
information
too
and
I'll
try
to
do
that
tomorrow
and
I
think
that's
it
for
the
2d.
If
I
can
attend
in
a
slightly
different
direction,
talking
with
Devon
we're
running
into
the
glitches
with
the
visas
right
now
and.
C
I
E
I
E
Yeah
yeah
and
that
can
be
processed
I
think
within
three
business
days.
So
my
last
my
last
visa
application
was
rejected,
I
mean
I,
contacted
them
and
said
I,
don't
know
where
I'm
staying
yet
I'm
part
of
a
visiting
group
from
a
university.
That's
basically
all
I
said,
and
but
they
get
on
it
pretty
quickly.
They
look
at
it
quite
quickly.
So
I
could
do
that.
I.
Think.
F
E
H
F
I
Yours
show
you
arriving
on
Friday
and,
of
course
we
leave
on
Friday,
but
we
don't
get
it
until
Saturday,
because
the
date
change,
Sookie,
yep
so
but
I
can
send
an
email
to
you
with
suggestions
for
the
hotel
or
even
send
the
hotel
or
booked
the
hotel
for
you,
and
then
that
would
be
a
convenient
one.
Oh
yeah.
E
I
E
A
One
one
question
with
respect
to
the
data
sets
that
you
sent
max:
is
there
somebody
who's
actually
championing
that
right
now,
I
want
to
make
sure
that
doesn't
get
sort
of
like
lost
in
the
void
of
you
know,
sending
it
to
four
different
people
in
an
email.
So
is
there
somebody
who's
willing
to
take
that
on
for
a
champion
working
that
through
yeah.
C
I
Process
set
C
is
the
software,
that's
free.
You
can
download
it
from
the
Irish
Center
just
type
in
process
process
and
then
you'll
very
quickly
see
that
you
can
actually
dump
it
into
UBC,
but
I'll
dump
it
into
the
UBC
format.
I,
don't
think
the
EBC
format
has
top
typography
in
it,
though
right,
so
you
have
to
bring
the
pharmacy
and
separately
I.
Think
there
is
a
there
is
a
typography
change
in
it,
but.
I
I
G
I
I
C
A
Just
a
couple
of
quick
updates
on
discourse
and
also
just
tracking
a
few
things
for
the
GWB
project,
and
then
we
can
also
go
back
if
there's
any
other
longer
discussion
items,
but
I
just
wanted
to
make
sure
we
we
got
to
these
two.
So
we
have
the
discourse
groups
set
up
and
there's
a
few
people
who
have
been
posting
on
there
already,
which
is
great
I,
asked
a
couple
of
people
who
have
sent
us
emails
with
a
bunch
of
questions
if
they
wouldn't
mind,
reposting
them
there,
and
so
what
I?
A
Well
we
found
out
is
that
for
brand
new
users
they
can
only
post
three
different
things,
so
I've
gone
in.
They
can
only
post
three
different
things
if
they're
a
brand
new
user,
so
I
went
in
and
updated
every
one
who's
currently
in
the
group
like
a
higher
level,
so
that
people
can
just
be
adding
more
content,
especially
early
on.
We
want
people
to
be
able
to
add
stuff,
so
yeah.
So
take
a
look.
A
If
there's
things
you
want
to
jump
in
and
comment
on,
Alexander
posted
some
stuff
on
the
e.l.f
and
I
saw
that
John
John
already
jumped
in
and
then
there's
a
woman
named
Jennifer
who
has
posted
some
questions
and
I'll
go
in
and
respond
to
those
because
I
already
like
she's,
basically
just
reposting
stuff
from
an
email.
But
if
you
have
other
comments
or
like
ideas
around
what
she's
written
feel
free
to
chime
in,
we
thought
I
also
change
the
layout.
A
A
So
this
is
what
it
looks
like
now
before
we
had
like
squares
for
each
of
these
topics.
But
now
what
we
have
is
basically
the
categories,
so
you
can
still
see
each
of
the
categories
that
we
have,
but
then
on.
