thdltools/Jskad
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity
Jskad/source/org/thdl/tib/text/ttt/EWTSTraits.java

518 lines
22 KiB
Java
Raw Normal View History

I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
/*
The contents of this file are subject to the THDL Open Community License
Version 1.0 (the "License"); you may not use this file except in compliance
with the License. You may obtain a copy of the License on the THDL web site
(http://www.thdl.org/).
Software distributed under the License is distributed on an "AS IS" basis,
WITHOUT WARRANTY OF ANY KIND, either express or implied. See the
License for the specific terms governing rights and limitations under the
License.
The Initial Developer of this software is the Tibetan and Himalayan Digital
Library (THDL). Portions created by the THDL are Copyright 2004 THDL.
All Rights Reserved.
Contributor(s): ______________________________________.
*/
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
// TODO(DLC)[EWTS->Tibetan]: TibetanMachineWeb has duplication of much of this!
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
package org.thdl.tib.text.ttt;
Two things: One, TMW->EWTS gives dbas and dngas instead of dabs and dangs because Chris Fynn's e-mail from today has dbas and dngas. Second, Down with ACIPRules. Long live ACIPTraits. EWTS->Tibetan conversion is closer still.
2005-02-22 04:36:54 +00:00
import java.util.ArrayList;
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
Tremendously better EWTS->Unicode and EWTS->TMW conversion, though still not tested end-to-end and without perfect unit tests. See EWTSTest.RUN_FAILING_TESTS, for example, to find imperfection.
2005-07-06 02:19:38 +00:00
import org.thdl.tib.text.tshegbar.UnicodeUtils;
Two things: One, TMW->EWTS gives dbas and dngas instead of dabs and dangs because Chris Fynn's e-mail from today has dbas and dngas. Second, Down with ACIPRules. Long live ACIPTraits. EWTS->Tibetan conversion is closer still.
2005-02-22 04:36:54 +00:00
import org.thdl.tib.text.DuffCode;
Added UI for EWTS->Tibetan conversions. GUI is disabled except in debug mode for now. I tested against a really simple-but-real document, found a bug with '*', tried to implement TMW vowel code but I don't trust it yet. Differentiated EWTS code from ACIP where needed. Several bugs in ewts->tibetan have been exposed; see the TODO comments.
2005-06-20 09:30:35 +00:00
import org.thdl.tib.text.THDLWylieConstants;
import org.thdl.tib.text.TibTextUtils;
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
import org.thdl.tib.text.TibetanMachineWeb;
import org.thdl.util.ThdlDebug;
Two things: One, TMW->EWTS gives dbas and dngas instead of dabs and dangs because Chris Fynn's e-mail from today has dbas and dngas. Second, Down with ACIPRules. Long live ACIPTraits. EWTS->Tibetan conversion is closer still.
2005-02-22 04:36:54 +00:00
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
/** A singleton class that should contain (but due to laziness and
* ignorance probably does not contain) all the traits that make EWTS
* transliteration different from other (say, ACIP) transliteration
* schemes. */
Two things: One, TMW->EWTS gives dbas and dngas instead of dabs and dangs because Chris Fynn's e-mail from today has dbas and dngas. Second, Down with ACIPRules. Long live ACIPTraits. EWTS->Tibetan conversion is closer still.
2005-02-22 04:36:54 +00:00
public final class EWTSTraits implements TTraits {
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
/** sole instance of this class */
private static EWTSTraits singleton = null;
/** Just a constructor. */
private EWTSTraits() { }
/** */
Two things: One, TMW->EWTS gives dbas and dngas instead of dabs and dangs because Chris Fynn's e-mail from today has dbas and dngas. Second, Down with ACIPRules. Long live ACIPTraits. EWTS->Tibetan conversion is closer still.
2005-02-22 04:36:54 +00:00
public static synchronized EWTSTraits instance() {
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
if (null == singleton) {
singleton = new EWTSTraits();
}
return singleton;
}
/** Returns ".". */
public String disambiguator() { return "."; }
/** Returns '.'. */
public char disambiguatorChar() { return '.'; }
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
// TODO(DLC)[EWTS->Tibetan]: isClearlyIllegal and hasSimpleError are different why?
