/* Part of SWI-Prolog Author: Jan Wielemaker E-mail: J.Wielemaker@vu.nl WWW: http://www.swi-prolog.org Copyright (c) 2008-2022, University of Amsterdam, VU University SWI-Prolog Solutions b.v. Amsterdam All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: 1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. 2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ :- module(terms, [ term_hash/2, % @Term, -HashKey term_hash/4, % @Term, +Depth, +Range, -HashKey term_size/2, % @Term, -Size term_variables/2, % @Term, -Variables term_variables/3, % @Term, -Variables, +Tail variant/2, % @Term1, @Term2 subsumes/2, % +Generic, @Specific subsumes_chk/2, % +Generic, @Specific cyclic_term/1, % @Term acyclic_term/1, % @Term term_subsumer/3, % +Special1, +Special2, -General term_factorized/3, % +Term, -Skeleton, -Subsitution mapargs/3, % :Goal, ?Term1, ?Term2 mapsubterms/3, % :Goal, ?Term1, ?Term2 mapsubterms_var/3, % :Goal, ?Term1, ?Term2 foldsubterms/4, % :Goal, +Term, +State0, -State foldsubterms/5, % :Goal, +Term1, ?Term2, +State0, -State same_functor/2, % ?Term1, ?Term2 same_functor/3, % ?Term1, ?Term2, -Arity same_functor/4 % ?Term1, ?Term2, ?Name, ?Arity ]). :- meta_predicate mapargs(2,?,?), mapsubterms(2,?,?), mapsubterms_var(2,?,?), foldsubterms(3,+,+,-), foldsubterms(4,+,?,+,-). :- autoload(library(rbtrees), [ rb_empty/1, rb_lookup/3, rb_insert/4, rb_new/1, rb_visit/2, ord_list_to_rbtree/2, rb_update/5 ]). :- autoload(library(error), [instantiation_error/1]). /** Term manipulation Compatibility library for term manipulation predicates. Most predicates in this library are provided as SWI-Prolog built-ins. @compat YAP, SICStus, Quintus. Not all versions of this library define exactly the same set of predicates, but defined predicates are compatible. */ %! term_size(@Term, -Size) is det. % % True if Size is the size in _cells_ occupied by Term on the % global (term) stack. A _cell_ is 4 bytes on 32-bit machines and % 8 bytes on 64-bit machines. The calculation does take _sharing_ % into account. For example: % % ``` % ?- A = a(1,2,3), term_size(A,S). % S = 4. % ?- A = a(1,2,3), term_size(a(A,A),S). % S = 7. % ?- term_size(a(a(1,2,3), a(1,2,3)), S). % S = 11. % ``` % % Note that small objects such as atoms and small integers have a % size 0. Space is allocated for floats, large integers, strings % and compound terms. term_size(Term, Size) :- '$term_size'(Term, _, Size). %! variant(@Term1, @Term2) is semidet. % % Same as SWI-Prolog =|Term1 =@= Term2|=. variant(X, Y) :- X =@= Y. %! subsumes_chk(@Generic, @Specific) % % True if Generic can be made equivalent to Specific without % changing Specific. % % @deprecated Replace by subsumes_term/2. subsumes_chk(Generic, Specific) :- subsumes_term(Generic, Specific). %! subsumes(+Generic, @Specific) % % True if Generic is unified to Specific without changing % Specific. % % @deprecated It turns out that calls to this predicate almost % always should have used subsumes_term/2. Also the name is % misleading. In case this is really needed, one is adviced to % follow subsumes_term/2 with an explicit unification. subsumes(Generic, Specific) :- subsumes_term(Generic, Specific), Generic = Specific. %! term_subsumer(+Special1, +Special2, -General) is det. % % General is the most specific term that is a generalisation of % Special1 and Special2. The implementation can handle cyclic % terms. % % @compat SICStus % @author Inspired by LOGIC.PRO by Stephen Muggleton % It has been rewritten by Jan Wielemaker to use the YAP-based % red-black-trees as mapping rather than flat lists and use arg/3 % to map compound terms rather than univ and lists. term_subsumer(S1, S2, G) :- cyclic_term(S1), cyclic_term(S2), !, rb_empty(Map), lgg_safe(S1, S2, G, Map, _). term_subsumer(S1, S2, G) :- rb_empty(Map), lgg(S1, S2, G, Map, _). lgg(S1, S2, G, Map0, Map) :- ( S1 == S2 -> G = S1, Map = Map0 ; compound(S1), compound(S2), functor(S1, Name, Arity), functor(S2, Name, Arity) -> functor(G, Name, Arity), lgg(0, Arity, S1, S2, G, Map0, Map) ; rb_lookup(S1+S2, G0, Map0) -> G = G0, Map = Map0 ; rb_insert(Map0, S1+S2, G, Map) ). lgg(Arity, Arity, _, _, _, Map, Map) :- !. lgg(I0, Arity, S1, S2, G, Map0, Map) :- I is I0 + 1, arg(I, S1, Sa1), arg(I, S2, Sa2), arg(I, G, Ga), lgg(Sa1, Sa2, Ga, Map0, Map1), lgg(I, Arity, S1, S2, G, Map1, Map). %! lgg_safe(+S1, +S2, -G, +Map0, -Map) is det. % % Cycle-safe version of the above. The difference is that we % insert compounds into the mapping table and check the mapping % table before going into a compound. lgg_safe(S1, S2, G, Map0, Map) :- ( S1 == S2 -> G = S1, Map = Map0 ; rb_lookup(S1+S2, G0, Map0) -> G = G0, Map = Map0 ; compound(S1), compound(S2), functor(S1, Name, Arity), functor(S2, Name, Arity) -> functor(G, Name, Arity), rb_insert(Map0, S1+S2, G, Map1), lgg_safe(0, Arity, S1, S2, G, Map1, Map) ; rb_insert(Map0, S1+S2, G, Map) ). lgg_safe(Arity, Arity, _, _, _, Map, Map) :- !. lgg_safe(I0, Arity, S1, S2, G, Map0, Map) :- I is I0 + 1, arg(I, S1, Sa1), arg(I, S2, Sa2), arg(I, G, Ga), lgg_safe(Sa1, Sa2, Ga, Map0, Map1), lgg_safe(I, Arity, S1, S2, G, Map1, Map). %! term_factorized(+Term, -Skeleton, -Substiution) % % Is true when Skeleton is Term where all subterms that appear % multiple times are replaced by a variable and Substitution is a % list of Var=Value that provides the subterm at the location Var. % I.e., After unifying all substitutions in Substiutions, Term == % Skeleton. Term may be cyclic. For example: % % == % ?- X = a(X), term_factorized(b(X,X), Y, S). % Y = b(_G255, _G255), % S = [_G255=a(_G255)]. % == term_factorized(Term, Skeleton, Substitutions) :- rb_new(Map0), add_map(Term, Map0, Map), rb_visit(Map, Counts), common_terms(Counts, Common), ( Common == [] -> Skeleton = Term, Substitutions = [] ; ord_list_to_rbtree(Common, SubstAssoc), insert_vars(Term, Skeleton, SubstAssoc), mk_subst(Common, Substitutions, SubstAssoc) ). add_map(Term, Map0, Map) :- ( primitive(Term) -> Map = Map0 ; rb_update(Map0, Term, Old, New, Map) -> New is Old+1 ; rb_insert(Map0, Term, 1, Map1), assoc_arg_map(1, Term, Map1, Map) ). assoc_arg_map(I, Term, Map0, Map) :- arg(I, Term, Arg), !, add_map(Arg, Map0, Map1), I2 is I + 1, assoc_arg_map(I2, Term, Map1, Map). assoc_arg_map(_, _, Map, Map). primitive(Term) :- var(Term), !. primitive(Term) :- atomic(Term), !. primitive('$VAR'(_)). common_terms([], []). common_terms([H-Count|T], List) :- !, ( Count == 1 -> common_terms(T, List) ; List = [H-_NewVar|Tail], common_terms(T, Tail) ). insert_vars(T0, T, _) :- primitive(T0), !, T = T0. insert_vars(T0, T, Subst) :- rb_lookup(T0, S, Subst), !, T = S. insert_vars(T0, T, Subst) :- functor(T0, Name, Arity), functor(T, Name, Arity), insert_arg_vars(1, T0, T, Subst). insert_arg_vars(I, T0, T, Subst) :- arg(I, T0, A0), !, arg(I, T, A), insert_vars(A0, A, Subst), I2 is I + 1, insert_arg_vars(I2, T0, T, Subst). insert_arg_vars(_, _, _, _). mk_subst([], [], _). mk_subst([Val0-Var|T0], [Var=Val|T], Subst) :- functor(Val0, Name, Arity), functor(Val, Name, Arity), insert_arg_vars(1, Val0, Val, Subst), mk_subst(T0, T, Subst). %! mapargs(:Goal, ?Term1, ?Term2) % % Term1 and Term2 have the same functor (name/arity) and for each % matching pair of arguments call(Goal, A1, A2) is true. mapargs(Goal, Term1, Term2) :- same_functor(Term1, Term2, Arity), mapargs_(1, Arity, Goal, Term1, Term2). mapargs_(I, Arity, Goal, Term1, Term2) :- I =< Arity, !, arg(I, Term1, A1), arg(I, Term2, A2), call(Goal, A1, A2), I2 is I+1, mapargs_(I2, Arity, Goal, Term1, Term2). mapargs_(_, _, _, _, _). %! mapsubterms(:Goal, +Term1, -Term2) is det. %! mapsubterms_var(:Goal, +Term1, -Term2) is det. % % Recursively map sub terms of Term1 into subterms of Term2 for every % pair for which call(Goal, ST1, ST2) succeeds. Procedurably, the % mapping