Merge "Add initial reordering rules for Myanmar" (61ddac28) · Commits · e / os / android_packages_inputmethods_LatinIME

java/src/com/android/inputmethod/event/MyanmarReordering.java

+200 −3

Original line number	Diff line number	Diff line
		@@ -16,23 +16,220 @@

		package com.android.inputmethod.event;

		import com.android.inputmethod.latin.Constants;
		import com.android.inputmethod.latin.utils.CollectionUtils;

		import java.util.ArrayList;
		import java.util.Arrays;

		/**
		* A combiner that reorders input for Myanmar.
		*/
		public class MyanmarReordering implements Combiner {
		// U+1031 MYANMAR VOWEL SIGN E
		private final static int VOWEL_E = 0x1031; // Code point for vowel E that we need to reorder
		// U+200C ZERO WIDTH NON-JOINER
		// U+200B ZERO WIDTH SPACE
		private final static int ZERO_WIDTH_NON_JOINER = 0x200B; // should be 0x200C

		private final ArrayList<Event> mCurrentEvents = CollectionUtils.newArrayList();

		// List of consonants :
		// U+1000 MYANMAR LETTER KA
		// U+1001 MYANMAR LETTER KHA
		// U+1002 MYANMAR LETTER GA
		// U+1003 MYANMAR LETTER GHA
		// U+1004 MYANMAR LETTER NGA
		// U+1005 MYANMAR LETTER CA
		// U+1006 MYANMAR LETTER CHA
		// U+1007 MYANMAR LETTER JA
		// U+1008 MYANMAR LETTER JHA
		// U+1009 MYANMAR LETTER NYA
		// U+100A MYANMAR LETTER NNYA
		// U+100B MYANMAR LETTER TTA
		// U+100C MYANMAR LETTER TTHA
		// U+100D MYANMAR LETTER DDA
		// U+100E MYANMAR LETTER DDHA
		// U+100F MYANMAR LETTER NNA
		// U+1010 MYANMAR LETTER TA
		// U+1011 MYANMAR LETTER THA
		// U+1012 MYANMAR LETTER DA
		// U+1013 MYANMAR LETTER DHA
		// U+1014 MYANMAR LETTER NA
		// U+1015 MYANMAR LETTER PA
		// U+1016 MYANMAR LETTER PHA
		// U+1017 MYANMAR LETTER BA
		// U+1018 MYANMAR LETTER BHA
		// U+1019 MYANMAR LETTER MA
		// U+101A MYANMAR LETTER YA
		// U+101B MYANMAR LETTER RA
		// U+101C MYANMAR LETTER LA
		// U+101D MYANMAR LETTER WA
		// U+101E MYANMAR LETTER SA
		// U+101F MYANMAR LETTER HA
		// U+1020 MYANMAR LETTER LLA
		// U+103F MYANMAR LETTER GREAT SA
		private static boolean isConsonant(final int codePoint) {
		return (codePoint >= 0x1000 && codePoint <= 0x1020) \|\| 0x103F == codePoint;
		}

		// List of medials :
		// U+103B MYANMAR CONSONANT SIGN MEDIAL YA
		// U+103C MYANMAR CONSONANT SIGN MEDIAL RA
		// U+103D MYANMAR CONSONANT SIGN MEDIAL WA
		// U+103E MYANMAR CONSONANT SIGN MEDIAL HA
		// U+105E MYANMAR CONSONANT SIGN MON MEDIAL NA
		// U+105F MYANMAR CONSONANT SIGN MON MEDIAL MA
		// U+1060 MYANMAR CONSONANT SIGN MON MEDIAL LA
		// U+1082 MYANMAR CONSONANT SIGN SHAN MEDIAL WA
		private static int[] MEDIAL_LIST = { 0x103B, 0x103C, 0x103D, 0x103E,
		0x105E, 0x105F, 0x1060, 0x1082};
		private static boolean isMedial(final int codePoint) {
		return Arrays.binarySearch(MEDIAL_LIST, codePoint) >= 0;
		}

		private static boolean isConsonantOrMedial(final int codePoint) {
		return isConsonant(codePoint) \|\| isMedial(codePoint);
		}

		private Event getLastEvent() {
		final int size = mCurrentEvents.size();
		if (size <= 0) {
		return null;
		}
		return mCurrentEvents.get(size - 1);
		}

		private CharSequence getCharSequence() {
		final StringBuilder s = new StringBuilder();
		for (final Event e : mCurrentEvents) {
		s.appendCodePoint(e.mCodePoint);
		}
		return s;
		}

