►
From YouTube: CHAOSS Augur Meeting November 21, 2022
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
C
B
It
over
to
here
now,
under
the
root
of
your
auger
directory
in
the
agar
New
Branch,
there
is
a
directory
called
auger
and
underneath
that
directory
is
a
directory
called
API.
B
A
B
Each
of
the
methods
signatures,
so
this
is
the
name
of
the
metric
and
each
of
the
method.
Signatures
takes
a
repo
ID,
a
repo,
a
repo
group
ID
or
a
repo
ID,
a
begin
date,
an
end
date
and
a
periodic
time
which
can
be
as
it
says,
down
here
in
the
documentation.
It
can
be
day
week,
month
or
year.
C
B
Right
and
then
the
the
final
thing
that's
standard
is:
if
somebody
does
not
set
the
date,
then
it
defaults
to
the
beginning
of
time
as
far
as
computers
are
concerned,
which
is
January
1st
1970
and
it
ends
at
the
moment
that
you're
running
the
query.
So,
basically
all
the
time,
usually,
we
set
the
Val
variable
that
we're
going
to
assign
to
the
SQL
text
equal
to
nothing
and
then
there's
a
if
block.
So
if
a
repo
ID
is
passed,
we
just
execute
that
block.
B
So
if
you,
if
somebody
were
to
paths,
pass
both
a
repo
ID
or
a
revote
which
is
impossible,
but
actually
it
would
only
execute
the
repo
ID
and
then
so
for
the
repo
ID
SQL.
We
have
this
and
you
can
see
that
typically
the
period,
so
we
use
this
date
trunk
function.
So
whatever
period
is
passed,
if
it's
the
default
of
month,
then
that's
what's
passed.
B
It
groups
everything
by
month
and
then
the
begin
date
and
the
end
date.
Parameters
are
assigned
down
in
here
and
the
repo
ID
parameter
is
assigned
here
and
then
we
do.
The
appropriate
group
buys
Etc
and
then
the
else
is
simply
in
the
case
where
a
repo
group
ID
is
assigned
and
it's
very
similar,
SQL
and
then.
Finally,
the
results
are
constructed.
B
Using
pandas
read:
SQL
passing
the
SQL
text,
the
database
engine,
any
parameters
that
were
assigned,
which
you
know
these
should
look
familiar
from
the
method,
signature
and
then
it
Returns
the
results
so
in
practice.
The
way
that
the
way
that
that
works
is.
It
also
creates
a
very
standard,
easy
to
remember
way
for
Developers
to
access
the
API.
B
You
put
in
a
repo
ID,
so
from
this
list
of
report
IDs,
you
know
the
sort
of
normal
application
flow
would
be
get
a
list
of
repo
IDs
and
then
those
repo,
IDs
or
repo
group
IDs
would
be
passed
to
subsequent
endpoint
calls.
So
in
this
case,
I'm
just
going
to
do
25440,
which
is
auger
and
committers.
B
And
so
that
was
that's
the
metric
that
we
were
just
looking
at
and
that
exact
pattern
is
what
gets
followed
when
you're
accessing
the
standard
metric
effectively.
It's
repos
or
repo
groups,
the
repo
reprogram
group
number
and
the
name
of
the
metric,
and
we
achieved
that
consistency
by
having
this
register
metric
function
in
the
standard
metric
standard
metrics
also
share
the
property
of
always
producing
time
series
data.
B
So
whatever
your
periodic
periodicity
choice
is
whether
it's
day
week
month
or
year,
it
will
give
you
a
Time
series
data
set
for
all
of
that,
which
of
course,
lends
itself
well
to
graphing
and
and
Trend
analysis.
C
Right
we
heard
like
I
wanted
to
ask
like
we
have
a
date
parameter
to
be
seen
over
here,
so
it's
like
specifically
on
that
date.
We
have
these
many
committers.
Is
that
what
the
metric
Google.
B
June
30th
2020
to
six
different
people
made
commits
for
the
week
of
or
I'm
sorry.
The
month
of
March
2022
26
people
made
commits.
C
C
Through
it,
so
these
are
the
unique
committers.
Yes,.
A
C
B
B
So
routes
routes
do
not
require
input
parameters,
although
they
can
take
them.
Routes
do
not
necessarily
and
I
think
most
in
most
cases
do
not
take
reap
date.
Time
parameters,
oftentimes
they're,
used
to
summarize
things,
so
you
can
create
like
one
that
took
a
date.
Time
parameter
like
in
the
case
of
this
query.
C
B
Yeah,
so
we
would
develop
that
all
out
put
those
parameters
in,
and
in
this
case
I
would
add
it
I
would
like
take
the
config
and
just
take
any.
Let's
see
metadata
is
probably
a
good
one.
