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

Now turns Tibetan into pa'am, not pa'm. Works with or without vowels

Browse source 
in the part preceding the 'am or 'ang, overcoming the inconsistency
that I'd put here for a short time.
This commit is contained in:
dchandler 2003-04-08 04:56:40 +00:00
parent eb71fb6075
commit 7dd67bbf6a
2 changed files with 63 additions and 72 deletions
Download patch file Download diff file Expand all files Collapse all files

129
source/org/thdl/tib/text/TibTextUtils.java
View file

@ -85,22 +85,26 @@ public class TibTextUtils implements THDLWylieConstants {
} }
/** /**
* Figures out how to arrange a list of characters into glyphs. For example, if the user types 'bsgr' * Figures out how to arrange a list of characters into glyphs. For
* using the Extended Wylie keyboard, this method figures out that this should be represented * example, if the user types 'bsgr' using the Extended Wylie keyboard,
* as a 'b' glyph followed by a 's-g-r' glyph. If you know that the characters do not * this method figures out that this should be represented as a 'b'
* contain Sanskrit stacks, or do not contain Tibetan stacks, then you can specify this * glyph followed by a 's-g-r' glyph. If you know that the characters
* to speed the process up. Otherwise, the method will first check to see if the characters * do not contain Sanskrit stacks, or do not contain Tibetan stacks,
* correspond to any Tibetan stacks, and if not, then it will check for Sanskrit stacks. * then you can specify this to speed the process up. Otherwise, the
* @param chars the list of Tibetan characters you want to find glyphs for * method will first check to see if the characters correspond to any
* @param areStacksOnRight whether stacking should try to maximize from right to left (true) * Tibetan stacks, and if not, then it will check for Sanskrit stacks.
* or from left to right (false). In the Extended Wylie keyboard, you try to stack from * @param chars the list of Tibetan characters you want to find glyphs
* right to left. Thus, the character sequence r-g-r would be stacked as r followed by gr, * for
* rather than rg followed by r. In the Sambhota and TCC keyboards, the stack direction * @param areStacksOnRight whether stacking should try to maximize from
* is reversed. * right to left (true) or from left to right (false). In the Extended
* @param definitelyTibetan should be true if the characters are known to be Tibetan and * Wylie keyboard, you try to stack from right to left. Thus, the
* not Sanskrit * character sequence r-g-r would be stacked as r followed by gr,
* @param definitelySanskrit should be true if the characters are known to be Sanskrit and * rather than rg followed by r. In the Sambhota and TCC keyboards, the
* not Tibetan * stack direction is reversed.
* @param definitelyTibetan should be true if the characters are known
* to be Tibetan and not Sanskrit
* @param definitelySanskrit should be true if the characters are known
* to be Sanskrit and not Tibetan
*/ */
public static List getGlyphs(List chars, boolean areStacksOnRight, boolean definitelyTibetan, boolean definitelySanskrit) { public static List getGlyphs(List chars, boolean areStacksOnRight, boolean definitelyTibetan, boolean definitelySanskrit) {
StringBuffer tibBuffer, sanBuffer; StringBuffer tibBuffer, sanBuffer;
@ -233,10 +237,10 @@ public class TibTextUtils implements THDLWylieConstants {
} }
/** /**
* Finds the first meaningful element to occur within a string of Extended Wylie. * Finds the first meaningful element to occur within a string of
* This could be a character, a vowel, * Extended Wylie. This could be a character, a vowel, punctuation, or
* punctuation, or formatting. For example, passed the string 'tshapo', * formatting. For example, passed the string 'tshapo', this method
* this method will return 'tsh'. * will return 'tsh'.
* @param wylie the String of wylie you want to scan * @param wylie the String of wylie you want to scan
* @return the next meaningful subpart of this string, or null if * @return the next meaningful subpart of this string, or null if
* no meaningful subpart can be found (for example 'x' has no equivalent * no meaningful subpart can be found (for example 'x' has no equivalent
@ -867,6 +871,7 @@ public class TibTextUtils implements THDLWylieConstants {
// prepend: // prepend:
tailEndWylie.insert(0, tailEndWylie.insert(0,
ACHUNG ACHUNG
+ aVowelToUseAfter(ACHUNG)