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
public boolean hasSimpleError(TPair p) {
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
if (pairHasBadWowel(p)) return true;
return (("a".equals(p.getLeft()) && null == p.getRight())
|| ("a".equals(p.getLeft())
&& null != p.getRight()
&& TibetanMachineWeb.isWylieVowel(p.getRight()))); // TODO(DLC)[EWTS->Tibetan]: or Unicode wowels? test "a\u0f74" and "a\u0f7e"
// TODO(DLC)[EWTS->Tibetan]: (a.e) is bad, one of (.a) or (a.) is bad
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
}
/** {tsh}, the longest consonant, has 3 characters, so this is
* three. */
public int maxConsonantLength() { return 3; }
Better, but still flawed, handling of EWTS [^] (i.e., U+0F39).
2005-07-06 22:26:55 +00:00
/** Wowels can be arbitrarily long via stacking. But each
* component is no longer, in characters, than this. [~M`] is
* the current winner. */
public int maxWowelLength() { return 3; }
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
public boolean isUnicodeConsonant(char ch) {
Tremendously better EWTS->Unicode and EWTS->TMW conversion, though still not tested end-to-end and without perfect unit tests. See EWTSTest.RUN_FAILING_TESTS, for example, to find imperfection.
2005-07-06 02:19:38 +00:00
return ((ch != '\u0f48' && ch >= '\u0f40' && ch <= '\u0f6a')
|| (ch != '\u0f98' && ch >= '\u0f90' && ch <= '\u0fbc')
// NOTE: \u0f88 is questionable, but we want EWTS
// [\u0f88+kha] to become "\u0f88\u0f91" and this does
// the trick.
|| ch == '\u0f88');
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
}
public boolean isUnicodeWowel(char ch) {
// TODO(DLC)[EWTS->Tibetan]: what about combiners that combine only with digits? TEST
return ((ch >= '\u0f71' && ch <= '\u0f84')
Better, but still flawed, handling of EWTS [^] (i.e., U+0F39).
2005-07-06 22:26:55 +00:00
|| '\u0f39' == ch
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
|| isUnicodeWowelThatRequiresAChen(ch));
}
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
// TODO(DLC)[EWTS->Tibetan]: u,e,i,o? If not, document the special treatment in this function's comment
public boolean isConsonant(String s) {
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
if (s.length() == 1 && isUnicodeConsonant(s.charAt(0))) return true;
if (aVowel().equals(s)) return false; // In EWTS, "a" is both a consonant and a vowel, but we treat it as just a vowel and insert the implied a-chen if you have a TPair ( . a) (TODO(DLC)[EWTS->Tibetan]: right?)
// TODO(DLC)[EWTS->Tibetan]: numbers are consonants?
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
// TODO(DLC)[EWTS->Tibetan]: just g for now
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
return TibetanMachineWeb.isWylieChar(s);
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
}
public boolean isWowel(String s) {
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
return (getUnicodeForWowel(s) != null);
/* TODO(DLC)[EWTS->Tibetan]: test ko+m+e etc.
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
// TODO(DLC)[EWTS->Tibetan]: all non-consonant combiners? 0f71 0f87 etc.?
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
if (s.length() == 1 && isUnicodeWowel(s.charAt(0))) return true;
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
return ("a".equals(s)
|| "e".equals(s)
|| "i".equals(s)
|| "o".equals(s)
|| "u".equals(s)
|| "U".equals(s)
|| "I".equals(s)
|| "A".equals(s)
|| "-i".equals(s)
|| "-I".equals(s)
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
|| "au".equals(s)
|| "ai".equals(s)
|| isWowelThatRequiresAChen(s));
// TODO(DLC)[EWTS->Tibetan]:???
*/
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
}
Two things: One, TMW->EWTS gives dbas and dngas instead of dabs and dangs because Chris Fynn's e-mail from today has dbas and dngas. Second, Down with ACIPRules. Long live ACIPTraits. EWTS->Tibetan conversion is closer still.
2005-02-22 04:36:54 +00:00
public String aVowel() { return "a"; }
public boolean isPostsuffix(String s) {
return ("s".equals(s) || "d".equals(s));
}
public boolean isPrefix(String l) {
return ("'".equals(l)
|| "m".equals(l)
|| "b".equals(l)
|| "d".equals(l)
|| "g".equals(l));
}
public boolean isSuffix(String l) {
return ("s".equals(l)
|| "g".equals(l)
|| "d".equals(l)
|| "m".equals(l)
|| "'".equals(l)
|| "b".equals(l)
|| "ng".equals(l)
|| "n".equals(l)
|| "l".equals(l)
|| "r".equals(l));
}
/** Returns l, since this is EWTS's traits class. */
public String getEwtsForConsonant(String l) { return l; }
/** Returns l, since this is EWTS's traits class. */
public String getEwtsForOther(String l) { return l; }
/** Returns l, since this is EWTS's traits class. */
public String getEwtsForWowel(String l) { return l; }
public TTshegBarScanner scanner() { return EWTSTshegBarScanner.instance(); }
Bulletproofed EWTS->Tibetan against nasty pseudo-EWTS like [RAM]. Renamed recoverACIP methods.