B
And
basically
so
in
a
non-standard
metric
in
the
routes
folder,
it
also
has
a
different
set
of
imports
and
an
auger
API
version
that
gets
declared
this.
This
gets
declared
here
because
it's
handled
by
The,
Decorator
and
standard
metrics
and
the
other
thing
that's
handled
by
The,
Decorator
and
standard
metrics
is
this
definition
crate
route
server.
So
essentially,
this
top
piece
is
all
unindented
and
then
every
metric
that
you
add,
every
route
that
you
add
gets
an
indentation
in
in
a
non-standard
metric
and
it
begins
with
declaring
at
server.app.route.
B
You
have
to
do
word,
wrap
there
as
well,
and
then
it
does
methods
get
which,
when
you're
retrieving
data
is
is
all
you
all
you
need
to
do
is
a
get
method,
and
then
you
define
in
the
next
line
the
name
of
of
the
metric
and
the
so
the
the
text,
and
so
here
we're
not
even
doing
a
data
collection
date,
I,
don't
think,
there's
any
parameters,
it's
just
giving
you
a
repo
info
for
all
of
the
repos
inside
of
an
instance.
B
So
there
it's
it's
actually
pretty
simple,
creating
our
standard
metric
or
a
route.
When
you
have
that
query
once
you
have
a
query
of
something
that
you
want
to
parameterize
and
certainly
like,
if
you
wanted
to
do
date,
if
you
wanted
to
add
date,
functions
to
this
query,
which
I'll
put
in
the
zoom
chat
here,.
B
If
you
wanted
to
create
a
endpoint
for
that
query,
you
could
parameterize
the
name
of
the
name
of
the
tag,
and
you
could
also
add
in
some
of
the
date
logic
that
exists
in
the
standard
metrics
and
you
can.
That
would
probably
be
as
far
as
you
would
want
to
go.
C
B
If,
if
the
date
windows,
don't
matter
and
I,
think
in
this
case
they
probably
do
because
I'd
want
to
see
the
the
growth
and
decline
in
the
use
of
different
yeah.
B
Yeah
that'd
be
a
great
contribution.
I
think
people
would
be
real
interested
in
seeing
a
bot
label
endpoint.
A
C
Or
any
any
labeling
point:
if
we
want
to
exactly
see
that
also
probably
we
can
have
some,
you
know
it
repeats
for
all
the
labels,
like
the
accounts
for
all
the
labels
for
a
repo
group
and
a
repo
ID
like
something
plotted
up
to
exactly
see
which
label
is
being
used.
The
most
yeah.
B
Yeah
there's
there
is
one
other
step
which
is
a
little
annoying
because
it's
another
step,
but
under
the
rest,
if
you
create
an
end
point,
there's
a
very
large
file,
but
it
does
follow
a
pattern
under
docs
Source,
rest
API.
B
You
basically
provide
a
description,
an
operation
ID
and
the
parameters
and
then
the
responses.
But
this
follows
a
very,
very
clear
structure,
so
you
could
just
copy
and
paste
one
of
these
and
then
fill
it
in
with
your
values.
B
Obviously,
the
the
API
doesn't
exist
for
most
people.
If
it's
not
documented.
B
B
We
could
write
the
test
for
you,
it's
preferable.
If
you
do
write
the
test,
there
is
a
pattern
there
and
they
do
run
so,
but
there's
because
this
is
a
data-driven
system.
B
B
C
A
C
C
And
there's
one
thing
more
on
my
mind:
if,
on
the
basis
of
like
when
we
are
deciding
on
newcomers
like,
we
do
have
something
which
I
went
through.
Like
probably,
you
know,
if
there's
a
contributor
who
did
more
than
two
commits
or
something
like
that,.
C
B
So
we
have
end
points
for
that,
but
those
endpoints
are
non-standard
metrics
their
routes
and
they're
characterized
by
the
word
reports
and
the
reports.
Endpoints
are
unique
or
different
than
the
others,
because
they
return
images
like
they
return.
Actually,
a
visualized
graph
instead
of
raw
Json.
A
B
It's
you
can
get
the
data,
but
it's
already
visualized
I
I
believe
one
of
there
may
be
I
know
Andrew
mentioned
he
was
gonna
make
it
so
that
you
could
pull
the
Jason
from
those
I.
Can
ask
him
really
quick
and
Discord.
A
B
B
I
just
asked
him:
if
we
ever
did
that
I,
but
something
he
mentioned
a
few
months
ago,
knowing
Andrew
he
it
depends
what
else
I
asked
him
to
do,
but
he
may
or
may
not
if,
if
he
he
may
have
done
it,
but
I
just
don't
know
the
ins
and
outs
of
of
calling
it
yet.
Okay.