The
right
is
the
latest
topics,
and
so
it's
easy
to
sort
of
see
what
questions
have
just
been
asked
and
where
there's
activity
versus
not
so
see
if
this
is
works
for
everyone,
I
think
it's
it's
the
same
layout
that
the
Jupiter
team
uses
so
take
a
look
see
if
it
makes
sense.
A
Yeah
and
then
the
other
thing
that
I
wanted
to
bring
up
is
this
issue
that
I
posted
for
tracking
tracking
issues
as
they
come
up
with
the
groundwater
notebooks,
and
so
here
what
we've
done
is
I've
got
started
elements
just
sort
of
like
a
bit
of
a
laundry
list
of
things.
I
went
in
this
morning
and
actually
fixed
the
the
question
that
we
had
with
the
data,
so
Doug
I,
don't
know
if
you
remember
and
and
Devin
this
will
be
relevant
to
you
earlier.
A
When
we
created
a
data
class,
you
could
not
set
the
uncertainty
in
the
instantiation.
Now
you
can,
and
so
that
should
make
that
a
little
easier
and
hopefully
a
bit
simpler
to
set
up
here.
If
you
want
to
create
the
uncertainties
and
set
that
by
a
keyword,
it
should
be
a
little
more
transparent,
now
and
I.
Think
Joe
is
going
through
and
making
sure
that
pull
requests,
sort
of
gets
up
straightened
to
the
rest
of
the
the
simulation
stuff.
A
Thank
you,
yeah
no
problem
and
then
there's
a
couple.
Other
comments.
Sagi
there
was
just
one
in
there
with
respect
to
the
one
DDC
I'm
changing
t2
thicknesses,
if
you'd
be
game
to
do
that.
That
would
be
that'd
be
great,
and
if
you
don't
have
the
time,
if
you
could
just
let
us
know-
and
somebody
else
can
yeah.
E
When
we
do
these
1d
simulations,
we
we
define
these
layers
as
being
from
plopped
down,
and
so
we
create
this
mesh,
which
technically
is
ordering
the
cells
and
the
opposite
order
that
a
standard,
sim,
peg
definition
and
I
just
I
wonder
what
I
don't
know
if
that's
sort
of
been
thought
of
or
addressed
or
if
anybody's
concerned
it
make.
If
you're
doing
a
1d
problem,
it
makes
complete
sense
to
define
thicknesses
from
from
top
down.
But
then
we
define
this
this
mesh
and
we
it's
not
being
defined
in
the
same
same
convention.
E
G
Nothing
particular
the
way
that
the
reasoning
was
how,
if
you
actually
write
down
the
equation,
that's
actually
much
more
natural
to
do
that
in
such
a
way,
so
yeah
yeah.
If
you
want
to
like
make
it
really
consistent,
then
we
can.
We
can
do
that,
but
I
think
that's
sort
of
typical
convention.
If
you
think
about
one
decode
and
set
down
and
positive,
you.
A
G
A
C
A
A
We
can
see
what
our
timelines
are
like,
because
we
need
the
simulation.
We
need
all
the
simulation
work
in
before
everyone
goes
to
Myanmar,
and
so
it's
a
bit
of
a
bit
of
a
timing
game,
but
at
the
same
time
I
just
wanted
to
bring
this
up,
and
we
can.
We
can
take
a
look
at
it.
So,
with
the
just
point
up
the
notes
again,
one
of
the
things
that
I
put
in
the
notes-
and
it's
also
in
that
github
issue-
is
serializing
and
deserializing.
A
So
basically,
we
create
a
synthetic
data,
object
in
some
forward
simulation,
and
then
we
should
be
able
to
load
that
up
into
a
new
notebook
and
run
an
inversion
with
that
with
the
D
serialization.
That
just
means
we
need
to
have
be
a
little
careful
with
how
classes
are
instantiated,
and
so
it
just
requires
a
little
bit
of
thought
and
a
little
bit
of
testing
so
I
put
in
an
example
in
the
notes
just
of
like
how
what
that
workflow
should
look
like.
