thdltools/Jskad
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

TMW->Wylie conversion now takes advantage of prefix rules, the rules

Browse source 
that say "ya can take a ga prefix" etc.

The ACIP->Unicode converter now gives warnings (optionally, and by
default, inline).  This converter now produces output even when
lexical errors occur, but the output has errors and warnings inline.
This commit is contained in:
dchandler 2003-08-23 22:03:37 +00:00
parent 21ef657921
commit d5ad760230
14 changed files with 678 additions and 270 deletions
Download patch file Download diff file Expand all files Collapse all files

84
source/org/thdl/tib/text/TGCPair.java
View file

@ -25,7 +25,7 @@ package org.thdl.tib.text;
context-insensitive THDL Extended Wylie representation. NOTE
WELL: this is not a real grapheme cluster; I'm misusing the term
(FIXME). It's actually whole or part of one. It's part of one
when this is a vowel or U+0F7F alone.
when this is U+0F7F alone.
@author David Chandler */
public class TGCPair {
@ -37,14 +37,84 @@ public class TGCPair {
public static final int SANSKRIT_WITHOUT_VOWEL = 5;
public static final int SANSKRIT_WITH_VOWEL = 6;
public String wylie;
public int classification;
public TGCPair(String wylie, int classification) {
this.wylie = wylie;
this.classification = classification;
public static final int TYPE_OTHER = 31;
public static final int TYPE_SANSKRIT = 32;
public static final int TYPE_TIBETAN = 33;
// Sanskrit or Tibetan consonant, or number, or oddball:
private String consonantWylie;
private String vowelWylie;
public String getConsonantWylie() {
return consonantWylie;
}
public String getVowelWylie() {
return vowelWylie;
}
/** Cludge. */
public void setWylie(String x) {
consonantWylie = x;
vowelWylie = null;
}
public String getWylie() {
StringBuffer b = new StringBuffer();
if (consonantWylie != null) {
// we may have {p-y}, but the user wants to see {py}.
for (int i = 0; i < consonantWylie.length(); i++) {
char ch = consonantWylie.charAt(i);
if ('-' != ch)
b.append(ch);
}
}
if (vowelWylie != null)
b.append(vowelWylie);
return b.toString();
}
public int classification;
/** Constructs a new TGCPair with (Tibetan or Sanskrit) consonant
* consonantWylie and vowel vowelWylie. Use
* classification==TYPE_OTHER for numbers, lone vowels, marks,
* etc. Use classification==TYPE_TIBETAN for Tibetan (not
* Tibetanized Sanskrit) and classification=TYPE_SANSKRIT for
* Tibetanized Sanskrit. */
public TGCPair(String consonantWylie, String vowelWylie, int classification) {
if ("".equals(vowelWylie))
vowelWylie = null;
// Technically, we don't need the following check, but it's
// nice for consistency's sake.
if ("".equals(consonantWylie))
consonantWylie = null;
// DLC FIXME: for speed, make these assertions:
if (classification != TYPE_OTHER
&& classification != TYPE_TIBETAN
&& classification != TYPE_SANSKRIT) {
throw new IllegalArgumentException("Bad classification " + classification + ".");
}
int realClassification = -37;
if (vowelWylie == null && classification == TYPE_TIBETAN)
realClassification = CONSONANTAL_WITHOUT_VOWEL;
if (vowelWylie != null && classification == TYPE_TIBETAN)
realClassification = CONSONANTAL_WITH_VOWEL;
if (vowelWylie == null && classification == TYPE_SANSKRIT)
realClassification = SANSKRIT_WITHOUT_VOWEL;
if (vowelWylie != null && classification == TYPE_SANSKRIT)
realClassification = SANSKRIT_WITH_VOWEL;
if (consonantWylie == null) {
if (classification != TYPE_OTHER)
throw new IllegalArgumentException("That's the very definition of a lone vowel.");
realClassification = LONE_VOWEL;
} else {
if (classification == TYPE_OTHER)
realClassification = OTHER;
}
this.consonantWylie = consonantWylie;
this.vowelWylie = vowelWylie;
this.classification = realClassification;
}
public String toString() {
return "<TGCPair wylie=" + wylie + " classification="
return "<TGCPair wylie=" + getWylie() + " classification="
+ classification + "/>";
}
}