0F880F90
+ 0F880F91
+ \u0F55\u0F39
+ \u0F56\u0F39
+
+TODO(DLC)[EWTS->Tibetan]:: also, There is one case where we break things up into two pair * lists if and only if specialHandlingForAppendages is true -- I @@ -65,19 +67,20 @@ class TPairListFactory { * @return an array of one or two pair lists, if the former, then * the second element will be null, if the latter, the second * element will have (* . ), (' . *) instead of (* . '*) which - * the former has @throws IllegalArgumentException if acip is too - * large for us to break into chunks (we're recursive, not - * iterative, so the boundary can be increased a lot if you care, - * but you don't) */ + * the former has + * @throws IllegalArgumentException if acip is too large for us + * to break into chunks (we're recursive, not iterative, so the + * boundary can be increased a lot if you care, but you don't) */ static TPairList[] breakACIPIntoChunks(String acip, boolean specialHandlingForAppendages) throws IllegalArgumentException { try { - TPairList a = breakHelper(acip, true, false); + TTraits ttraits = ACIPTraits.instance(); + TPairList a = breakHelperACIP(acip, true, false, ttraits); TPairList b = null; if (specialHandlingForAppendages) - b = breakHelper(acip, false, false); + b = breakHelperACIP(acip, false, false, ttraits); if (null != b && a.equals(b)) return new TPairList[] { a, null }; else @@ -88,6 +91,22 @@ class TPairListFactory { throw new IllegalArgumentException("Input too large[2]: " + acip); } } + + /** TODO(DLC)[EWTS->Tibetan]: doc */ + static TPairList[] breakEWTSIntoChunks(String ewts) + throws IllegalArgumentException + { + try { + return new TPairList[] { + breakHelperEWTS(ewts, EWTSTraits.instance()), null + }; + } catch (StackOverflowError e) { + throw new IllegalArgumentException("Input too large[1]: " + ewts); + } catch (OutOfMemoryError e) { + throw new IllegalArgumentException("Input too large[2]: " + ewts); + } + } + /** Helps {@link #breakACIPIntoChunks(String,boolean)}. * @param tickIsVowel true if and only if you want to treat the * ACIP {'} as an U+0F71 vowel instead of the full-sized @@ -96,7 +115,9 @@ class TPairListFactory { * @param weHaveSeenVowelAlready true if and only if, in our * recursion, we've already found one vowel (not a disambiguator, * but a vowel like "A", "E", "Um:", "m", "'U", etc.) */ - private static TPairList breakHelper(String acip, boolean tickIsVowel, boolean weHaveSeenVowelAlready) { + private static TPairList breakHelperACIP(String acip, boolean tickIsVowel, + boolean weHaveSeenVowelAlready, + TTraits ttraits) { // base case for our recursion: if ("".equals(acip)) @@ -104,7 +125,7 @@ class TPairListFactory { StringBuffer acipBuf = new StringBuffer(acip); int howMuchBuf[] = new int[1]; - TPair head = getFirstConsonantAndVowel(acipBuf, howMuchBuf); + TPair head = getFirstConsonantAndVowel(acipBuf, howMuchBuf, ttraits); int howMuch = howMuchBuf[0]; if (!tickIsVowel && null != head.getLeft() @@ -122,26 +143,63 @@ class TPairListFactory { TPairList tail; if ((tail - = breakHelper(acipBuf.substring(howMuch), - tickIsVowel, - weHaveSeenVowelAlready - || (head.getRight() != null - && !"+".equals(head.getRight()) - && !"-".equals(head.getRight())))).hasSimpleError()) { + = breakHelperACIP(acipBuf.substring(howMuch), + tickIsVowel, + weHaveSeenVowelAlready + || (head.getRight() != null + && !"+".equals(head.getRight()) + && !"-".equals(head.getRight())), + ttraits)).hasSimpleError(ttraits)) { for (int i = 1; i < howMuch; i++) { // try giving i characters back if that leaves us with // a legal head and makes the rest free of simple // errors. TPairList newTail = null; TPair newHead; - if ((newHead = head.minusNRightmostACIPCharacters(i)).isLegal() + if ((newHead = head.minusNRightmostTransliterationCharacters(i)).isLegal() && !