		/**
		* Clears the currently combining stream of events and returns the resulting software text
		* event corresponding to the stream. Optionally adds a new event to the cleared stream.
		* @param newEvent the new event to add to the stream. null if none.
		* @return the resulting software text event. Null if none.
		*/
		private Event clearAndGetResultingEvent(final Event newEvent) {
		final CharSequence combinedText;
		if (mCurrentEvents.size() > 0) {
		combinedText = getCharSequence();
		mCurrentEvents.clear();
		} else {
		combinedText = null;
		}
		if (null != newEvent) {
		mCurrentEvents.add(newEvent);
		}
		return null == combinedText ? null
		: Event.createSoftwareTextEvent(combinedText, Event.NOT_A_KEY_CODE);
		}

		@Override
		public Event processEvent(ArrayList<Event> previousEvents, Event event) {
		return event;
		public Event processEvent(ArrayList<Event> previousEvents, Event newEvent) {
		final int codePoint = newEvent.mCodePoint;
		if (VOWEL_E == codePoint) {
		final Event lastEvent = getLastEvent();
		if (null == lastEvent) {
		mCurrentEvents.add(newEvent);
		return null;
		} else if (isConsonantOrMedial(lastEvent.mCodePoint)) {
		final Event resultingEvent = clearAndGetResultingEvent(null);
		mCurrentEvents.add(Event.createSoftwareKeypressEvent(ZERO_WIDTH_NON_JOINER,
		Event.NOT_A_KEY_CODE,
		Constants.NOT_A_COORDINATE, Constants.NOT_A_COORDINATE,
		false /* isKeyRepeat */));
		mCurrentEvents.add(newEvent);
		return resultingEvent;
		} else { // VOWEL_E == lastCodePoint. But if that was anything else this is correct too.
		return clearAndGetResultingEvent(newEvent);
		}
		} if (isConsonant(codePoint)) {
		final Event lastEvent = getLastEvent();
		if (null == lastEvent) {
		mCurrentEvents.add(newEvent);
		return null;
		} else if (VOWEL_E == lastEvent.mCodePoint) {
		final int eventSize = mCurrentEvents.size();
		if (eventSize >= 2
		&& mCurrentEvents.get(eventSize - 2).mCodePoint == ZERO_WIDTH_NON_JOINER) {
		// We have a ZWJN before a vowel E. We need to remove the ZWNJ and then
		// reorder the vowel with respect to the consonant.
		mCurrentEvents.remove(eventSize - 1);
		mCurrentEvents.remove(eventSize - 2);
		mCurrentEvents.add(newEvent);
		mCurrentEvents.add(lastEvent);
		return null;
		}
		// If there is already a consonant, then we are starting a new syllable.
		for (int i = eventSize - 2; i >= 0; --i) {
		if (isConsonant(mCurrentEvents.get(i).mCodePoint)) {
		return clearAndGetResultingEvent(newEvent);
		}
		}
		// If we come here, we didn't have a consonant so we reorder
		mCurrentEvents.remove(eventSize - 1);
		mCurrentEvents.add(newEvent);
		mCurrentEvents.add(lastEvent);
		return null;
		} else { // lastCodePoint is a consonant/medial. But if it's something else it's fine
		return clearAndGetResultingEvent(newEvent);
		}
		} else if (isMedial(codePoint)) {
		final Event lastEvent = getLastEvent();
		if (null == lastEvent) {
		mCurrentEvents.add(newEvent);
		return null;
		} else if (VOWEL_E == lastEvent.mCodePoint) {
		final int eventSize = mCurrentEvents.size();
		// If there is already a consonant, then we are in the middle of a syllable, and we
		// need to reorder.
		boolean hasConsonant = false;
		for (int i = eventSize - 2; i >= 0; --i) {
		if (isConsonant(mCurrentEvents.get(i).mCodePoint)) {
		hasConsonant = true;
		break;
		}
		}
		if (hasConsonant) {
		mCurrentEvents.remove(eventSize - 1);
		mCurrentEvents.add(newEvent);
		mCurrentEvents.add(lastEvent);
		return null;
		}
		// Otherwise, we just commit everything.
		return clearAndGetResultingEvent(null);
		} else { // lastCodePoint is a consonant/medial. But if it's something else it's fine
		return clearAndGetResultingEvent(newEvent);
		}
		} else if (Constants.CODE_DELETE == newEvent.mKeyCode) {
		if (mCurrentEvents.size() > 0) {
		mCurrentEvents.remove(mCurrentEvents.size() - 1);
		return null;
		}
		}
		// This character is not part of the combining scheme, so we should reset everything.
		if (mCurrentEvents.size() > 0) {
		// If we have events in flight, then add the new event and return the resulting event.
		mCurrentEvents.add(newEvent);
		return clearAndGetResultingEvent(null);
		} else {
		// If we don't have any events in flight, then just pass this one through.
		return newEvent;
		}
		}

