Mercurial > hg > mpdl-group

comparison software/mpdl-services/mpiwg-mpdl-lt/src/de/mpg/mpiwg/berlin/mpdl/lt/text/norm/lang/MpdlNormalizerLexFR.lex @ 19:4a3641ae14d2

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Erstellung
author Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
date Wed, 09 Nov 2011 15:32:05 +0100
parents
children e845310098ba
comparison
equal deleted inserted replaced
18:dc5e9fcb3fdc 19:4a3641ae14d2
1 /*
2 * Normalization rules for French text
3 * [this is a JFlex specification]
4 *
5 * Wolfgang Schmidle
6 * version 2011-07-12
7 *
8 */
9
10 package de.mpg.mpiwg.berlin.mpdl.lt.analyzer.lang;
11
12 %%
13
14 %public
15 %class MpdlNormalizerLexFR
16 %type java.lang.String
17 %unicode
18
19 // French: fr
20
21 %states DISP, DICT, SEARCH
22 %state CELEX
23
24 %{
25 private String original = "";
26 private String normalized = "";
27 private int problem = 0;
28
29 private void add (String norm) {
30 original += yytext();
31 normalized += norm;
32 }
33
34 private static final String LB = "[\u002d\u00ad] ";
35 %}
36
37 hyphen = [-\u{00ad}] // hyphen and soft hyphen
38 LB = {hyphen} \u0020
39 // lb = ({hyphen} \u0020)?
40
41 END = \n
42
43 Alphabet = [abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ]
44
45 %%
46
47 <DISP, DICT, SEARCH, CELEX> {
48
49 ſ { add("s"); }
50 ß { add("ss"); }
51 æ { add("ae"); }
52
53 }
54
55 <CELEX> {
56
57 [œŒ] { add("oe"); }
58 [áàâ] { add("a"); }
59 [éèê] { add("e"); }
60 [íìî] { add("i"); }
61 [óòô] { add("o"); }
62 [úùû] { add("u"); }
63 ’ { add(""); }
64
65 {Alphabet} { add(yytext()); }
66
67 . { problem = 1; add(yytext()); } // in particular "@"
68
69 }
70
71 // default
72
73 @ { problem = 1; add(yytext()); }
74 {LB} { add(yytext()); }
75 . { add(yytext()); }
76
77
78 <DISP> {
79
80 {END} {
81 switch (problem) {
82 case 1: return original;
83 default: return normalized;
84 }
85 }
86 }
87
88 <DICT, CELEX> {
89
90 {END} {
91 switch (problem) {
92 case 1: return "";
93 default: return normalized.replaceAll(LB, "");
94 }
95 }
96 }
97
98 <SEARCH> {
99
100 {END} {
101 switch (problem) {
102 case 1: return original;
103 default: return normalized.replaceAll(LB, "").toLowerCase();
104 }
105 }
106 }
107
108
109 /*
110
111 Annahmen:
112 - die Routine wird wortweise aufgerufen, mit einem \n am Ende des Strings
113 - Zeilenumbrüche innerhalb des Wortes sind als hyphen/soft hyphen plus space markiert
114
115 TO DO:
116
117 FR: richtig? vollständig?
118
119 */