2005-07-07 02:54:36 +00:00
public void getDuffForWowel(ArrayList duff, DuffCode preceding, String wowel)
throws IllegalArgumentException
{
Added UI for EWTS->Tibetan conversions. GUI is disabled except in debug mode for now. I tested against a really simple-but-real document, found a bug with '*', tried to implement TMW vowel code but I don't trust it yet. Differentiated EWTS code from ACIP where needed. Several bugs in ewts->tibetan have been exposed; see the TODO comments.
2005-06-20 09:30:35 +00:00
// TODO(DLC)[EWTS->Tibetan]: I have no confidence in this! test, test, test.
EWTS->TMW fixes. Wowel handling still isn't perfect but I'm lazy. Jskad now uses the new EWTS->TMW routine, not the old, and thus the "(Buggy)" label is [unfairly, perhaps] dropped.
2005-07-07 01:30:03 +00:00
// TODO(DLC)[EWTS->Tibetan]: ko+o doesn't work. kai+-i doesn't work.
// TODO(DLC)[EWTS->Tibetan]: kai doesn't work.
Added UI for EWTS->Tibetan conversions. GUI is disabled except in debug mode for now. I tested against a really simple-but-real document, found a bug with '*', tried to implement TMW vowel code but I don't trust it yet. Differentiated EWTS code from ACIP where needed. Several bugs in ewts->tibetan have been exposed; see the TODO comments.
2005-06-20 09:30:35 +00:00
// Order matters here.
boolean context_added[] = new boolean[] { false };
if (wowel.equals(THDLWylieConstants.WYLIE_aVOWEL)) {
TibTextUtils.getVowel(duff, preceding, THDLWylieConstants.WYLIE_aVOWEL, context_added);
} else {
// TODO(DLC)[EWTS->Tibetan]: test vowel stacking
if (wowel.indexOf(THDLWylieConstants.U_VOWEL) >= 0) {
TibTextUtils.getVowel(duff, preceding, THDLWylieConstants.U_VOWEL, context_added);
}
if (wowel.indexOf(THDLWylieConstants.reverse_I_VOWEL) >= 0) {
TibTextUtils.getVowel(duff, preceding, THDLWylieConstants.reverse_I_VOWEL, context_added);
} else if (wowel.indexOf(THDLWylieConstants.I_VOWEL) >= 0) {
TibTextUtils.getVowel(duff, preceding, THDLWylieConstants.I_VOWEL, context_added);
}
if (wowel.indexOf(THDLWylieConstants.A_VOWEL) >= 0) {
TibTextUtils.getVowel(duff, preceding, THDLWylieConstants.A_VOWEL, context_added);
}
if (wowel.indexOf(THDLWylieConstants.ai_VOWEL) >= 0) {
TibTextUtils.getVowel(duff, preceding, THDLWylieConstants.ai_VOWEL, context_added);
EWTS->TMW fixes. Wowel handling still isn't perfect but I'm lazy. Jskad now uses the new EWTS->TMW routine, not the old, and thus the "(Buggy)" label is [unfairly, perhaps] dropped.
2005-07-07 01:30:03 +00:00
} else if (wowel.indexOf(THDLWylieConstants.reverse_i_VOWEL) >= 0) {
Added UI for EWTS->Tibetan conversions. GUI is disabled except in debug mode for now. I tested against a really simple-but-real document, found a bug with '*', tried to implement TMW vowel code but I don't trust it yet. Differentiated EWTS code from ACIP where needed. Several bugs in ewts->tibetan have been exposed; see the TODO comments.
