text2pa4chinese

 text2pa4chinese: Convert a sentence into phonetic alphabets (in Hanyu Pinyin) using maximum matching
    Usage: hanyuSeq=senence2hanyu(sentence, wpaList)

    For example:
        sentence='「海角七號」的茂伯是國寶ㄟ！';
        wpaList=wpaRead('chinese.hanyu.wpa');
        hanyuSeq=text2pa4chinese(sentence, wpaList);
        fprintf('Input sentence = %s\n', sentence);
        fprintf('Output:\n');
        for i=1:length(hanyuSeq)
            fprintf('%s ===> %s\n', hanyuSeq(i).word, hanyuSeq(i).pa);
        end

0001 function hanyuSeq=text2pa4chinese(sentence, wpaList)
0002 % text2pa4chinese: Convert a sentence into phonetic alphabets (in Hanyu Pinyin) using maximum matching
0003 %    Usage: hanyuSeq=senence2hanyu(sentence, wpaList)
0004 %
0005 %    For example:
0006 %        sentence='「海角七號」的茂伯是國寶ㄟ！';
0007 %        wpaList=wpaRead('chinese.hanyu.wpa');
0008 %        hanyuSeq=text2pa4chinese(sentence, wpaList);
0009 %        fprintf('Input sentence = %s\n', sentence);
0010 %        fprintf('Output:\n');
0011 %        for i=1:length(hanyuSeq)
0012 %            fprintf('%s ===> %s\n', hanyuSeq(i).word, hanyuSeq(i).pa);
0013 %        end
0014 
0015 % Roger Jang, 20081025
0016 
0017 if nargin<1, selfdemo; return; end
0018 if nargin<2, wpaList=wpaRead('chinese.hanyu.wpa'); end
0019 wordList={wpaList.word};
0020 
0021 % Method: Forward maximum-matching
0022 outIndex=[];
0023 pos=1;
0024 while (pos<=length(sentence))
0025     found=0;
0026     for len=5:-1:1
0027         start=pos;
0028         stop=min(pos+len-1, length(sentence));
0029         str=sentence(start:stop);
0030     %    fprintf('%s ', str);
0031         index=find(strcmp(wordList, str));
0032         if ~isempty(index)
0033             outIndex=[outIndex, index(1)];
0034             pos=pos+len;
0035             found=1;
0036     %        fprintf('===> bingo!\n');
0037             break;
0038         end
0039     end
0040     if ~found
0041         pos=pos+1;
0042     end
0043 end
0044 
0045 for i=1:length(outIndex)
0046     hanyuSeq(i).word=wpaList(outIndex(i)).word;
0047     temp=split(wpaList(outIndex(i)).pa, '#');
0048     hanyuSeq(i).pa=temp{1};    % 注意：抓第一個高頻字，但對於單字詞，還是可能抓到不正確的破音字！
0049 end
0050 
0051 % ====== Self demo
0052 function selfdemo
0053 sentence='我到廟口吃小吃';
0054 hanyuSeq=text2pa4chinese(sentence);
0055 fprintf('Input = %s\n', sentence);
0056 fprintf('Output = %s\n', join({hanyuSeq.pa}, '#'));
0057 sentence='我們三人參加會議';
0058 hanyuSeq=text2pa4chinese(sentence);
0059 fprintf('Input = %s\n', sentence);
0060 fprintf('Output = %s\n', join({hanyuSeq.pa}, '#'));
0061 sentence='「海角七號」的茂伯是國寶ㄟ！';
0062 hanyuSeq=text2pa4chinese(sentence);
0063 fprintf('Input = %s\n', sentence);
0064 fprintf('Output = %s\n', join({hanyuSeq.pa}, '#'));
0065