gusaSlhFN-4

so lover and welcome to another tutorial
my name is Henry Bakshi and this time
we're going to be going over how you can
create a Watson conversation application
which includes contextual features which
essentially piece well let me actually
simplify this so to begin we're going to
be using IBM Watson conversation
more specifically the context feature of
the service and let's just dive right
into what this means today so just
arriving in though I like to say a big
thank you to one of my subscribers who
Daiya for sending in this question and
so of course thank you much for
reminding me to create this video so
this is actually kind of a part two I
guess you could call it to my earlier
drone videos where you could ask loss in
conversation for example what's the
battery percentage in the drone and it
would be able to answer you that was
using IBM Watson conversations context
feature so let's take a look at what
that is and how it works so to begin
let's say we've got our Watson
conversation Watson conversation little
service over here okay so we've got our
service and let's also say that we've
got our node.js application and of
course we've got our drill now I'm not
going to get into how you know the
Raspberry Pi is able to send messages
from the drone into the a ask them back
instead I'm just going to simplify it
all by saying that the drone just sends
nodejs it's valeria for it's but a drone
keep sending no js' a lot of information
actually about itself but one of those
keys would be battery percentage okay
and so what's happening is that lets say
that Watson conversation let's say
actually that you say that let's say the
users input to Watson conversation would
be
um what's the battery level okay and
they were to send this string - no - as
input what would happen is before
sending the request to IBM Watson now
it's not just going to do that what's
going to happen it is going to form a
request and so it'll create this little
request and that request will be sent to
lots of conversation the request
contains quite a few things namely the
input
it contains the input which would be of
course this string and we've also got of
course something else like for example a
context which is a JSON and also this is
a string and there are a few more things
but I will get into that right now so
these are two piece that are very
important to us so no js' will form a
request which contains the input in the
context and these two are sent to watson
conversation so how does this work
though well essentially the context is a
JSON value and this JSON value contains
a key called battery P okay and this so
if I just draw this contains something
called a key called battery P and the
value is really however much battery is
left in the drone for example 65 for
first 65% we're just going to say 65
sorry and then we're going to close our
curly brace and this is essentially our
JSON value we've got a battery P 65 and
so this is the context that's being sent
to IBM Watson conversation and Watson
conversation like the dialogue that I've
created here what it does is let's say
it needs to send let's say it needs to
find the battery percentage the it'll
actually go through its contacts and if
it contains
the battery p-value then it'll give
something else that contains that
battery P tell you but if it doesn't
have this in its context where this is
equal to a null value what's going to
happen is it's going to go through
another dialogue chain but it it can it
can return back to no just no J sorry
essentially that sorry I don't have the
battery information right now
please try asking again later or make
sure that your turns and drum turned on
that's a good thing
same thing for altitude information same
thing for all this type of information
that lots and conversation can give you
if you ask for it but then again without
any further ado let's get straight to
the Mac part now we're I'm going to be
showing you how you can code in a simple
nodejs application using the same pizza
example that I gave you a little while
ago in my last Watson conversation video
and now we're going to be incorporating
context so it Watson actually knows
about the pizza that you're building and
is then able to send that pizza over to
really any sort of a nodejs API so that
they can actually request the pizza
company to create the pizza for you all
right
let's actually do that now let's get
straight to the knock part where I'm
going to be showing you how to build
that nodejs application let's go now
shall we alright so welcome back to the
Mac part and now we're going to be
covering how exactly you can create this
Watson conversation but so now remember
the entire point of this Watson
conversation boss is to essentially tell
you the current hour and minute and of
course say hello and goodbye to you and
so let's begin shall we
so you can see over here I've got a
terminal window open right here at
figure something agency oh and then of
course over here on this side I've got
the lots and conversation workspace open
here so now let's actually begin by
going over the Swanson conversation
workspace it's actually quite simple I
just have a by intent which has by
goodbye I must go I will go now see you
soon time has current time I eaten all
the time please tell me the time time
what's the time when is it and then of
course greeting contains hello hey hi
how are you how do you do nice to meet
you now these are all the intents that
we have in this application quite simple
of course we have no entities because
there's nothing complex that we're doing
here but the dialogue is extremely
simple as well we basically just have a
greeting where Watson can respond with
hello a by intent where wasps can
respond with by I hope to see you soon
and then of course we've got another
intent which is okay if you want to know
the current time and I currently have
time meaning that the our context
variable is greater than negative 1 and
the minute context variable is greater
than negative 1 now before I explain
what that means that I know this is
essentially a tutorial that will teach
you how to actually use context with IBM
Watson before I tell you how this know
it works here let me show you a demo of
this application in action so again this
is coded in nodejs here so if I run this
application as you can see I've actually
salvaged some of this code from my
Facebook live event with IBM so you can
actually go check that up and check that
out if you like there will be a link in
the description now once more so it's
essentially Watson says hello to us as
you saw in this greeting node and then
we can say to Watson hey what's the time
Watson responds with the time is
currently 2/3 one amp all right so now
we can also ask you to tell us by and
Watson says bye I hope to see you soon
then you can exit the application
and so that is essentially how this
application works now as you can see
that was actually the correct time over
here if you look at the watch now one
more thing to note here is that what's
happening is every time you send a
request to Watson whether it's saying
bye or what's the time or anything you
are sending the current time to IBM
Watson ok in this current time is that
used by IBM Watson to either give a
response or it's just story ok and so
essentially the point is here that let's
just say I wanted to use this
application do ring let's say I mean
inside of this workspace ok so it says
let's just clear this and so let's just
say I say
or hi in in this workspace it knows that
this is a greeting intent and it says
hello but let's just say I say what's
the time it's going to give me an error
okay or or let's just say sorry your
application is not provided me with a
current time the reason is because this
workspace over here doesn't have the
code that I've implemented in order to
give Watson conversation the context
which is the current time however over
here are no js' application it is
sending Watson the current time in the
context and so Watson is able to parse
that out give it to you as a response if
you've asked for it and just because you
don't work here let's take a look at the
source code itself so once again as you
can see quite simply just with any other
you know just with any type of node.js
application we are declaring our imports
or requires and in this case we're just
importing Watson developer cloud library
we are also importing or we're actually
defining the variable conversation which
is equal to a new Watson conversation
we're giving it the correct username and
password we're giving a good version and
a version date and this is where the fun
part comes in so as you can see I'm
actually declaring an object here
okay this object is called context and
inside of context of setting the hour to
negative one and the minute to negative
one this means that if Watson has have
not been provided with the current hour
a minute for some reason even though it
should be then they will be negative one
and as you can see from this node over
here we are checking if these values are
91 if they're negative one then we're
going to say that sorry your application
is not provided me with this information
but if they're greater than negative one
then we know that we can actually idiot
we have this current latest information
and we can give it to the user next in
our nodejs application we've got a
prompt function man this prep function
again this is actually been taken from a
stack overflow post and so there will be
a link to this prompt function down in
the description essentially what this
allows me to do is just essentially ask
the user for input and so essentially
this will allow me to say okay let's say
I want input from users
you want some conversation as you can
see I just say you as what I'm asking
the user and then the user types in what
they want to send to Watson that's
simple but of course we've also got a
time convert function again very very
simple this is from a stacker or so I'll
post the link to the stack overflow post
will be in the description and this will
essentially convert 24-hour time to a.m.