2005-06-20 09:30:35 +00:00
TibTextUtils.getVowel(duff, preceding, THDLWylieConstants.reverse_i_VOWEL, context_added);
} else if (wowel.indexOf(THDLWylieConstants.i_VOWEL) >= 0) {
TibTextUtils.getVowel(duff, preceding, THDLWylieConstants.i_VOWEL, context_added);
}
if (wowel.indexOf(THDLWylieConstants.e_VOWEL) >= 0) {
TibTextUtils.getVowel(duff, preceding, THDLWylieConstants.e_VOWEL, context_added);
}
if (wowel.indexOf(THDLWylieConstants.o_VOWEL) >= 0) {
TibTextUtils.getVowel(duff, preceding, THDLWylieConstants.o_VOWEL, context_added);
}
EWTS->TMW fixes. Wowel handling still isn't perfect but I'm lazy. Jskad now uses the new EWTS->TMW routine, not the old, and thus the "(Buggy)" label is [unfairly, perhaps] dropped.
2005-07-07 01:30:03 +00:00
if (wowel.indexOf(THDLWylieConstants.au_VOWEL) >= 0) {
TibTextUtils.getVowel(duff, preceding, THDLWylieConstants.au_VOWEL, context_added);
} else if (wowel.indexOf(THDLWylieConstants.u_VOWEL) >= 0) {
Added UI for EWTS->Tibetan conversions. GUI is disabled except in debug mode for now. I tested against a really simple-but-real document, found a bug with '*', tried to implement TMW vowel code but I don't trust it yet. Differentiated EWTS code from ACIP where needed. Several bugs in ewts->tibetan have been exposed; see the TODO comments.
2005-06-20 09:30:35 +00:00
TibTextUtils.getVowel(duff, preceding, THDLWylieConstants.u_VOWEL, context_added);
}
if (wowel.indexOf("~X") >= 0) { // TODO(DLC)[EWTS->Tibetan]: introduce THDLWylieConstants.blah
duff.add(TibetanMachineWeb.getGlyph("~X"));
} else if (wowel.indexOf("X") >= 0) { // TODO(DLC)[EWTS->Tibetan]: introduce THDLWylieConstants.blah
duff.add(TibetanMachineWeb.getGlyph("X"));
}
}
// FIXME: Use TMW9.61, the "o'i" special combination, when appropriate.
EWTS->TMW fixes. Wowel handling still isn't perfect but I'm lazy. Jskad now uses the new EWTS->TMW routine, not the old, and thus the "(Buggy)" label is [unfairly, perhaps] dropped.
2005-07-07 01:30:03 +00:00
if (wowel.indexOf(THDLWylieConstants.BINDU) >= 0
// TODO(DLC)[EWTS->Tibetan]: This is really ugly... we
// rely on the fact that we know every Wylie wowel that
// contains 'M'. Let's, instead, parse the wowel.
&& wowel.indexOf(THDLWylieConstants.U0F82) < 0
&& wowel.indexOf(THDLWylieConstants.U0F83) < 0) {
Added UI for EWTS->Tibetan conversions. GUI is disabled except in debug mode for now. I tested against a really simple-but-real document, found a bug with '*', tried to implement TMW vowel code but I don't trust it yet. Differentiated EWTS code from ACIP where needed. Several bugs in ewts->tibetan have been exposed; see the TODO comments.
2005-06-20 09:30:35 +00:00
DuffCode last = null;
Fixed important EWTS->Tibetan conversion bugs. [g.yogs] (and maybe [hUM^]) are not yet converting correctly. I have not yet committed the end-to-end test that I'm manually doing to find these problems. It will be another document for TMW_RTF_TO_THDL_WYLIETest.java. Note that thdl.debug=true is essential to access the GUI for the EWTS->* converters.
2005-07-06 07:46:21 +00:00
if (!context_added[0]) {
last = preceding;
} else if (duff.size() > 0) {
Added UI for EWTS->Tibetan conversions. GUI is disabled except in debug mode for now. I tested against a really simple-but-real document, found a bug with '*', tried to implement TMW vowel code but I don't trust it yet. Differentiated EWTS code from ACIP where needed. Several bugs in ewts->tibetan have been exposed; see the TODO comments.
2005-06-20 09:30:35 +00:00
last = (DuffCode)duff.get(duff.size() - 1);
duff.remove(duff.size() - 1); // getBindu will add it back...
// TODO(DLC)[EWTS->Tibetan]: is this okay???? when is a bindu okay to be alone???
}
TibTextUtils.getBindu(duff, last);
EWTS->TMW fixes. Wowel handling still isn't perfect but I'm lazy. Jskad now uses the new EWTS->TMW routine, not the old, and thus the "(Buggy)" label is [unfairly, perhaps] dropped.