(newTail - = breakHelper(acipBuf.substring(howMuch - i), - tickIsVowel, - weHaveSeenVowelAlready - || (newHead.getRight() != null - && !"+".equals(newHead.getRight()) - && !"-".equals(newHead.getRight())))).hasSimpleError()) { + = breakHelperACIP(acipBuf.substring(howMuch - i), + tickIsVowel, + weHaveSeenVowelAlready + || (newHead.getRight() != null + && !"+".equals(newHead.getRight()) + && !"-".equals(newHead.getRight())), + ttraits)).hasSimpleError(ttraits)) { + newTail.prepend(newHead); + return newTail; + } + } + // It didn't work. Return the first thing we'd thought + // of: head appended with tail. (I.e., fall through.) + } + tail.prepend(head); + return tail; + } + + // TODO(DLC)[EWTS->Tibetan]: doc + private static TPairList breakHelperEWTS(String ewts, TTraits ttraits /* TODO(DLC)[EWTS->Tibetan]: use */) { + + // base case for our recursion: + if ("".equals(ewts)) + return new TPairList(); + + StringBuffer ewtsBuf = new StringBuffer(ewts); + int howMuchBuf[] = new int[1]; + TPair head = getFirstConsonantAndVowel(ewtsBuf, howMuchBuf, ttraits); + int howMuch = howMuchBuf[0]; + + TPairList tail; + if ((tail = breakHelperEWTS(ewtsBuf.substring(howMuch), + ttraits)).hasSimpleError(ttraits)) { + for (int i = 1; i < howMuch; i++) { + // try giving i characters back if that leaves us with + // a legal head and makes the rest free of simple + // errors. + TPairList newTail = null; + TPair newHead; + if ((newHead = head.minusNRightmostTransliterationCharacters(i)).isLegal() + && !(newTail + = breakHelperEWTS(ewtsBuf.substring(howMuch - i), ttraits)).hasSimpleError(ttraits)) { newTail.prepend(newHead); return newTail; } @@ -162,14 +220,16 @@ class TPairListFactory { * A later phase will need to turn that into {N+YE} or an error * or whatever you like. howMuch[0] will be set to the number of * characters of acip that this call has consumed. */ - private static TPair getFirstConsonantAndVowel(StringBuffer acip, - int howMuch[]) { + private static TPair getFirstConsonantAndVowel(StringBuffer acip, // TODO(DLC)[EWTS->Tibetan]: function name needs ACIP in it? + int howMuch[], + TTraits ttraits) { // Note that it is *not* the case that if acip.substring(0, N) // is legal (according to TPair.isLegal()), then // acip.substring(0, N-1) is legal for all N. For example, - // think of {shA} and {KshA}. However, 's' is the only tricky - // fellow, so it is true that acip.substring(0, N-1) is either - // legal or ends with 's' if acip.substring(0, N) is legal. + // think of ACIP's {shA} and {KshA}. However, 's' is the only + // tricky fellow, so it is true that acip.substring(0, N-1) is + // either legal or ends with 's' if acip.substring(0, N) is + // legal. // // We don't, however, use this approach. We just try to find // a consonant of length 3, and then, failing that, of length @@ -180,9 +240,9 @@ class TPairListFactory { howMuch[0] = 0; return new TPair(null, null); } - if (acip.charAt(0) == '-') { + if (acip.charAt(0) == ttraits.disambiguatorChar()) { howMuch[0] = 1; - return new TPair(null, "-"); + return new TPair(null, ttraits.disambiguator()); } char ch = acip.charAt(0); @@ -190,32 +250,33 @@ class TPairListFactory { // like seeing 1-2-3-4. if (ch >= '0' && ch <= '9') { howMuch[0] = 1; // not 2... - return new TPair(acip.substring(0, 1), (xl == 1) ? null : "-"); + return new TPair(acip.substring(0, 1), (xl == 1) ? null : ttraits.disambiguator()); } String l = null, r = null; - for (i = Math.min(ACIPRules.MAX_CONSONANT_LENGTH, xl); i >= 1; i--) { + for (i = Math.min(ttraits.maxConsonantLength(), xl); i >= 1; i--) { String t = null; - if (ACIPRules.isConsonant(t = acip.substring(0, i))) { + if (ttraits.isConsonant(t = acip.substring(0, i))) { l = t; break; } } int ll = (null == l) ? 