Package com.jackdaw.chatwithnpc.openaiapi.function

Class CustomFunction

java.lang.Object
com.jackdaw.chatwithnpc.openaiapi.function.CustomFunction
Direct Known Subclasses:
FunctionManager.NoCallableFunction, QueryGroupFunction
public abstract class CustomFunction extends Object

Custom Function Constructor

First, you need to expend this class and implement the execute method.

You will also need to add the description and properties fields.

The description field is a string that describes what the function does.

The properties field is a map of strings that describe the parameters of the function.

The key is the name of the parameter and the value is the description of the parameter.

For example:

     public class MyFunction extends CustomFunction {
         public MyFunction() {
             description = "This function does something";
             properties = Map.of(
                 "param1", Map.of(
                      "description", "This is the first parameter",
                      "type", "string"
                 ),
                 "param2", Map.of(
                      "description", "This is the second parameter"
                      "type", "string"
                 )
             );
             required = new String[] { "param1", "param2" };
         }

     public Map< String, String > execute(@NotNull ConversationHandler conversation, @NotNull Map args) {
         // Do something with the arguments
         }
     }
Then register the function in the FunctionManager class. 
     FunctionManager.registerFunction("myFunction", new MyFunction());
Finally, the function will be automatically called by the OpenAI Assistant (If you register the function for an NPC).

  • Field Details

  • Constructor Details

    • CustomFunction

      public CustomFunction()

  • Method Details

    • execute

      public abstract Map<String,String> execute(@NotNull @NotNull ConversationHandler conversation, @NotNull @NotNull Map<String,Object> args)
      Execute the function. This method will be called by the OpenAI Assistant in a conversation.
      Parameters:
      conversation - The conversation handler
      args - The arguments
      Returns:
      The result you want to tell the OpenAI assistant