2005-07-07 01:30:03 +00:00
context_added[0] = true;
}
if (!context_added[0]) {
duff.add(preceding);
Added UI for EWTS->Tibetan conversions. GUI is disabled except in debug mode for now. I tested against a really simple-but-real document, found a bug with '*', tried to implement TMW vowel code but I don't trust it yet. Differentiated EWTS code from ACIP where needed. Several bugs in ewts->tibetan have been exposed; see the TODO comments.
2005-06-20 09:30:35 +00:00
}
if (wowel.indexOf('H') >= 0)
duff.add(TibetanMachineWeb.getGlyph("H"));
EWTS->TMW fixes. Wowel handling still isn't perfect but I'm lazy. Jskad now uses the new EWTS->TMW routine, not the old, and thus the "(Buggy)" label is [unfairly, perhaps] dropped.
2005-07-07 01:30:03 +00:00
int ix;
if ((ix = wowel.indexOf(THDLWylieConstants.WYLIE_TSA_PHRU)) >= 0) {
// This likely won't look good! TMW has glyphs for [va]
// and [fa], so use that transliteration if you care, not
// [ph^] or [b^].
duff.add(TibetanMachineWeb.getGlyph(THDLWylieConstants.WYLIE_TSA_PHRU));
StringBuffer sb = new StringBuffer(wowel);
sb.replace(ix, ix + THDLWylieConstants.WYLIE_TSA_PHRU.length(), "");
wowel = sb.toString();
}
if ((ix = wowel.indexOf(THDLWylieConstants.U0F82)) >= 0) {
duff.add(TibetanMachineWeb.getGlyph(THDLWylieConstants.U0F82));
StringBuffer sb = new StringBuffer(wowel);
sb.replace(ix, ix + THDLWylieConstants.U0F82.length(), "");
wowel = sb.toString();
}
if ((ix = wowel.indexOf(THDLWylieConstants.U0F83)) >= 0) {
duff.add(TibetanMachineWeb.getGlyph(THDLWylieConstants.U0F83));
StringBuffer sb = new StringBuffer(wowel);
sb.replace(ix, ix + THDLWylieConstants.U0F83.length(), "");
wowel = sb.toString();
}
Added UI for EWTS->Tibetan conversions. GUI is disabled except in debug mode for now. I tested against a really simple-but-real document, found a bug with '*', tried to implement TMW vowel code but I don't trust it yet. Differentiated EWTS code from ACIP where needed. Several bugs in ewts->tibetan have been exposed; see the TODO comments.
2005-06-20 09:30:35 +00:00
// TODO(DLC)[EWTS->Tibetan]: verify that no part of wowel is discarded! acip does that. 'jam~X I think we screw up, e.g.
// TODO(DLC)[EWTS->Tibetan]:: are bindus are screwed up in the unicode output? i see (with tmuni font) lone bindus without glyphs to stack on
Two things: One, TMW->EWTS gives dbas and dngas instead of dabs and dangs because Chris Fynn's e-mail from today has dbas and dngas. Second, Down with ACIPRules. Long live ACIPTraits. EWTS->Tibetan conversion is closer still.
2005-02-22 04:36:54 +00:00
}
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
public String getUnicodeForWowel(String wowel) {
if ("a".equals(wowel))
return "";
return helpGetUnicodeForWowel(wowel);
}
private String helpGetUnicodeForWowel(String wowel) {
if ("a".equals(wowel))
return null; // ko+a+e is invalid, e.g.
if (wowel.length() == 1 && isUnicodeWowel(wowel.charAt(0)))
return wowel;
// handle o+u, etc.
int i;
if ((i = wowel.indexOf("+")) >= 0) {
// recurse.
// Chris Fynn says \u0f7c\u0f7c is different from \u0f7d.
// So o+o is not the same as au. e+e is not the same as
// ai.
String left = helpGetUnicodeForWowel(wowel.substring(0, i));
String right = helpGetUnicodeForWowel(wowel.substring(i + 1));
if (null != left && null != right)
return left + right;
else
return null;
} else {
// Handle vowels. (TODO(dchandler): tibwn.ini has this
// info, use that instead of duplicating it in this code.)
if ("i".equals(wowel)) return "\u0f72";
if ("u".equals(wowel)) return "\u0f74";
if ("A".equals(wowel)) return "\u0f71";
if ("U".equals(wowel)) return "\u0f71\u0f74"; // \u0f75 is discouraged
if ("e".equals(wowel)) return "\u0f7a";
if ("o".equals(wowel)) return "\u0f7c";
if ("-i".equals(wowel)) return "\u0f80";
if ("ai".equals(wowel)) return "\u0f7b";
if ("au".equals(wowel)) return "\u0f7d";
if ("-I".equals(wowel)) return "\u0f81";
if ("I".equals(wowel)) return "\u0f71\u0f72"; // \u0f73 is discouraged
Better, but still flawed, handling of EWTS [^] (i.e., U+0F39).
