annotate software/mpdl-services/mpiwg-mpdl-lt/bin/de/mpg/mpiwg/berlin/mpdl/lt/text/norm/lang/MpdlNormalizerLexTemplate.lex @ 23:e845310098ba

diverse Korrekturen
author Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
date Tue, 27 Nov 2012 12:35:19 +0100
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
23
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
1 /*
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
2 * Template for normalization rules
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
3 * [this is a JFlex specification]
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
4 *
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
5 * Wolfgang Schmidle
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
6 * version 2011-07-12
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
7 *
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
8 */
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
9
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
10 package de.mpg.mpiwg.berlin.mpdl.lt.analyzer.lang;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
11
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
12 %%
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
13
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
14 %public
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
15 %class MpdlNormalizerLexTemplate
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
16 %type java.lang.String
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
17 %unicode
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
18
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
19 // Language: list of ISO codes
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
20
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
21 %states DISP, DICT, SEARCH
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
22
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
23 %{
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
24 private String original = "";
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
25 private String normalized = "";
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
26 private int problem = 0;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
27
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
28 private void add (String norm) {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
29 original += yytext();
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
30 normalized += norm;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
31 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
32
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
33 private static final String LB = "[\u002d\u00ad] ";
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
34 %}
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
35
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
36 hyphen = [-\u{00ad}] // hyphen and soft hyphen
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
37 LB = {hyphen} \u0020
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
38 // lb = ({hyphen} \u0020)?
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
39
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
40 END = \n
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
41
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
42 %%
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
43
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
44 <DISP, DICT, SEARCH> {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
45
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
46 ſ { add("s"); } // sample rule
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
47
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
48 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
49
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
50
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
51 // default rules
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
52
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
53 @ { problem = 1; add(yytext()); }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
54 {LB} { add(yytext()); }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
55 . { add(yytext()); }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
56
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
57
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
58 // at the end, determine which string to return
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
59
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
60 <DISP> {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
61
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
62 {END} {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
63 switch (problem) {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
64 case 1: return original;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
65 default: return normalized;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
66 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
67 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
68 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
69
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
70 <DICT> {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
71
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
72 {END} {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
73 switch (problem) {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
74 case 1: return "";
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
75 default: return normalized.replaceAll(LB, "");
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
76 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
77 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
78 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
79
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
80 <SEARCH> {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
81
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
82 {END} {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
83 switch (problem) {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
84 case 1: return original;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
85 default: return normalized.replaceAll(LB, "").toLowerCase();
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
86 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
87 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
88 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
89