Graphlogic.ai documentation

How to create a bot

Creating a Bot with Weather Service Integration

The

@Platform⁠

offers integration capabilities with various external systems, such as CRM or other services with open APIs. While the article

Creating a Simple Survey Bot⁠

taught us how to create simple push-button bots, this article delves into integrating a bot with a weather service. Similarly, the bot can be integrated with CRM systems.

⁠

Creating External Requests

The first step is to create an

@External Request⁠

— this is a functionality enabling the

@Agent⁠

to access external systems via API. Using

@External Request⁠

s, we will send data to the weather service and receive the information required for the

@Script⁠

https://home.openweathermap.org/users/sign_up⁠

. This service allows you to make up to 1000 API calls per day for free.

The requests we need can be found at

https://openweathermap.org/api/geocoding-api#direct⁠

— this request enables the retrieval of city coordinates, and

https://openweathermap.org/current⁠

— this request facilitates weather retrieval based on location coordinates.

@Platform⁠

or log into an existing account using the link

https://app.chatbots.gms-worldwide.ch/platform⁠

. You will be directed to the Company

@Project⁠

s page. More information about registration and authorization:

Registration and authorization⁠

⁠

To create your first

@External Request⁠

, go to the

@Resourses Tab⁠

by clicking Resources on the upper panel.

⁠

Go to the section with

@External Request⁠

s. More information about the tab:

External Requests tab⁠

⁠

Create a new

@External Request⁠

by clicking the Create new button.

⁠

Name the request “weather”, keep the request method as GET, and input {{ url }} in the Endpoint field. This format is used to specify

@User Context Variables⁠

in this field. To learn more about

@Context Variable⁠

s please refer to:

Context Variables⁠

. For further details on

@User Context Variables⁠

see:

User context variables⁠

. Later we will create this

@Context Variable⁠

in our

@Agent⁠

@Script⁠

⁠

In the

@Platform⁠

, there is a

@Limit on the size of the received body in response to an External Request ⁠

. If the body exceeds the limits, then the

@raw_response⁠

variable, which contains the external system’s response to the

@External Request⁠

, records not the response, but the error text. To get the response text, you need to configure its parsing by writing it to the

@User Context Variables⁠

on the Response tab. To do this, go to the Response tab, designated for parsing the response to a request, and input “result” in the Variable field and {{ body }} in the Value field. This way, the entire response body will be stored in the result variable and will be available for use in the

@Agent⁠

@Script⁠

. For additional information regarding the Response tab:

External Requests tab⁠

⁠

Save the request by clicking the CREATE button.

Create another

@External Request⁠

by clicking the Create new button. Name it “weather 2”, keep the request method as GET, and input {{ url2 }} in the Endpoint field. Later we will create this

@Context Variable⁠

in our

@Agent⁠

@Script⁠

⁠

Go to the Response tab and and input “result2” in the Variable field and {{ body }} in the Value field to save the entire response body in the result2 variable and use it in the

@Agent⁠

@Script⁠

⁠

Save the request by clicking the CREATE button.

Creating an Agent Script

The

@External Request⁠

s are ready, now you need to create the

@Agent⁠

and its

@Script⁠

Go to

@Projects Tab⁠

by clicking Dashboard button on the upper panel.

⁠

Create a new

@Agent⁠

, as described in the article

Creating a Simple Survey Bot⁠

⁠

Go to

@BotBuilder⁠

and create a

@Text⁠

@Slot⁠

with the text “Hello! I'm a meteorologist bot, I can tell you the weather in any city. Do you want to try?".

⁠

Next, add the “Yes” and “No” buttons.

⁠

After the “No” button, add a

@Text⁠

@Slot⁠

with the text “You can come back at any time.” and a

@Wait For Reaction⁠

@Slot⁠

⁠

After the

@fallback⁠

@Subslot⁠

, add a

@Text⁠

@Slot⁠

with the text “Please select one of the buttons.” and

@Jump⁠

@Slot⁠

with transition back to the button menu.

⁠

After the “Yes” button, add a

@Text⁠

@Slot⁠

with the text “Then write the name of the city.” and a

@Wait For Reaction⁠

@Slot⁠

⁠

Next, add a Memory Slot. This Slot enables you to write data to

@User Context Variables⁠

. Learn more about the

@Memory⁠

@Slot⁠

Creating and Saving Context Variables | Memory slot⁠

⁠

Find your API key by going to

https://home.openweathermap.org/api_keys⁠

and copy the value from the Key field.