2005-07-06 22:26:55 +00:00
// TODO(DLC)[EWTS->Tibetan]: test, test, test.
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
if ("M".equals(wowel)) return "\u0f7e";
if ("H".equals(wowel)) return "\u0f7f";
if ("?".equals(wowel)) return "\u0f84";
if ("~M".equals(wowel)) return "\u0f83";
if ("~M`".equals(wowel)) return "\u0f82";
if ("X".equals(wowel)) return "\u0f37";
if ("~X".equals(wowel)) return "\u0f35";
Better, but still flawed, handling of EWTS [^] (i.e., U+0F39).
2005-07-06 22:26:55 +00:00
if ("^".equals(wowel)) return "\u0f39";
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
return null;
}
}
public String getUnicodeFor(String l, boolean subscribed) {
Added UI for EWTS->Tibetan conversions. GUI is disabled except in debug mode for now. I tested against a really simple-but-real document, found a bug with '*', tried to implement TMW vowel code but I don't trust it yet. Differentiated EWTS code from ACIP where needed. Several bugs in ewts->tibetan have been exposed; see the TODO comments.
2005-06-20 09:30:35 +00:00
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
// First, handle "\u0f71\u0f84\u0f86", "", "\u0f74", etc.
{
boolean already_done = true;
for (int i = 0; i < l.length(); i++) {
Added UI for EWTS->Tibetan conversions. GUI is disabled except in debug mode for now. I tested against a really simple-but-real document, found a bug with '*', tried to implement TMW vowel code but I don't trust it yet. Differentiated EWTS code from ACIP where needed. Several bugs in ewts->tibetan have been exposed; see the TODO comments.
2005-06-20 09:30:35 +00:00
char ch = l.charAt(i);
if ((ch < '\u0f00' || ch > '\u0fff')
Tremendously better EWTS->Unicode and EWTS->TMW conversion, though still not tested end-to-end and without perfect unit tests. See EWTSTest.RUN_FAILING_TESTS, for example, to find imperfection.
2005-07-06 02:19:38 +00:00
&& SAUVASTIKA != ch
&& SWASTIKA != ch
&& (ch < PUA_MIN || ch > PUA_MAX) // TODO(DLC)[EWTS->Tibetan]: give a warning, though? PUA isn't specified by the unicode standard after all.
Added UI for EWTS->Tibetan conversions. GUI is disabled except in debug mode for now. I tested against a really simple-but-real document, found a bug with '*', tried to implement TMW vowel code but I don't trust it yet. Differentiated EWTS code from ACIP where needed. Several bugs in ewts->tibetan have been exposed; see the TODO comments.
2005-06-20 09:30:35 +00:00
&& '\n' != ch
&& '\r' != ch) {
// TODO(DLC)[EWTS->Tibetan]: Is this the place
// where we want to interpret how newlines work???
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
already_done = false;
break;
}
}
if (already_done)
return l; // TODO(dchandler): \u0fff etc. are not valid code points, though. Do we handle that well?
}
// TODO(DLC)[EWTS->Tibetan]:: vowels !subscribed could mean (a . i)???? I doubt it but test "i"->"\u0f68\u0f72" etc.
if (subscribed) {
if ("R".equals(l)) return "\u0fbc";
if ("Y".equals(l)) return "\u0fbb";
if ("W".equals(l)) return "\u0fba";
// g+h etc. should not be inputs to this function, but for
// completeness they're here.
