我正在尝试结合以前堆栈溢出问题中的一些纯谓词来制作我自己的谓词。
我想给出一个 c 列表(与它们相关联的事实 - 'ats')和一个 'feature' 术语,它有一个运算符和一个 'at' 的阈值。我想对 c 的列表进行分区,如果 c 没有来自“功能”的相应“at”,它将进入错误分区,否则操作员将测试该“c”的“at”并拆分c 合适。
例如:
?-cpgpartition_ts_fs_feature([c1,c2,c3],Ts,Fs,feature(at2,_,>=,10)).
应该导致:
Ts = [c3], %c3 has an at2 >= 10
Fs = [c1,c2]. %c1 has at2 <10 and c2 does not have an at2
这是我的代码:
:-use_module(library(clpfd)).
cpgpartition_ts_fs_feature([],[],[],_).
cpgpartition_ts_fs_feature([X|Xs0],Ts,Fs,Feature):-
Feature = feature(At,_,Op,FValue),
cpg_ats_i(X,AtList),
atom_concat(#,Op,Op2), %make clpfd operator
Test =..[Op2,AtValue3,FValue],
if_(memberd_t(attribute(At,AtValue3),AtList),
(
if_(call(Test), (Ts=[X|Ts0],Fs=Fs0),
( Ts =Ts0,Fs=[X|Fs0]))
)
,Fs=[X|Fs0]),
cpgpartition_ts_fs_feature(Xs0,Ts0,Fs0,Feature).
if_(If_1, Then_0, Else_0) :-
call(If_1, T),
( T == true -> call(Then_0)
; T == false -> call(Else_0)
; nonvar(T) -> throw(error(type_error(boolean,T),_))
; /* var(T) */ throw(error(instantiation_error,_))
).
bool01_t(1,true).
bool01_t(0,false).
=(X, Y, T) :-
( X == Y -> T = true
; X \= Y -> T = false
; T = true, X = Y
; T = false,
dif(X, Y) % ISO extension
% throw(error(instantiation_error,_)) % ISO strict
).
#=<(X,Y,Truth) :- X #=< Y #<==> B, bool01_t(B,Truth).
#<( X,Y,Truth) :- X #< Y #<==> B, bool01_t(B,Truth).
#>( X,Y,Truth) :- X #> Y #<==> B, bool01_t(B,Truth).
#>=(X,Y,Truth) :- X #>= Y #<==> B, bool01_t(B,Truth).
list_memberd_t([] ,_,false).
list_memberd_t([Y|Ys],X,Truth) :-
if_(X=Y, Truth=true, list_memberd_t(Ys,X,Truth)).
list_memberd_truth(Xs,X,Truth) :- list_memberd_t(Xs,X,Truth).
memberd_t(X,Xs,Truth) :- list_memberd_t(Xs,X,Truth).
value_intvalue(attribute(_A,X),attribute(_A,Y)):-
AtValue2 is X *100, %Convert decimal number to integer.
Y is integer(AtValue2).
cpg_ats_i(C,AtList):-
cpg_ats(C,Ats),
maplist(value_intvalue,Ats,AtList).
cpg_ats(c1,[attribute(at1,0.5),attribute(at2,0.03)]).
cpg_ats(c2,[attribute(at1,0.02)]).
cpg_ats(c3,[attribute(at2,0.1),attribute(at3,0.04),attribute(at4,0.08)]).
尝试测试查询时,我得到:
cpgpartition_ts_fs_feature([c1,c2,c3],Ts,Fs,feature(at2,_,>=,10)).
Fs = [c1, c2] ;
Fs = [c1, c2, c3] ;
Fs = [c1, c2] ;
Fs = [c1, c2, c3].
有趣的是,如果 clist 的顺序不同,结果会发生变化。
?- cpgpartition_ts_fs_feature([c3,c1,c2],Ts,Fs,feature(at2,_,>=,10)).
Ts = [c3|_12950],
Fs = [c1, c2] ;
Ts = [c3|_12950],
Fs = [c1, c2] ;
Fs = [c3, c1, c2] ;
Fs = [c3, c1, c2].
我认为这是因为以下查询返回带有dif/2
约束的结果,这些约束似乎不适合我正在尝试做的事情,我只想要具体的解决方案。
?- cpg_ats_i(C,Ats), if_(memberd_t(attribute(at2,AtValue),Ats),Q=true,Q=false).
C = c1,
Ats = [attribute(at1, 50), attribute(at2, 3)],
AtValue = 3,
Q = true ;
C = c1,
Ats = [attribute(at1, 50), attribute(at2, 3)],
Q = false,
dif(AtValue, 3) ;
C = c2,
Ats = [attribute(at1, 2)],
Q = false ;
C = c3,
Ats = [attribute(at2, 10), attribute(at3, 4), attribute(at4, 8)],
AtValue = 10,
Q = true ;
C = c3,
Ats = [attribute(at2, 10), attribute(at3, 4), attribute(at4, 8)],
Q = false,
dif(AtValue, 10).
此外,目标是让这段代码在大量数据上运行,c 的列表将有数十万的长度,每个 c 可能有 50k 的 ats,我该如何计算内存需求?使用不纯谓词的不同方法是否可能占用更少的内存?