⁠

In this

@Memory⁠

@Slot⁠

the Endpoint for the first of our

@External Request⁠

s will be compiled. The URL format for the first request is: http://api.openweathermap.org/geo/1.0/direct?q={city name}&appid={API key}. We need to compose a URL by substituting the name of the city entered by the

@Bot User⁠

and your personal API key.

Name the

@Slot⁠

“compose url for request”.

Enter the url value in the Key field, and the value 'http://api.openweathermap.org/geo/1.0/direct?q={{ client_message }}&appid=API key' in the Value field, where API key is the key from the

previous step⁠

, and {{ client_message }} is the

@client_message⁠

@Context Variable⁠

, which contains the

@Bot User⁠

‘s message. In the Memory Slot, The Context Variable intended for substitution must be specified in the particular format: it needs to be enclosed within double curly quotes. The entire field value must be enclosed in quotes in order for it to be written to the variable as a string. In this way, the URL for the first request will be compiled and stored in the url

@Context Variable⁠

, which we previously specified in the

@External Request⁠

as the Endpoint.

Save the

@Slot⁠

⁠

After the

@Memory⁠

@Slot⁠

, add an

@External Request⁠

@Slot⁠

— it will be used to integrate with the weather service. Enter the name “receiving coordinates” in the Name field. In the Request field, select

@External Request⁠

called “weather”.

⁠

If the

@Bot User⁠

specifies a non-existent city name, we will receive an empty response from the weather service, i.e. The

@Context Variable⁠

@raw_response⁠

will have the value []. We will use this in the

@Script⁠

to prompt the

@Bot User⁠

to enter a different name. To do this, add a

@Transition Rule⁠

@Slot⁠

after the

@External Request⁠

@Slot⁠

— it enables

@Agent⁠

@Script⁠

branching depending on the specified conditions. Read more:

Transitions to random parts of the script | Jump slot⁠

. Name the

@Slot⁠

“city check”.

⁠

When creating a

@Transition Rule⁠

, two

@condition⁠

@Subslot⁠

s are automatically generated. If a condition in a

@Subslot⁠

is true, the Agent will proceed to the Script Branch following this

@condition⁠

@Subslot⁠

⁠

We need to check whether the value of the

@Context Variable⁠

@raw_response⁠

is []. Open the "main"

@Subslot⁠

and enter {{ raw_response == '[]' }} in the Condition field. Then, clear the Name field so that the expression from the Condition field will be displayed in the

@Script Tree⁠

on the

@condition⁠

@Subslot⁠

⁠

After this

@Subslot⁠

, add a

@Text⁠

@Slot⁠

with the text “I don’t know such a city. Please enter a different name.” and the

@Jump⁠

@Slot⁠

with a transition to the

@Wait For Reaction⁠

@Slot⁠

, where the

@Agent⁠

waits for the

@Bot User⁠

to enter the name of the city.

⁠

If the city is found in the weather service, a response in the following format will be received: [

{

"name":"city name",

"local_names":{

city name in different languages

"lat": latitude,

"lon": longitude,

"country":"code of the country",

"state":"name of the country"

}

]

After the FALLBACK_CONDITION

@Subslot⁠

, add a

@Memory⁠

@Slot⁠

to extract the coordinates of the city from the response received from the service.

Name the

@Slot⁠

“saving coordinates”.

⁠

Enter the value lat in the Key field, and the value {{ result[0].lat }} in the Value field.

⁠

Add another Key - Value pair by clicking Add new field.

⁠

Enter the value lon in the Key field, and the value {{ result[0].lon }} in the Value field.

⁠

This is how the first element of the result array and the variables lat and lon contained in it are accessed — the latitude and longitude values. More information about the syntax

@Platform⁠

Syntax⁠

In the same

@Slot⁠

we can collect the Endpoint for the second

@External Request⁠

. The URL format for the second request is: https://api.openweathermap.org/data/2.5/weather?lat={lat}&lon={lon}&appid={API key}. We need to compose a URL by substituting the coordinates of the city entered by the

@Bot User⁠

and our personal API key.

Add another Key - Value pair. Enter the value url2 in the Key field, and the value 'https://api.openweathermap.org/data/2.5/weather?lat={{ lat }}&lon={{ lon }}&appid=API key' in the Value field , where API key is the key value from

step 8⁠

, and {{ lat }} and {{ lon }} are coordinate variables. This will compose the URL for the second request and store it in the url2

@Context Variable⁠

, which we previously specified in the

@External Request⁠

as the Endpoint.

