context recall vs x apply op
Now, here is a key part of the project that I haven't mentioned yet (though there are plenty of others!). We have a wrapper around context.recall(). I don't exactly recall the reason why, since I implemented this so long ago, but there we have it.
So in the new_context() class (this being the beasty that stores all state in a convoluted hash table, something I'm hoping to replace with something more elegant later), we have two key pieces: context.learn(a,b,c) and context.recall(a,b).
The first of these does the hard work in learning knowledge.
The second of these does the hard work when answering questions:
coeff = label.value if type(label) == ket else 1
# coeff = 1 # use this to switch off the multiply(coeff) feature
op.label.split("op: ")[-1] if type(op) == ket else op
label = label.label if type(label) == ket else label
op is the operator, either ket("op: friends") or a direct string: "friends"
label is the ket label from our learn rule, either ket("Fred") or a direct string "Fred"
active is a variable that keeps track of when we want to activate stored rules.
OK. So we have that. Now in the ket() class we have:
And in the superposition() class we have:
result = superposition()
for x in self.data:
result += context.recall(op,x,True)
where in the superposition class, self.data stores the list of kets (perhaps later we will re-implement this list of kets as an ordered dictionary because the list representation sometimes has terrible big-O), and so this for loop is what implements the linearity of operators.
Where by "linearity of operators", recall:
op1 |x> => |a> + |b> + |c> + |d> + |e>
op2 op1 |x>
= op2 (|a> + |b> + |c> + |d> + |e>)
= op2 |a> + op2 |b> + op2 |c> + op2 |d> + op2 |e>
So what point am I trying to make? Just trying to make it clearer what cell.apply_op() meant in the walking-our-grid post.
Anyway, a couple of examples:
sa: friends |Fred>
is translated to this python:
op (2|x> + 3|y> + |z>)
is translated to this python (noting that when you add kets you get a superposition):
(ket("x",2) + ket("y",3) + ket("z")).apply_op(context,"op")
I guess that is enough for now. I hope things are a little clearer!
previous: temperature conversion
next: introducing file recall
by Garry Morrison
email: garry -at- semantic-db.org