A
C
C
Probably
we
need
to
actually
make
a
lot
of
updates
to
the
documentation
because,
like
everything
has
changed,
we
don't
have
workers
in
the
new
release.
Yeah.
B
Yeah
I
know
there
needs
to
be
some
updates
to
the
docs.
That's
on
my
list
of
things
to
do
today.
B
Yeah
yeah,
the
the
you
know,
the
developers
that
I
pay
will
be
less
active
starting
next
week,
because
they'll
have
the
last
two
weeks
of
courses
and
then
exams,
and
then
most
of
them
will
take
the
holiday
we'll
take
some
time
off
before
the
holidays,
but
they'll
work.
B
You
know
either
the
let's
start
the
work
starting
the
last
week
of
the
year.
You
know
so
they'll
work
three
or
four
weeks
pretty.
But
it's
full
time
then,
after
that,
but
yeah
there'll
be
a
three
or
four
weeks
of
much
less
activity.
A
C
C
For
my
case,
it's
like
I'm
already
going
through
my
exams
presently
and
December
is
something
like
when
we
are
really
free.
So
you
know
for
having
open
source
contributions.
I
guess
December
is
something
yeah
yeah.
A
And
let's
see.
A
C
There
I
am
really
commented
to
working
on
the
street
I.
Guess,
let's
see,
how
can
I
improve
it
plus
I
wanted
to
ask
me
what,
in
the
tool,
is
exactly
taking
the
time
when.
A
B
So
the
the
data
collection
takes
the
most
time
it
takes
about
three
percent
or
less
of
the
time
that
it
used
to
take
because
we're
using
multi-threaded
queuing
system,
so
we
can
send
out
and
gigantic
number
of
jobs
all
at
once.
The
timeouts
that
we're
having
to
put
in
are
actually
because
we're
hitting
the
database
too
fast
too
often.
C
B
Know
very
rarely
we'll
get
a
GitHub
timeout,
but
no
more
or
less
than
we
were
getting
in
the
old
version,
principally
we're
just
pounding
the
database
so
hard
because
we're
inserting
data
so
fast
and
so
yeah
it
just
it
collects
data
like
way
faster
than
it
used
to.
B
So
that's
that's
the
biggest
advantage
and
that
you
really
would
only
notice
it
if
you
were
collecting
more
than
like
3,
000
or
4
000
repositories,
because
that's
where
you
go
from
days
to
weeks
in
the
old
version
of
auger
and
now,
we've
gotten
all
the
data
for
12
000
repositories
in
a
week,
which
is
pretty
phenomenal
honestly
because
it's
not
just
collected,
but
it's
also
validated
and
structured
in
a
relational
style
that
people
can
use
so
right.
You
know,
there's
a
lot
going
on
behind.
A
B
Yeah
we
did,
we
did
I
need
to
make
a
video
for
installation
so
that
we
don't
have
to,
because
I've
got
to
set
up
a
machine.
I
decided
I'm
gonna.
Do
a
video
I
meant
to
do
it
last
week,
but
I'll
do
it
this
week,
so
that
and
I'll
promote
it
I'm,
trying
to
to
time
it
with
the
release
and
then
use
a
a
bear.
B
So
and
else
yeah
so
right
right
now,
I
would
say:
I
want
I
want
to
make
sure
that
the
installation
is
perfectly
stable
before
I
do
the
demo.
But
in
you
know
the
installation,
instructions
that
you
would
follow
for
the
current
version
of
auger
are
still
valid
for
auger
new
okay,
so
you'll
you'll
get
fewer,
prompts
and
that's
about
it.
But
the
process
of
starting
auger
is
exactly
the
same.
B
So
all
of
there
I
guess,
there's
one
dependency
in
that.
You
need
to
have
redis
installed
on
your
computer
on
the
computer
you're
installing
it
to
because
we
use
redis
for
persisting
a
queue.
C
B
I
mean
well:
well,
that's
one
of
the
reasons
I
need
to
do
the
dry
run
on
an
ec2
instance,
but
yeah
auger
auger
is
supposed
to
install
it
now
I've
seen
it
install
it
I
think
the
challenge
is
if,
because
I
install
it
so
much
that's
why
I
want
to
do
it
on
a
bare
server,
so
I,
don't
so
that
I
catch
everything
that
any
new
person
would
catch.
C
B
Believe
it
I've
fixed
that
in
auger
knew
I
know
I
fixed
the.
There
were
some
issues
with
the
score,
the
ossf
scorecard
because
they
changed
their
code
so
often,
but
I
fixed.
That.
C
And
probably
it's
just
one
last
thing
that
I
wanted
to
ask
yeah
the
machine
learning
workers
that
we
have
already.
Are
they
using
tensorflow?