0 : l.length(); - if (null != l && xl > ll && acip.charAt(ll) == '-') { + if (null != l && xl > ll && acip.charAt(ll) == ttraits.disambiguatorChar()) { howMuch[0] = l.length() + 1; - return new TPair(l, "-"); + return new TPair(l, ttraits.disambiguator()); } if (null != l && xl > ll && acip.charAt(ll) == '+') { howMuch[0] = l.length() + 1; return new TPair(l, "+"); } - for (i = Math.min(ACIPRules.MAX_VOWEL_LENGTH, xl - ll); i >= 1; i--) { + for (i = Math.min(ttraits.maxWowelLength(), xl - ll); i >= 1; i--) { String t = null; - if (ACIPRules.isVowel(t = acip.substring(ll, ll + i)) + if (ttraits.isWowel(t = acip.substring(ll, ll + i)) // Or these, which we massage into "Am", "Am:", and // "A:" because I didn't think {Pm} should be treated // like {PAm} originally: + // TODO(DLC)[EWTS->Tibetan]: NOW NIGHTMARE || "m".equals(t) || "m:".equals(t) || ":".equals(t)) { r = t; break; @@ -224,14 +285,14 @@ class TPairListFactory { // Treat {BATA+SA'I} like {BAT+SA'I}: int z; - if (null != l && "A".equals(r) && ((z = ll + "A".length()) < xl) + if (null != l && /* TODO(DLC)[EWTS->Tibetan]: */"A".equals(r) && ((z = ll + /* TODO(DLC)[EWTS->Tibetan]: */"A".length()) < xl) && acip.charAt(z) == '+') { acip.deleteCharAt(z-1); howMuch[0] = l.length() + 1; return new TPair(l, "+"); } - // Allow Pm to mean PAm, P: to mean PA:, Pm: to mean PAm:. + // Allow Pm to mean PAm, P: to mean PA:, Pm: to mean PAm:. /* TODO(DLC)[EWTS->Tibetan]: */ int mod = 0; if ("m".equals(r)) { r = "Am"; mod = -1; } if (":".equals(r)) { r = "A:"; mod = -1; } @@ -239,19 +300,20 @@ class TPairListFactory { if (":m".equals(r)) { r = "A:m"; mod = -1; } // not seen, though... - // what if we see a character that's not part of any vowel or + // what if we see a character that's not part of any wowel or // consonant? We return it. if (null == l && null == r) { howMuch[0] = 1; // not 2... - // add a '-' to avoid exponentials: - return new TPair(acip.substring(0, 1), (xl == 1) ? null : "-"); + // add a disambiguator to avoid exponential running time: + return new TPair(acip.substring(0, 1), + (xl == 1) ? null : ttraits.disambiguator()); } howMuch[0] = (((l == null) ? 0 : l.length()) + ((r == null) ? 0 : r.length()) + mod); return new TPair(l, r); - } + } // TODO(DLC)[EWTS->Tibetan]: } diff --git a/source/org/thdl/tib/text/ttt/TParseTree.java b/source/org/thdl/tib/text/ttt/TParseTree.java index 55c3cf4..2dba84a 100644 --- a/source/org/thdl/tib/text/ttt/TParseTree.java +++ b/source/org/thdl/tib/text/ttt/TParseTree.java @@ -246,7 +246,7 @@ class TParseTree { return al.toString(); } - /** Returns true if and only if either x is an TParseTree + /** Returns true if and only if either x is a TParseTree * object representing the same TPairLists in the same order * or x is a String that is equals to the result of {@link * #toString()}. */ diff --git a/source/org/thdl/tib/text/ttt/TStackList.java b/source/org/thdl/tib/text/ttt/TStackList.java index 5a74e36..6007acf 100644 --- a/source/org/thdl/tib/text/ttt/TStackList.java +++ b/source/org/thdl/tib/text/ttt/TStackList.java @@ -100,7 +100,7 @@ class TStackList { return al.toString(); } - /** Returns true if and only if either x is an TStackList + /** Returns true if and only if either x is a TStackList * object representing the same TPairLists in the same * order or x is a String that is equals to the result of {@link * #toString()}. */ diff --git a/source/org/thdl/tib/text/ttt/TStackListList.java b/source/org/thdl/tib/text/ttt/TStackListList.java index 6e20840..fe60ab6 100644 --- a/source/org/thdl/tib/text/ttt/TStackListList.java +++ b/source/org/thdl/tib/text/ttt/TStackListList.java @@ -62,7 +62,7 @@ class TStackListList { return al.toString(); } - /** Returns true if and only if either x is an TStackListList + /** Returns true if and only if either x is a TStackListList * object representing the same TStackList objects in the same * order or x is a String that is equals to the result of {@link * #toString()}. */ diff --git a/source/org/thdl/tib/text/ttt/TString.java