		@Override
		public CharSequence getCombiningStateFeedback() {
		return "";
		return getCharSequence();
		}

		@Override
		public void reset() {
		mCurrentEvents.clear();
		}
		}

java/src/com/android/inputmethod/latin/settings/SettingsValues.java

+2 −1

Original line number	Diff line number	Diff line
		@@ -205,7 +205,8 @@ public final class SettingsValues {
		}

		public boolean isWordCodePoint(final int code) {
		return Character.isLetter(code) \|\| isWordConnector(code);
		return Character.isLetter(code) \|\| isWordConnector(code)
		\|\| Character.COMBINING_SPACING_MARK == Character.getType(code);
		}

		public boolean isUsuallyPrecededBySpace(final int code) {

tests/src/com/android/inputmethod/latin/InputLogicTestsReorderingMyanmar.java

0 → 100644

+232 −0

Original line number	Diff line number	Diff line
		/*
		* Copyright (C) 2012 The Android Open Source Project
		*
		* Licensed under the Apache License, Version 2.0 (the "License");
		* you may not use this file except in compliance with the License.
		* You may obtain a copy of the License at
		*
		* http://www.apache.org/licenses/LICENSE-2.0
		*
		* Unless required by applicable law or agreed to in writing, software
		* distributed under the License is distributed on an "AS IS" BASIS,
		* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
		* See the License for the specific language governing permissions and
		* limitations under the License.
		*/

		package com.android.inputmethod.latin;

		import android.test.suitebuilder.annotation.LargeTest;
		import android.util.Pair;

