Chaining di funzioni - SK

SK è stato sviluppato con lo spirito delle pipeline di UNIX per prendere un comando e effettuare lo streaming del suo output al comando successivo. Questo concetto si vede dall’uso del parametro $INPUT.

Ad esempio, possiamo creare tre semantic functions concatenate e mettere in stringa i loro output nell’input successivo.

string myJokePrompt = """
Tell a short joke about {{$INPUT}}.
""";
 
string myPoemPrompt = """
Take this "{{$INPUT}}" and convert it to a nursery rhyme.
""";
 
string myMenuPrompt = """
Make this poem "{{$INPUT}}" influence the three items in a coffee shop menu. 
The menu reads in enumerated form:
1.
""";
 
var myJokeFunction = myKernel.CreateSemanticFunction(myJokePrompt, maxTokens: 500);
var myPoemFunction = myKernel.CreateSemanticFunction(myPoemPrompt, maxTokens: 500);
var myMenuFunction = myKernel.CreateSemanticFunction(myMenuPrompt, maxTokens: 500);
 
var myOutput = await myKernel.RunAsync(
    new ContextVariables("Charlie Brown"),
    myJokeFunction,
    myPoemFunction,
    myMenuFunction);
 
Console.WriteLine(myOutput);

E il risultato sarà

1. Charlie Brown's Surprise - A sweet and creamy latte with a hint of caramel 
2. Good Grief! - A bold espresso with a dash of cinnamon 
3. Wide Smile - A smooth cappuccino with a sprinkle of nutmeg

Se invece togliamo l’utlima funzione:

var myOutput = await myKernel.RunAsync(
    new ContextVariables("Charlie Brown"),
    myJokeFunction,
    myPoemFunction);

Otterremo

Charlie Brown got a present one day
He said "Oh good grief!" in dismay
He opened it up with a smile so wide
But it wasn't what he had in mind