if ("k".equals(l)) return "\u0F90";
if ("kh".equals(l)) return "\u0F91";
if ("g".equals(l)) return "\u0F92";
if ("g+h".equals(l)) return "\u0F93";
if ("ng".equals(l)) return "\u0F94";
if ("c".equals(l)) return "\u0F95";
if ("ch".equals(l)) return "\u0F96";
if ("j".equals(l)) return "\u0F97";
if ("ny".equals(l)) return "\u0F99";
if ("T".equals(l)) return "\u0F9A";
if ("Th".equals(l)) return "\u0F9B";
if ("D".equals(l)) return "\u0F9C";
if ("D+h".equals(l)) return "\u0F9D";
if ("N".equals(l)) return "\u0F9E";
if ("t".equals(l)) return "\u0F9F";
if ("th".equals(l)) return "\u0FA0";
if ("d".equals(l)) return "\u0FA1";
if ("d+h".equals(l)) return "\u0FA2";
if ("n".equals(l)) return "\u0FA3";
if ("p".equals(l)) return "\u0FA4";
if ("ph".equals(l)) return "\u0FA5";
if ("b".equals(l)) return "\u0FA6";
if ("b+h".equals(l)) return "\u0FA7";
if ("m".equals(l)) return "\u0FA8";
if ("ts".equals(l)) return "\u0FA9";
if ("tsh".equals(l)) return "\u0FAA";
if ("dz".equals(l)) return "\u0FAB";
if ("dz+h".equals(l)) return "\u0FAC";
if ("w".equals(l)) return "\u0FAD"; // TODO(DLC)[EWTS->Tibetan]:: ???
if ("zh".equals(l)) return "\u0FAE";
if ("z".equals(l)) return "\u0FAF";
if ("'".equals(l)) return "\u0FB0";
if ("y".equals(l)) return "\u0FB1";
if ("r".equals(l)) return "\u0FB2";
if ("l".equals(l)) return "\u0FB3";
if ("sh".equals(l)) return "\u0FB4";
if ("Sh".equals(l)) return "\u0FB5";
if ("s".equals(l)) return "\u0FB6";
if ("h".equals(l)) return "\u0FB7";
if ("a".equals(l)) return "\u0FB8";
if ("k+Sh".equals(l)) return "\u0FB9";
Better, but still flawed, handling of EWTS [^] (i.e., U+0F39).
2005-07-06 22:26:55 +00:00
if ("f".equals(l)) return "\u0FA5\u0F39";
if ("v".equals(l)) return "\u0FA6\u0F39";
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
return null;
} else {
if ("R".equals(l)) return "\u0f6a";
if ("Y".equals(l)) return "\u0f61";
if ("W".equals(l)) return "\u0f5d";
EWTS->Unicode for // now produces \u0f0e as it should.
2005-07-10 05:01:03 +00:00
if ("//".equals(l)) return "\u0f0e";
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
if (!TibetanMachineWeb.isKnownHashKey(l)) {
Tremendously better EWTS->Unicode and EWTS->TMW conversion, though still not tested end-to-end and without perfect unit tests. See EWTSTest.RUN_FAILING_TESTS, for example, to find imperfection.
2005-07-06 02:19:38 +00:00
// System.err.println("Getting unicode for the following is hard: '"
// + l + "' (pretty string: '"
// + UnicodeUtils.unicodeStringToPrettyString(l)
// + "'");
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
ThdlDebug.noteIffyCode();
return null;
}
String s = TibetanMachineWeb.getUnicodeForWylieForGlyph(l);
if (null == s)
ThdlDebug.noteIffyCode();
return s;
}
}
public String shortTranslitName() { return "EWTS"; }
private boolean pairHasBadWowel(TPair p) {
return (null != p.getRight()
&& !disambiguator().equals(p.getRight())
&& !"+".equals(p.getRight())
&& null == getUnicodeForWowel(p.getRight()));
}
public boolean isClearlyIllegal(TPair p) {
if (pairHasBadWowel(p)) return true;
if (p.getLeft() == null
&& (p.getRight() == null ||
(!disambiguator().equals(p.getRight())
&& !isWowel(p.getRight()))))
return true;
if ("+".equals(p.getLeft()))
return true;
if (p.getLeft() != null && isWowel(p.getLeft())
&& !aVowel().equals(p.getLeft())) // achen
return true;
return false;
}
public TPairList[] breakTshegBarIntoChunks(String tt, boolean sh) {
if (sh) throw new IllegalArgumentException("Don't do that, silly!");
try {
return TPairListFactory.breakEWTSIntoChunks(tt);
} catch (StackOverflowError e) {
throw new IllegalArgumentException("Input too large[1]: " + tt);
} catch (OutOfMemoryError e) {
throw new IllegalArgumentException("Input too large[2]: " + tt);
}
}
public boolean isACIP() { return false; }
public boolean vowelAloneImpliesAChen() { return true; }
public boolean vowelsMayStack() { return true; }
public boolean isWowelThatRequiresAChen(String s) {
// TODO(DLC)[EWTS->Tibetan]: fix me!
return ((s.length() == 1 && (isUnicodeWowelThatRequiresAChen(s.charAt(0))
Better, but still flawed, handling of EWTS [^] (i.e., U+0F39).