		/*
		* Relevant characters for this test :
		* Spurs the need to reorder :
		* U+1031 MYANMAR VOWEL SIGN E : ေ
		* U+1004 U+103A U+1039 Kinzi. It's a compound character.
		*
		* List of consonants :
		* U+1000 MYANMAR LETTER KA က
		* U+1001 MYANMAR LETTER KHA ခ
		* U+1002 MYANMAR LETTER GA ဂ
		* U+1003 MYANMAR LETTER GHA ဃ
		* U+1004 MYANMAR LETTER NGA င
		* U+1005 MYANMAR LETTER CA စ
		* U+1006 MYANMAR LETTER CHA ဆ
		* U+1007 MYANMAR LETTER JA ဇ
		* U+1008 MYANMAR LETTER JHA ဈ
		* U+1009 MYANMAR LETTER NYA ဉ
		* U+100A MYANMAR LETTER NNYA ည
		* U+100B MYANMAR LETTER TTA ဋ
		* U+100C MYANMAR LETTER TTHA ဌ
		* U+100D MYANMAR LETTER DDA ဍ
		* U+100E MYANMAR LETTER DDHA ဎ
		* U+100F MYANMAR LETTER NNA ဏ
		* U+1010 MYANMAR LETTER TA တ
		* U+1011 MYANMAR LETTER THA ထ
		* U+1012 MYANMAR LETTER DA ဒ
		* U+1013 MYANMAR LETTER DHA ဓ
		* U+1014 MYANMAR LETTER NA န
		* U+1015 MYANMAR LETTER PA ပ
		* U+1016 MYANMAR LETTER PHA ဖ
		* U+1017 MYANMAR LETTER BA ဗ
		* U+1018 MYANMAR LETTER BHA ဘ
		* U+1019 MYANMAR LETTER MA မ
		* U+101A MYANMAR LETTER YA ယ
		* U+101B MYANMAR LETTER RA ရ
		* U+101C MYANMAR LETTER LA လ
		* U+101D MYANMAR LETTER WA ဝ
		* U+101E MYANMAR LETTER SA သ
		* U+101F MYANMAR LETTER HA ဟ
		* U+1020 MYANMAR LETTER LLA ဠ
		* U+103F MYANMAR LETTER GREAT SA ဿ
		*
		* List of medials :
		* U+103B MYANMAR CONSONANT SIGN MEDIAL YA ျ
		* U+103C MYANMAR CONSONANT SIGN MEDIAL RA ြ
		* U+103D MYANMAR CONSONANT SIGN MEDIAL WA ွ
		* U+103E MYANMAR CONSONANT SIGN MEDIAL HA ှ
		* U+105E MYANMAR CONSONANT SIGN MON MEDIAL NA ၞ
		* U+105F MYANMAR CONSONANT SIGN MON MEDIAL MA ၟ
		* U+1060 MYANMAR CONSONANT SIGN MON MEDIAL LA ၠ
		* U+1082 MYANMAR CONSONANT SIGN SHAN MEDIAL WA ႂ
		*
		* Other relevant characters :
		* U+200C ZERO WIDTH NON-JOINER
		* U+200B ZERO WIDTH SPACE
		*/

		@LargeTest
		public class InputLogicTestsReorderingMyanmar extends InputTestsBase {
		// The tests are formatted as follows.
		// Each test is an entry in the array of Pair arrays.

		// One test is an array of pairs. Each pair contains, in the `first' member,
		// the code points that the next key press should contain. In the `second'
		// member is stored the string that should be in the text view after this
		// key press.

		private static final Pair[][] TESTS = {

		// Tests for U+1031 MYANMAR VOWEL SIGN E : ေ
		new Pair[] { // Type : U+1031 U+1000 U+101F ေ က ဟ
		Pair.create(new int[] { 0x1031 }, "\u1031"), // ေ
		Pair.create(new int[] { 0x1000 }, "\u1000\u1031"), // ကေ
		Pair.create(new int[] { 0x101F }, "\u1000\u1031\u101F") // ကေဟ
		},

		new Pair[] { // Type : U+1000 U+1031 U+101F က ေ ဟ
		Pair.create(new int[] { 0x1000 }, "\u1000"), // က
		Pair.create(new int[] { 0x1031 }, "\u1000\u200B\u1031"), // က‌ေ
		Pair.create(new int[] { 0x101F }, "\u1000\u101F\u1031") // ကဟေ
		},

		new Pair[] { // Type : U+1031 U+101D U+103E U+1018 ေ ဝ ှ ဘ
		Pair.create(new int[] { 0x1031 }, "\u1031"), // ေ
		Pair.create(new int[] { 0x101D }, "\u101D\u1031"), // ဝေ
		Pair.create(new int[] { 0x103E }, "\u101D\u103E\u1031"), // ဝှေ
		Pair.create(new int[] { 0x1018 }, "\u101D\u103E\u1031\u1018") // ဝှေဘ
		},