p.m. time okay
and this a.m. p.m. time is actually
pretty important because of course I
mean for this example I wanted to give
you 24 I mean sorry 12-hour time not
24-hour time and since no js' by default
gives us 20 that 24-hour time we have to
convert that to 12-hour so this is again
just a very simple function
not too much related to the core
workings of this application or Watson
context then we have one more function
called conversation message which takes
a message and so this is actually the
core function that allows us to well try
to use this application and so here we
did good as you can see this logic over
here basically gets the current date and
it gets a 12 hour time format using the
time convert function okay then in the
context object that as you know we have
set right here we are telling the hour
to be we're just you know parsing that
out of the T convert result and then of
course we're setting the minute to
another you know we're parsing out the
minute from the Tekin result once more
once that's done we're calling
conversation dot message on this
conversation variable that we had
defined in the beginning of this
application essentially this will be
given a workspace ID some input which is
your current message which was passed to
conversation message and then of course
the context and this context contains
the I guess you could say well it's
quite literally context that I need
Watson should know to continue from if
they're asked something related to that
context in one of the dialog notes or as
the intent and so once we've sent this
request to Watson and we've given it the
new context and the workspace ID and the
input we wait for its response and if we
got an error we just print out the error
and continue but if we got our response
and then we're going to print out this
is what Watson said that we're going to
parse out Watson's response and then
we're going to ask for another prompt
again our input is you we want the
question to be you and so basically the
callback is once the user has provided
some input which is in this input
variable we will call conversation
message with that input which becomes
undefined I type it will keep asking me
questions
you give Watson some input it'll give
you output and it continues after that
we set the context to the responses
context but why we do this well
essentially let's just say that in the
dialogue note Watson had actually
modified the context or there's some
update in the context well in that case
we want to set our current context to
this new context in order to keep both
applications in sync both Watson and
ourselves and that is basically how it
works and just to you know kick start
the conversation we call conversation
message with hi and that is basically
how this entire application works in
fact let's actually take a look at one
more demo here let's just ask lots of
what's the time bye or and as you can
see it says the time is currently 247
a.m. all right as you can see then of
course it continues in that infinite
loop it asks me for more input but what
I can do is just ctrl C and exit that in
finite loop now one more thing that's
absolutely critical to understanding
context with Watson conversation is how
it's actually used in the dialogue so as
you can see I'm in the dialogue pane of
the Watson workspace here if I and if I
click on this node which handles giving
time to the user as you can see this
note is triggered by if their intent is
to find the time and if the our context
variable is greater than night negative
1 now you already know what the greater
than negative 1 means great and negative
1 is the default which means that at
this point we do not have time if it's
negative 1 ok
but it's mo it is if you're writing from
the node.js application with
code then it's for sure that will always
be greater than negative one
we're spreading listen as a protective
measure and so essentially what this
means is as you can see over here if I
just zoom in maybe a little bit yeah so
as you can see I'm actually calling I'm
actually getting this context variable
with a dollar sign here now for an
intent you put a hashtag symbol for an
entity would put an at the rate symbol
but for this context variable we're
putting a dollar sign as you can see
right at the source code here you can
see in the context we're setting the
variables hour and minute and that's
what we're doing dollar sign our dollar
sign minute those are our context
variables and if you can say if you have
these variables that means that Watson
is getting this context correctly and of
course we are checking if it's greater
than negative one then we know in that
case that we currently do have a time
and in the response with the actual
response here we can see that the time
is currently and then give it the
context variable per hour and the
context variable per minute in order to
give the user the current correct time
and that is essentially how this entire
application works and how Watson
conversation deals with context of
course thanks to Diane for sending in
this question originally I hope this was
able to help you out but of course
that's going to be it for this tutorial
all right so if you have any more
questions suggestions or feedback please
do feel free to leave that in the
comments down below she emailed them to
me at Teddy mini gmail.com or tweet them
to me at Taji betty if you like you can
also subscribe to the channel as it
really does help out a lot and also like
the video and share if you think it
could help someone else you know
alright so that's gonna get from this
video today thank you very much goodbye