coding Session Log 10-22-19

We can talk here if needed.

Sure

Okay So far I have implemented all
the functionality. We now need to
abstract it for midca.

How do we go about doing that?

The abstractions should be done in 
the perceive phase. 

Are we using ROS right now?
No. No need for it if its only job 
is for communication..

ok so lets write the communication
in the percieve phase. I will 
copy a function for reference 
right now. just a minute

Just a minute. I am looking at the
states in grace midca. 

ok

Here are the states that we need
knows(fumin, pooldepth)
at_bottom(grace)
at_surface(grace)
here's the question
how do we know when grace is surfaced?

It is based on the depth and we 
can also use an acknowledge message
from the person who recieved 
communication, then we know we are
at surface because radio works 
and depth is almost 0

ok ! so the code in percieve should be
if (depth ==  0 or message):
	at_surface(grace)
so can you write this code
"if (depth ==  0 or message)" ?
I will point you where it should be 
written. is it ok ?

yes. we can do that? the message
flag will most likely require that
we write to a file when the message
is recieved or wait a set period
to see if a message was recieved

I have a clarification question. 
midca controls grace right? then
from your previous message, did you
say that if grace can succesfully
communicate with a person then it
is surfaced right? 

Yes

So, we now have two options ! we
can either check for the flag 
or wait certain amount of time 
for the flag to be set, is that correct?

Exactly.

midca is a continous loop, so as
of now I dont think we should wait
for the message. If it misses in
one cycle. It sure can get the message
in the next cycle.

Great. So I will set it up where
we check a file that tells us if
we recieved an acknowledge message

Sure ! Before you do that. I will just
give you the template in the 
percieve.py. So you know where you
can write. Does that sound good?


Perfect!

give me 5 minutes.

I have created Graceobserver class
and had some comments where the
code should be written. Let me
know if you have any questions?
one more thing this class contains
init and ru functions 
init is only run when midca is
initialized and run functions is
executed iteratively 

Ok. I think I can work with this.
It may take 30 minutes or so. So
you can leave and get other work done
I will text when I am finished
or have questions.

Sure ! sounds good ! 

Nice ! How about pool depth? MY
idea is to have certain depths
mapped to the words
for example if it is at 40 then
pooldepth1
60 then pooldepth2 and so on..

I don't particularly like that idea
because depth is a continous 
quantity that we are trying to 
determine. So we won't know it
ahead of time

How about mapping a range of depths
to a certain area?

We can do that for now, at least to 
get a working example. We can 
work out the fine details after the
code is able to run.

Sure ! let us have one pooldepth1 
you choose the range? does it sound ok

Can pooldepth just be a variable
without a numeric value? Just at
bottom let's us know that we know
it?

I didnt quite get that ! do you mean 
if the grace is not at the surface or
at bottom then it is at pooldepth.
do you want to assume something like that?

That would work. Then we have
three states that cover the continium

Sure. this would get us started.
One more thing, after that is done
I will have to modify some part of the
code in midca, which I will do it 
today. But inorder to execute, I
think we need to meet again tomorrow
is it ok ?

Yes. Same time works for me.

How about 11 ?

yes.

Sure that works. Let me know when you
are done. I will push the code to
github and work on my computer and pull 
it back. Through vpn, it is a bit slow
working on the code. is it ok?

Yep, that works. This is mostly done.
Are there any other changes you would
like to add before pushing it to github?

Not I can think of. I feel everything
is in place right now. I will let
you know If I missed anything.
I am pretty sure I got
what i needed for the initial start.

What about the async_3d.py? Should 
something similar to this be done 
there?


I have thought about that, before
I want you to write the feedback
function for every action. But now
since we can percieve each and
every action in the percieve phase
we are already getting the feedback
this can be used in the asynch_3d.py
You have completed functions for
all the actions right?

Yes. Let's look to make sure though
These functions interface with the
low level actions of the robot,
but I don't have them as the 
AsyncAction type classes like 
GraceSense(...)


Ok ! I will work on it. Its not a 
problem. I can use those functions
to execute the actions. I think
there is an action to sensedepth
as well. which should read the 
value from the depth sensor or
some sensor

Yes. I moved it to the percieve modulr
though. I can put it back here as well.

Its ok lets leave it like that. 
sensing can be considered as
both action and perception, but 
lets leave it like that. Can
you please comment that part of
the code in percieve. So I can
use it if required 

Yes. Give me a minute.
Good?

Perfect ! I will now push the code to 
github. Is it ok?

Yes.

If you want to push the code through
your account so that you will
get a commit, please do so.
If you dont care about that I c
an use my credentials. Let me know

You can use yours.

Sure !

Done. Thank you 

No problem!

See you tomorrow.