for each (sub) term pair `T1/T2` is defined as: % % - If `T1` is a variable % - mapsubterms/3 unifies `T2` with `T1`. % - mapsubterms_var/3 treats variables as other terms. % - If call(Goal, T1, T2) succeeds we are done. Note that the % mapping does not continue in `T2`. If this is desired, `Goal` % must call mapsubterms/3 explicitly as part of its conversion. % - If `T1` is a dict, map all values, i.e., the _tag_ and _keys_ % are left untouched. % - If `T1` is a list, map all elements, i.e., the list structure % is left untouched. % - If `T1` is a compound, use same_functor/3 to instantiate `T2` % and recurse over the term arguments left to right. % - Otherwise `T2` is unified with `T1`. % % Both predicates are implemented using foldsubterms/5. mapsubterms(Goal, Term1, Term2) :- foldsubterms(map2(Goal), Term1, Term2, _, _). mapsubterms_var(Goal, Term1, Term2) :- foldsubterms(map2_var(Goal), Term1, Term2, _, _). map2(Goal, Term1, Term2, _, _) :- nonvar(Term1), call(Goal, Term1, Term2). map2_var(Goal, Term1, Term2, _, _) :- call(Goal, Term1, Term2). %! foldsubterms(:Goal3, +Term1, +State0, -State) is semidet. %! foldsubterms(:Goal4, +Term1, ?Term2, +State0, -State) is semidet. % % The predicate foldsubterms/5 calls call(Goal4, SubTerm1, SubTerm2, % StateIn, StateOut) for each subterm, including variables, in Term1. % If this call fails, `StateIn` and `StateOut` are the same. This % predicate may be used to map subterms in a term while collecting % state about the mapped subterms. The foldsubterms/4 variant does not % map the term. foldsubterms(Goal, Term1, State0, State) :- foldsubterms(fold1(Goal), Term1, _, State0, State). fold1(Goal, Term1, _Term2, State0, State) :- call(Goal, Term1, State0, State). foldsubterms(Goal, Term1, Term2, State0, State) :- call(Goal, Term1, Term2, State0, State), !. foldsubterms(Goal, Term1, Term2, State0, State) :- is_dict(Term1), !, dict_pairs(Term1, Tag, Pairs1), fold_dict_pairs(Pairs1, Pairs2, Goal, State0, State), dict_pairs(Term2, Tag, Pairs2). foldsubterms(Goal, Term1, Term2, State0, State) :- is_list(Term1), !, fold_some(Term1, Term2, Goal, State0, State). foldsubterms(Goal, Term1, Term2, State0, State) :- compound(Term1), !, same_functor(Term1, Term2, Arity), foldsubterms_(1, Arity, Goal, Term1, Term2, State0, State). foldsubterms(_, Term, Term, State, State). fold_dict_pairs([], [], _, State, State). fold_dict_pairs([K-V0|T0], [K-V|T], Goal, State0, State) :- foldsubterms(Goal, V0, V, State0, State1), fold_dict_pairs(T0, T, Goal, State1, State). fold_some([], [], _, State, State). fold_some([H0|T0], [H|T], Goal, State0, State) :- foldsubterms(Goal, H0, H, State0, State1), fold_some(T0, T, Goal, State1, State). foldsubterms_(I, Arity, Goal, Term1, Term2, State0, State) :- I =< Arity, !, arg(I, Term1, A1), arg(I, Term2, A2), foldsubterms(Goal, A1, A2, State0, State1), I2 is I+1, foldsubterms_(I2, Arity, Goal, Term1, Term2, State1, State). foldsubterms_(_, _, _, _, _, State, State). %! same_functor(?Term1, ?Term2) is semidet. %! same_functor(?Term1, ?Term2, -Arity) is semidet. %! same_functor(?Term1, ?Term2, ?Name, ?Arity) is semidet. % % True when Term1 and Term2 are terms that have the same functor % (Name/Arity). The arguments must be sufficiently instantiated, which % means either Term1 or Term2 must be bound or both Name and Arity % must be bound. % % If Arity is 0, Term1 and Term2 are unified with Name for % compatibility. % % @compat SICStus same_functor(Term1, Term2) :- same_functor(Term1, Term2, _Name, _Arity). same_functor(Term1, Term2, Arity) :- same_functor(Term1, Term2, _Name, Arity). same_functor(Term1, Term2, Name, Arity) :- ( nonvar(Term1) -> functor(Term1, Name, Arity, Type), functor(Term2, Name, Arity, Type) ; nonvar(Term2) -> functor(Term2, Name, Arity, Type), functor(Term1, Name, Arity, Type) ; functor(Term2, Name, Arity), functor(Term1, Name, Arity) ).