b/source/org/thdl/tib/text/ttt/TString.java index 26636b1..17c1656 100644 --- a/source/org/thdl/tib/text/ttt/TString.java +++ b/source/org/thdl/tib/text/ttt/TString.java @@ -53,7 +53,7 @@ public class TString { return encoding; } - /** Returns true if and only if an TString with type type + /** Returns true if and only if a TString with type type * is to be converted to something other than Tibetan text. * (Chinese Unicode, Latin, etc. all qualify as non-Tibetan.) */ public boolean isLatin() { diff --git a/source/org/thdl/tib/text/ttt/TTGCList.java b/source/org/thdl/tib/text/ttt/TTGCList.java index c0f4c5a..0a97971 100644 --- a/source/org/thdl/tib/text/ttt/TTGCList.java +++ b/source/org/thdl/tib/text/ttt/TTGCList.java @@ -53,7 +53,7 @@ class TTGCList implements TGCList { /** Returns the number of TGCPairs in this list. */ public int size() { return al.size(); } - /** Returns a zero-based index of an TPairList inside the stack + /** Returns a zero-based index of a TPairList inside the stack * list from which this list was constructed. This pair list is * the one that caused the TGCPair at index tgcPairIndex to come * into existence. */ diff --git a/source/org/thdl/tib/text/ttt/TTraits.java b/source/org/thdl/tib/text/ttt/TTraits.java new file mode 100644 index 0000000..41bcb66 --- /dev/null +++ b/source/org/thdl/tib/text/ttt/TTraits.java @@ -0,0 +1,70 @@ +/* +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 2003-2004 THDL. +All Rights Reserved. + +Contributor(s): ______________________________________. +*/ + +package org.thdl.tib.text.ttt; + +/** A TTraits object encapsulates all the things that make a + * particular Roman transliteration scheme unique. If both EWTS and + * ACIP transliterations have a property in common, then it's likely + * encoded in a manner that's hard to modify. But if they differ in + * some respect, then that difference should be encoded in a TTraits + * object. + * + *
It is very likely that classes that implement this interface + * will choose to use the design pattern 'singleton'. */ +interface TTraits { + /** Returns the disambiguator for this transliteration scheme, + * which had better be a string containing just one character + * lest {@link #disambiguatorChar()} become nonsensical for + * certain transliteration systems. A disambiguator is the + * string that separates two bits of transliteration that would + * otherwise be interpreted differently. The canonical example + * is EWTS's {gya} versus {g.ya}, which, due to the disambiguator + * in the latter, are different sequences of Unicode. */ + String disambiguator(); + + /** For convenience, a convenience that is possible because ACIP's + * and EWTS's disambiguator strings both have length one, this + * method returns the sole character in the string returned by + * {@link #disambiguator()}. */ + char disambiguatorChar(); + + /** Returns the maximum number of characters of transliteration + * required to denote a Tibetan consonant. */ + int maxConsonantLength(); + + /** Returns the maximum number of characters of transliteration + * required to denote a Tibetan wowel, i.e. a vowel or one or + * more hangers-on like U+TODO(DLC)[EWTS->Tibetan]:_NOW or both. */ + int maxWowelLength(); + // TODO(DLC)[EWTS->Tibetan]: use the term 'wowel' everywhere, never "vowel" unless you mean just {e,i,o,u} + + /** Returns true if and only if s is a stretch of + * transliteration corresponding to a Tibetan consonant (without + * any wowel) */ + boolean isConsonant(String s); + + /** Returns true if and only if s is a stretch of + * transliteration corresponding to a Tibetan wowel (without any + * [achen or other] consonant) */ + boolean isWowel(String s); + + /** Returns true if and only if the pair given has a simple error + * other than being a mere disambiguator. */ + boolean hasSimpleError(TPair p); +}