		new Pair[] { // Type : U+1031 U+1014 U+1031 U+1000 U+102C U+1004 U+103A U+1038 U+101C
		// U+102C U+1038 U+104B ေ န ေ က ာ င ် း လ ာ း ။
		Pair.create(new int[] { 0x1031 }, "\u1031"), // ေ
		Pair.create(new int[] { 0x1014 }, "\u1014\u1031"), // နေ
		Pair.create(new int[] { 0x1031 }, "\u1014\u1031\u1031"), // နေ‌ေ
		Pair.create(new int[] { 0x1000 }, "\u1014\u1031\u1000\u1031"), // နေကေ
		Pair.create(new int[] { 0x102C }, "\u1014\u1031\u1000\u1031\u102C"), // နေကော
		Pair.create(new int[] { 0x1004 }, "\u1014\u1031\u1000\u1031\u102C\u1004"), // နေကောင
		Pair.create(new int[] { 0x103A }, // နေကောင်
		"\u1014\u1031\u1000\u1031\u102C\u1004\u103A"),
		Pair.create(new int[] { 0x1038 }, // နေကောင်း
		"\u1014\u1031\u1000\u1031\u102C\u1004\u103A\u1038"),
		Pair.create(new int[] { 0x101C }, // နေကောင်းလ
		"\u1014\u1031\u1000\u1031\u102C\u1004\u103A\u1038\u101C"),
		Pair.create(new int[] { 0x102C }, // နေကောင်းလာ
		"\u1014\u1031\u1000\u1031\u102C\u1004\u103A\u1038\u101C\u102C"),
		Pair.create(new int[] { 0x1038 }, // နေကောင်းလား
		"\u1014\u1031\u1000\u1031\u102C\u1004\u103A\u1038\u101C\u102C\u1038"),
		Pair.create(new int[] { 0x104B }, // နေကောင်းလား။
		"\u1014\u1031\u1000\u1031\u102C\u1004\u103A\u1038\u101C\u102C\u1038\u104B")
		},

		new Pair[] { // Type : U+1031 U+1031 U+1031 U+1000 ေ ေ ေ က
		Pair.create(new int[] { 0x1031 }, "\u1031"), // ေ
		Pair.create(new int[] { 0x1031 }, "\u1031\u1031"), // ေေ
		Pair.create(new int[] { 0x1031 }, "\u1031\u1031\u1031"), // U+1031ေေေ
		Pair.create(new int[] { 0x1000 }, "\u1031\u1031\u1000\u1031") // ေေကေ
		},

		new Pair[] { // Type : U+1031 U+1001 U+103B U+103D U+1038 ေ ခ ျ ွ း
		Pair.create(new int[] { 0x1031 }, "\u1031"), // ေ
		Pair.create(new int[] { 0x1001 }, "\u1001\u1031"), // ခေ
		Pair.create(new int[] { 0x103B }, "\u1001\u103B\u1031"), // ချေ
		Pair.create(new int[] { 0x103D }, "\u1001\u103B\u103D\u1031"), // ချွေ
		Pair.create(new int[] { 0x1038 }, "\u1001\u103B\u103D\u1031\u1038") // ချွေး
		},

		// Tests for Kinzi U+1004 U+103A U+1039 :

		/* Kinzi reordering is not implemented yet. Uncomment these tests when it is.

		new Pair[] { // Type : U+1021 U+1002 (U+1004 U+103A U+1039)
		// U+101C U+1014 U+103A အ ဂ (င ် ္) လ န ်
		Pair.create(new int[] { 0x1021 }, "\u1021"), // အ
		Pair.create(new int[] { 0x1002 }, "\u1021\u1002"), // အဂ
		Pair.create(new int[] { 0x1004, 0x103A, 0x1039 }, // အင်္ဂ
		"\u1021\u1004\u103A\u1039\u1002"),
		Pair.create(new int[] { 0x101C }, // အင်္ဂလ
		"\u1021\u1004\u103A\u1039\u1002\u101C"),
		Pair.create(new int[] { 0x1014 }, // အင်္ဂလန
		"\u1021\u1004\u103A\u1039\u1002\u101C\u1014"),
		Pair.create(new int[] { 0x103A }, // အင်္ဂလန်
		"\u1021\u1004\u103A\u1039\u1002\u101C\u1014\u103A")
		},