+ TibetanMachineWeb.getWylieForGlyph((DuffCode)glyphList.get(effectiveSize + 1))); + TibetanMachineWeb.getWylieForGlyph((DuffCode)glyphList.get(effectiveSize + 1)));
effectiveSize -= 2; effectiveSize -= 2;
} }
@ -938,27 +943,34 @@ public class TibTextUtils implements THDLWylieConstants {
* 9 words doesn't have * 9 words doesn't have
* any ending with d -- * any ending with d --
* all end with s. */) { * all end with s. */) {
/* Yes, this is ambiguous. How do we handle it? See this from Andres: /* Yes, this is ambiguous. How do we handle
* it? See this from Andres:
I'm posting this upon David Chandler's request. According to Lobsang *
Thonden in Modern Tibetan Grammar Language (page 42), with regards to * I'm posting this upon David Chandler's
identifying the root letter in 3 lettered words there are only 23 * request. According to Lobsang Thonden in
ambiguous cases. He writes: * Modern Tibetan Grammar Language (page 42),
* with regards to identifying the root letter
If the last letter is 'sa' and the first two letters are affixes, then * in 3 lettered words there are only 23
the SECOND ONE is the root letter in the following 9 WORDS ONLY: * ambiguous cases. He writes:
*
gdas gnas gsas dgas dmas bdas mdas 'gas 'das * If the last letter is 'sa' and the first
* two letters are affixes, then the SECOND
And the FIRST is the root letter in the following 14 WORDS ONLY: * ONE is the root letter in the following 9
* WORDS ONLY:
rags lags nags bags bangs gangs rangs langs nangs sangs *
babs rabs rams nams * gdas gnas gsas dgas dmas bdas mdas 'gas
* 'das
As I mentioned before, I think that the best solution for now is to *
hard-wire these cases. Even if the list is not exhaustive, at least * And the FIRST is the root letter in the
we'll have most cases covered. * following 14 WORDS ONLY:
*
* rags lags nags bags bangs gangs rangs langs
* nangs sangs babs rabs rams nams
*
* As I mentioned before, I think that the
* best solution for now is to hard-wire these
* cases. Even if the list is not exhaustive,
* at least we'll have most cases covered.
*/ */
/* FIXME: these constants are hard-wired here, /* FIXME: these constants are hard-wired here,
@ -978,36 +990,6 @@ public class TibTextUtils implements THDLWylieConstants {
wylie2)); wylie2));
} }
// DLC FIXME: what about ambiguity between
// wa-zur and wa? dwa vs. d.wa, e.g.?
// DLC FIXME: disambiguators are needed for
// this case too, as b.lag vs. blag
// illustrates. Use something based on this,
// from LegalTshegBar.java:
//
// boolean disambiguatorNeeded = false;
// char prefix = getPrefix();
// sb.append(UnicodeCodepointToThdlWylie.getThdlWylieForUnicodeCodepoint(prefix));
// if (!hasHeadLetter()) {
// if (EWC_ya == rootLetter) {
// if (isConsonantThatTakesYaBtags(prefix))
// disambiguatorNeeded = true;
// } else if (EWC_ra == rootLetter) {
// if (isConsonantThatTakesRaBtags(prefix))
// disambiguatorNeeded = true;
// } else if (EWC_la == rootLetter) {
// if (isConsonantThatTakesLaBtags(prefix))
// disambiguatorNeeded = true;
// } else if (EWC_wa == rootLetter) {
// if (isConsonantThatTakesWaZur(prefix))
// disambiguatorNeeded = true;
// }
// }
// if (disambiguatorNeeded)
// sb.append(WYLIE_DISAMBIGUATING_KEY);
} else { } else {
/* no ambiguity. the "a" vowel comes after /* no ambiguity. the "a" vowel comes after
* wylie1. */ * wylie1. */
@ -1076,6 +1058,11 @@ public class TibTextUtils implements THDLWylieConstants {
&& currWylie.equals(ACHEN))) && currWylie.equals(ACHEN)))
sb.append(WYLIE_DISAMBIGUATING_KEY); sb.append(WYLIE_DISAMBIGUATING_KEY);
/* le'ang, not le'ng, to be consistent w.r.t. pa'am
* vs. pa'm: */
if (lastWylie.equals(ACHUNG))
sb.append(WYLIE_aVOWEL);
sb.append(currWylie); sb.append(currWylie);
lastWylie = currWylie; lastWylie = currWylie;

4
source/org/thdl/tib/text/TibetanMachineWeb.java
View file

@ -628,6 +628,10 @@ public static boolean isVowel(String s) {
* transliteration, false if not * transliteration, false if not
*/ */
public static boolean isAmbiguousWylie(String x, String y) { public static boolean isAmbiguousWylie(String x, String y) {
// What about ambiguity between wa-zur and wa? dwa vs. d.wa, e.g.?
// Doesn't matter, because that's illegal. wa doesn't take any
// prefixes.
return (("g".equals(x) && "y".equals(y)) return (("g".equals(x) && "y".equals(y))
|| ("b".equals(x) && "l".equals(y)) || ("b".equals(x) && "l".equals(y))
|| ("b".equals(x) && "r".equals(y))); || ("b".equals(x) && "r".equals(y)));