2005-07-06 22:26:55 +00:00
|| "?MHX^".indexOf(s.charAt(0)) >= 0))
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
|| "~X".equals(s)
|| "~M".equals(s)
|| "~M`".equals(s)
);
}
public boolean isUnicodeWowelThatRequiresAChen(char ch) {
// TODO(DLC)[EWTS->Tibetan]: ask if 18 19 3e 3f combine only with digits
My last commit left the tests broken. Doh. Also, I'm enabling EWTS->Tibetan converters in the GUI so that I can ask folks to try them out.
2005-07-06 22:55:19 +00:00
return "\u0f39\u0f35\u0f37\u0f18\u0f19\u0f3e\u0f3f\u0f86\u0f87\u0fc6".indexOf(ch) >= 0;
I really hesitate to commit this because I'm not sure what it brings to the table exactly and I fear that it makes the ACIP->Tibetan converter code a lot uglier. The TODO(DLC)[EWTS->Tibetan] comments littered throughout are part of the ugliness; they point to the ugliness. If each were addressed, cleanliness could perhaps be achieved. I've largely forgotten exactly what this change does, but it attempts to improve EWTS->Tibetan conversion. The lexer is probably really, really primitive. I concentrate here on converting a single tsheg bar rather than a whole document. Eclipse was used during part of my journey here and some imports were reorganized merely because I could. :) (Eclipse was needed when the usual ant build failed to run a new test EWTSTest. And I wanted its debugger.) Next steps: end-to-end EWTS tests should bring many problems to light. Fix those. Triage all the TODO comments. I don't know that I'll ever really trust the implementation. The tests are valuable, though. A clean implementation of EWTS->Tibetan in Jython might hold enough interest for me; I'd like to learn Python.
2005-06-20 06:18:00 +00:00
}
public boolean couldBeValidStack(TPairList pl) {
StringBuffer hashKey = new StringBuffer();
boolean allHavePlus = true;
for (int i = 0; i < pl.size(); i++) {
if (i + 1 < pl.size() && !"+".equals(pl.get(i).getRight()))
allHavePlus = false;
if (0 != hashKey.length())
hashKey.append('-');
hashKey.append(pl.get(i).getLeft());
}
return (allHavePlus
|| TibetanMachineWeb.hasGlyph(hashKey.toString())); // TODO(DLC)[EWTS->Tibetan]: test with smra and tsma and bdgya
}
Tremendously better EWTS->Unicode and EWTS->TMW conversion, though still not tested end-to-end and without perfect unit tests. See EWTSTest.RUN_FAILING_TESTS, for example, to find imperfection.
2005-07-06 02:19:38 +00:00
public boolean stackingMustBeExplicit() { return true; }
public String U0F7F() { return "H"; }
public String U0F35() { return "~X"; }
public String U0F37() { return "X"; }
/** The EWTS standard mentions this character specifically. See
http://www.symbols.com/encyclopedia/15/155.html to learn about
its meaning as relates to Buddhism.
*/
static final char SAUVASTIKA = '\u534d';
/** The EWTS standard mentions this character specifically. See
http://www.symbols.com/encyclopedia/15/151.html to learn about
its meaning as relates to Buddhism.
*/
static final char SWASTIKA = '\u5350';
/** EWTS has some glyphs not specified by Unicode in the
* private-use area (PUA). EWTS puts them in the range [PUA_MIN,
* PUA_MAX]. (Note that \uf042 is the highest in use as of July
* 2, 2005.) */
static final char PUA_MIN = '\uf021';
/** EWTS has some glyphs not specified by Unicode in the
* private-use area (PUA). EWTS puts them in the range [PUA_MIN,
* PUA_MAX]. (Note that \uf042 is the highest in use as of July
* 2, 2005.) */
static final char PUA_MAX = '\uf0ff';
I did this stuff back in August. It's all in support of EWTS->Tibetan conversion. The tag 'TODO(DLC)[EWTS->Tibetan]' exists all over the place. EWTS->Tibetan isn't here yet; lexing isn't here yet; this is mainly a refactoring so that the ACIP->Tibetan code can be reused to do EWTS->Tibetan. I'm committing this because tests pass (it shouldn't be breaking anything), because I want a checkpoint, and because the laptop this sandbox was on isn't my preferred development environment.
2005-02-21 01:16:10 +00:00
}
Reference in a new issue Copy permalink