Yes,.
B
Yeah
in
the
message
insights
worker
uses,
tensorflow
and
yeah.
The
message
insights
worker
is
the
one
that's
using
tensorflow.
B
B
So
if,
if
there's
a
message
or
group
of
messages
that
are
kind
of
noticeably
different
than
other
messages
in
a
repo,
those
can
get
flagged
as
well,
just
to
make
sure
they're
not
spam,
mostly
or
to
determine
that
perhaps
they
are
identifying
some
completely
new
thing,
but
messages
that
fall
out
of
a
normal
band
would
would
be
ones
that
a
maintainer
would
want
to
pay
more.
You
know
give
special
attention
to
just
to
see.
C
And
this
Nick
benevolo
I,
want
to
like
add
about
to
be
specific
or
like
I,
have
an
idea
whenever,
like
in
future,
presently,
we've
discussed
one
of
them,
but
if
I
have
to
add
a
route
so
like.
What's
the
process,
I
should
open
an
issue
with
the
entire
description
and
get
it
approved
or
like
I
should
ping
you
on
Slack.
C
B
If,
if
the
route
is
is
not
already
in
existence,
and
it
either
fits
a
metric
or
is
obviously
useful,
as
in
the
case
of
the
bot
metric,
then
my
recommendation
would
be
to
go
ahead
and
build
the
metric
and
issue
a
pull
request.
A
C
B
Know
and
then,
if,
if
you
know,
if
there's
questions,
then
we
can
have
the
discussion
in
the
pull
request
review,
which
I
I
find
to
be
more,
both
more
efficient
and
more
fun
than
trying
to
negotiate
all
of
the
details
in
advance
that
you
know
what
I
mean
yeah.
B
And,
of
course,
you're
welcome
to
open
an
issue
and-
and
you
may
get
some
comments
that
are
helpful,
so
I,
wouldn't
you
know,
I,
don't
discourage
you
from
opening
an
issue,
but
I
would
also
say:
don't
like
wait
for
the
issue
to
fully
resolve
all
of
your
questions
before
just
building
the
metric.
You
know
we'll
do
the
best
we
can,
but
sometimes
talking
about
things.
Abstractly
is
more
difficult
than
just
looking
at
the
code.
C
I
was
going
through
what
exactly
the
pattern
is,
and
even
for
the
metric
there's
a
pattern
that
you
already
discussed.
That's
very
much
clear
yeah.
C
How
do
we
decide
they're
like
there
is
a
core
dot
by
I
guess,
along
with
the
task?
It's
not
wrong.
B
C
B
C
B
Isaac
is
Isaac,
was
gonna,
provide
an
explanation
of
that
and
an
example
template,
but
I
don't
have
that
from
him.
Yet
so.
B
Yeah
I
just
I,
just
pinged
him
on
that
again
he's
he's
aware
of
it.
He
was
on
the
call
last
week,
but
this
Pat
the
last
week
was
our
last
week
before
our
haul,
our
Thanksgiving
holiday
break
here
in
the
U.S,
so
I
think
a
lot
of
professors
loaded
their
students
down
with
tons
of
work.
So.
C
C
B
B
How
I
create
a
new
data
collection
task
and
what
the
because,
because
I
know
it's
as
patterned
as
API
endpoints,
and
so
there
are
just
certain
things
like
you
need
to
know,
of
course,
the
GitHub
API
that
you're
calling
and
you
need,
but
then
thing
there's
things
that
are
already
built
like
pagination
token
passing
so
a
lot
of
the
work
that
you
would
have
to
do.
If
you
were
writing
it
from
scratch,
is
already
baked
into
auger.
So
writing
a
task
to
collect
data
in
auger.
Is
you
know
five
times
as
easy?
B
As
you
know,
writing
a
bare
naked.
You
know
GitHub
API
call
it's
it's
just
I
just
need.
We
just
need
to
you
and
I
both
need
that
example,
so
that
we.
B
Be
awesome,
thank
you
very
much
meet.
Thank
you,
yeah
and
ping
me
on
slack.
If
you
have
any
questions
otherwise,
I'll
be
here
every
Monday
for
the
next
month
through
the
first
week
of
January,
and
we
can
talk
then.
C
B
I'm,
sorry,
for
the
last
minute
change
of
times,
I
just
had
my
department
chair,
decided
that
we
were
going
to
have
a
9
A.M
meeting
for
a
task
force
that
I've
been
assigned
to
so
I
had
to
I
had
to
Pivot.
C
That
kind
of
it's
too
late
at
night
for
us
eight.
B
B
Time
all
right-
well,
maybe
I'll
just
make
them
all
earlier
than.
B
I
can
I
can
do
that
I'll
change
them
all
to
eight
yeah
sounds.