⁠

Save the

@Slot⁠

Next, create an

@External Request⁠

@Slot⁠

named “receive weather” and select the “weather 2”

@External Request⁠

in it.

⁠

In response to the request, we expect to receive a response in the following format (the response has a Dictionary data format): {

"coord": {

city coordinates

"weather": [

{

"id": 501,

"main": "weather",

"description": "weather description",

"icon": "10d"

}

"base": "stations",

"main": {

"temp": temperature in Kelvin degrees,

"feels_like": what the temperature feels like in Kelvin degrees,

"temp_min": minimal temperature in Kelvin degrees,

"temp_max": maximal temperature in Kelvin degrees,

"pressure": atmospheric pressure,

"humidity": air humidity,

"sea_level": atmospheric pressure on sea level,

"grnd_level": atmospheric pressure on ground level

"visibility": visibility in meters,

"wind": {

data about wind

"rain": {

data about rain

"clouds": {

data about clouds

"dt": 1661870592,

"sys": {

system data

"timezone": timezone,

"id": 3163858,

"name": "name of the city",

"cod": system parameter

}

The

@Agent⁠

will display temperature data to the

@Bot User⁠

. To do this, add a

@Memory⁠

@Slot⁠

after the

@External Request⁠

@Slot⁠

Give the

@Slot⁠

the name “saving weather data”.

⁠

First, let's save the value of the "main" element, which contains temperature data, into a variable. The service response is contained in the result2 variable, which has the format of a

dictionary⁠

. A dictionary element is accessed through square brackets indicating the name of the element. Enter the value “var1” in the Key field, and the value {{ result2['main'] }} in the Value field.

⁠

Next, save the values of the actual and felt air temperature. To do this, add two more Key - Value pairs. In the Key fields, enter the values var2 and var3, and in the Value fields — {{ var1['temp'] }} and {{ var1['feels_like'] }}.

⁠

Since the service returns temperature data in degrees Kelvin, let’s convert them to degrees Celsius. Temperature in degrees Celsius is calculated using the formula {temperature in degrees Kelvin}-273.15. Add two more Key - Value pairs: in the Key fields, write “temperature” and “feels_like”, and in the Value fields, input {{ var2 - 273.15 }} and {{ var3 - 273.15 }}, respectively. More information about mathematical operations in the syntax:

Mathematical operations⁠

⁠

After the

@Memory⁠

@Slot⁠

, add a

@Text⁠

@Slot⁠

with the text “The temperature in {{ client_message }} is {{ temperature }} ºC at the moment. Feels like {{ feels_like }}ºC. Would you like to try with another city?” and

@Jump⁠

@Slot⁠

with а transition to the button menu.

⁠

This concludes the creation of the Agent Script. You should now have the following Script:

⁠

Agent training and testing

After creating the

@Agent⁠

@Script⁠

, you need to train the

@Agent⁠

and test its

@Script⁠

To train the

@Agent⁠

, click the TRAIN button located in the upper right corner of

@BotBuilder⁠

. It will initiate the

@Script Validation⁠

process. If the

@Script⁠

meets the

requirements⁠

, the

@Agent⁠

will be trained. Otherwise, the

@Script Validation⁠

will fail and an error text will appear in the lower right corner. Read more:

Requirements for successful Agent training⁠

⁠

If errors are detected in the

@Script⁠

, identify problem areas using the guidelines outlined in the article

Requirements for successful Agent training⁠

and fix the errors. After this, train the

@Agent⁠

again.

After successful

@Agent⁠

Training, open the

@Debug Widget⁠

by clicking the Debug button.

⁠

In the

@Debug Widget⁠

that opens, send a message to the

@Agent⁠

to start the

@Script⁠

⁠

Communicate with the

@Agent⁠

, going through all the steps of the

@Script⁠

, while checking that there are no errors in the texts and the logic of the

@Script⁠

is correct.

If you find errors, fix them, remembering to retrain the

@Agent⁠

after each change so that it takes effect. If everything is correct, congratulations, your

@Agent⁠

is ready!

⁠

Support-Weather-2024-04-15-15-33.cfg

5 kB

⁠

Creating External Requests

Creating an Agent Script

Agent training and testing

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.