[DOC] add lists as part of a belief by constructing a literal with a tuple representing the list

[ccarstens]

A belief containing a list as an argument can be constructed by passing tuples representing the list to the Literal constructor. e.g

my_literal = Literal("my_functor", ((12, 45), ))

is equivalent to the literal within ASL

my_functor([12, 45])

[niklasf]

That's correct. The embedding is as follows:

Python	Agentspeak
int, float	number
string	string
tuple, agentspeak.LinkedList	list
agentspeak.Literal	literal
agentspeak.Var, agentspeak.Wildcard	variable
any other object*	opaque

*) That means it's valid to create a Literal("foo", (MyObject(), )). On the AgentSpeak side MyObject is completely opaque, but it can be used in unification and passed to Python actions that might be able to do something with it.

[ccarstens]

About agentspeak.LinkedList, I saw that the constructor takes head and tail as arguments.

LinkedList([1, 2, 3], [4, 5, 6])

results in

[
    [1, 2, 3],
    [4, 5, 6]
]

and not in

[1, 2, 3, 4, 5, 6]

*) That means it's valid to create a Literal("foo", (MyObject(), )). On the AgentSpeak side MyObject is completely opaque, but it can be used in unification and passed to Python actions that might be able to do something with it.

That is incredibly good to know! Thank you! :)

[niklasf]

Avoid LinkedList, if you can.

It is only required to represent the [Head|Tail] syntax from AgentSpeak. We cannot use tuples for that, because Tail matches a list of unknown length.

For example:

>>> [Head|Tail] = [1,2,3,4].
Head = 1
Tail = [2, 3, 4]

The proper way to construct a LinkedList would be:

LinkedList(1, (2, 3, 4))
LinkedList(1, LinkedList(2, (3, 4)))
LinkedList(1, LinkedList(2, LinkedList(3, (4, ))))

[ccarstens]

Oh okay, I see. I guess using tuples and knowing that they'll be cast to agent speak lists does the job! :)