parser.pro

经典的基于Context-Free_Grammer 的语法分析器。

% Author:
% Date: 11/29/2005
:-module(parser,[parse/2,parse/0]).
:-use_module('rule.pro').
:-use_module('tokenizer.pro').
:-op(1000,xfy,'=>').


%testing command.
%parse:-forall(parse('a mouse chase the pig',X),(write(X),nl)).
parse:-forall(parse('ASF/SF2 is a protein that plays important roles in premRNA splicing',X),(write(X),nl)).

%parse(+Sentence_list,-Output), input a list of words as a sentence
parse(SentStr,S):-  tokenize(SentStr,Sentence),
                    reset, %erase all the edges.
                    process(Sentence,1),
                    getlength(Sentence,Length), % get the length of the sentence
                   % Length1 is Length+1,
                    edge(_,_,S),
                    getType(S,[a(_)]) .
getType(X,X).

reset:-retractall(edge(_,_,_)).

                    
process([Head|Tail],L):- L1 is L+1,
                         addedge(edge(L,L1,[Head])),
                         scan(L1,L1,[]),
                         process(Tail,L1),
                         fail.

process(_,_).

%scan(+Start,+L,+Clist), Start: the start position; L: scan location; Clist: completed list
scan(Start,L,Clist):- foreach((Clist=>Clist_up),
                              (addedge(edge(L,Start,Clist_up)),scan(Start,L,Clist_up))),
                      edge(L1,L,List),/*read edge C*/

                      append(List,Clist,NClist), /*append C to Clist*/
                      % write([a,NClist]),nl,
                      /*for each rule that start from NClist, add edge and start new scan*/
                      foreach((NClist=>NClist_up),
                              (addedge(edge(L1,Start,NClist_up)),scan(Start,L1,NClist_up))),

                      scan(Start,L1,NClist),fail.
scan(_,_,_).

addedge(X):- \+(X),assertz(X).


getlength([_|Tail],L):- getlength(Tail,L1),L is L1+1.
getlength([],0).

seeEdge:-edge(X,Y,Z),write([X,Y,Z]),nl,fail.
seeEdge.

foreach(X,Y):-X,once(Y),fail.
foreach(_,_).