A
So
as
folks
are
refactoring,
the
simulation
are
like
injecting
the
simulation
into
whatever
sort
of
area
or
whatever
physics
you're
looking
at.
If
you
wouldn't
mind
adding
a
test
for
this,
that
would
be
much
appreciated
if
you
run
into
any
questions
or
drilling
up
ten,
what's
going
on,
feel
free
to
post
in
slack
or
on
discourse,
and
we
can
jump
in
and
try
and
help
out
there
so
at
least
for
the
DC.
Let's
try
and
get
this
up
and
running.
If
we
don't
get
it
across
everything,
that's
that's!
A
Ok,
we
can
inject
it
later
on,
but
at
least
for
the
DC
I
think
it's
important
that
this
works
so
that
you
can
just
generate
some
synthetic
data
in
one
notebook
and
load
it
up
in
another
that'll,
greatly
simplify
I'm
sort
of
like
the
idea
of
doing
like
teaching
inversions
so,
and
there's
a
note
about
that.
Also
in
that
in
that
issue,
but
there's
an
example
of
how
to
do
that
with
the
simple
class
or
at
least
how
to
do
the
testing
with
discretizing.
A
I
C
I
One
which
I'll
put
those
up
there
I'll
put
the
UBC
dump
file
and
a
data
file
in
the
top
will
file.
Those
directories
that
I
had
there
and
the
second
thing
and
I
know
we
talked
about
this
before
and
I,
not
100%
clear
on
what
is
the
front.
End
processing
that's
available
with
the
wholesome
bank
package
and
is
I'm,
assuming
that
it
may
be
that
using
process
is
the
easier
way
to
go
right
for
the
data
set
to
make
sure
you
have
a
good
data
set,
that's
going
into
initially
into
the
modeling
programs.
Yeah.
K
A
A
A
I
A
A
F
At
this
point
were
kind
of
okay,
getting
together
with
Joe
tomorrow,
when
maybe
when
the
snow
clears
up,
and
we
all
get
in
into
into
UBC
and
yeah
kind
of
step
through
the
books.
That
Joe
is
as
cleaned
up
and
then
the
2d
I'm
not
sure
whether
you've
had
a
chance
to
test
drive
the
2d
codes.
Joe.
Have
you
done
that
I'm.
C
C
I
E
E
Well
and
everything
just
because
the
the
datasets
that
we
were
provided
were
slumbrous
jetty
and
everything
was.
You
gave
us
the
a
B
over
2
and
M
n
over
2
values
and
then
the
beta
values,
and
so
I
just
did
everything
kind
of
based
on
that.
Assuming
that
was
the
the
survey
configuration
they
wanted
to
use.
F
E
You
wanted
to
do
yeah,
it
would
be
I
mean
I!
Guess
if
you
technically,
if
you
gave
it
a
data
file
such
that
a
B
over
T
and
then
over
to
dimensions
or
in
a
winner
array,
go
in
and
work
that,
and
it
would
just
be
a
winner
right.
So
you
could
put
in
a
B
over
2
and
M
n
over
2
values
that
follow
the
winner.
E
A
Just
for
reference,
especially
for
kolten
in
entre
I,
put
the
link
for
the
notebooks
that
were
talking
about
in
both
the
chat
and
then
next
to
Joe's
name
in
the
notes.
So
if
you
want
to
take
a
look
and
see
the
groundwater
notebooks
that
we're
playing
around
with
they're
just
in
this,
in
that
research,
we
go
there.
D
This
may
be
something
quick
sure
I
mentioned
that
doing
the
release
of
the
0.13
on
the
slack,
but
there
was
like
many
ways:
I'd
show
it
went
unnoticed,
but
basically
so
I
released
like
so
like
the
tag.
What
was
tagged
as
the
open
13,
but
we
had
at
least
two
or
three
PR
in
on
the
master
branch
since
so
I
was
wondering
like
how
close
is
a
simulation
class
to
be
murder
on
master?