		new Pair[] { //Type : kinzi after a whole syllable U+101E U+1001 U+103B U+102D U+102F
		// (U+1004 U+103A U+1039) U+1004 U+103A U+1038 သ ခ ျ ိ ု င ် ္ င ် း
		Pair.create(new int[] { 0x101E }, "\u101E"), // သခ
		Pair.create(new int[] { 0x1001 }, "\u101E\u1001"), // သခ
		Pair.create(new int[] { 0x103B }, "\u101E\u1001\u103B"), // သချ
		Pair.create(new int[] { 0x102D }, "\u101E\u1001\u103B\u102D"), // သချိ
		Pair.create(new int[] { 0x102F }, "\u101E\u1001\u103B\u102D\u102F"), // သချို
		Pair.create(new int[] { 0x1004, 0x103A, 0x1039}, // သင်္ချို
		"\u101E\u1004\u103A\u1039\u1001\u103B\u102D\u102F"),
		Pair.create(new int[] { 0x1004 }, // သင်္ချိုင
		"\u101E\u1004\u103A\u1039\u1001\u103B\u102D\u102F\u1004"),
		Pair.create(new int[] { 0x103A }, // သင်္ချိုင်
		"\u101E\u1004\u103A\u1039\u1001\u103B\u102D\u102F\u1004\u103A"),
		Pair.create(new int[] { 0x1038 }, // သင်္ချိုင်း
		"\u101E\u1004\u103A\u1039\u1001\u103B\u102D\u102F\u1004\u103A\u1038")
		},

		new Pair[] { // Type : kinzi after the consonant U+101E U+1001 (U+1004 U+103A U+1039)
		// U+103B U+102D U+102F U+1004 U+103A U+1038 သ ခ င ် ္ ျ ိ ု င ် း
		Pair.create(new int[] { 0x101E }, "\u101E"), // သခ
		Pair.create(new int[] { 0x1001 }, "\u101E\u1001"), // သခ
		Pair.create(new int[] { 0x1004, 0x103A, 0x1039 }, // သင်္ခ
		"\u101E\u1004\u103A\u1039\u1001"),
		Pair.create(new int[] { 0x103B }, // သင်္ချ
		"\u101E\u1004\u103A\u1039\u1001\u103B"),
		Pair.create(new int[] { 0x102D }, // သင်္ချိ
		"\u101E\u1004\u103A\u1039\u1001\u103B\u102D"),
		Pair.create(new int[] { 0x102F }, // သင်္ချို
		"\u101E\u1004\u103A\u1039\u1001\u103B\u102D\u102F"),
		Pair.create(new int[] { 0x1004 }, // သင်္ချိုင
		"\u101E\u1004\u103A\u1039\u1001\u103B\u102D\u102F\u1004"),
		Pair.create(new int[] { 0x103A }, // သင်္ချိုင်
		"\u101E\u1004\u103A\u1039\u1001\u103B\u102D\u102F\u1004\u103A"),
		Pair.create(new int[] { 0x1038 }, // သင်္ချိုင်း
		"\u101E\u1004\u103A\u1039\u1001\u103B\u102D\u102F\u1004\u103A\u1038")
		},
		*/
		};

		private void doMyanmarTest(final int testNumber, final Pair[] test) {
		int stepNumber = 0;
		for (final Pair<int[], String> step : test) {
		++stepNumber;
		final int[] input = step.first;
		final String expectedResult = step.second;
		if (input.length > 1) {
		mLatinIME.onTextInput(new String(input, 0, input.length));
		} else {
		type(input[0]);
		}
		assertEquals("Myanmar reordering test " + testNumber + ", step " + stepNumber,
		expectedResult, mEditText.getText().toString());
		}
		}

		public void testMyanmarReordering() {
		int testNumber = 0;
		changeLanguage("mm_MY");
		for (final Pair[] test : TESTS) {
		// Small trick to reset LatinIME : setText("") and send updateSelection with values
		// LatinIME has never seen, and cursor pos 0,0.
		mEditText.setText("");
		mLatinIME.onUpdateSelection(1, 1, 0, 0, -1, -1);
		doMyanmarTest(++testNumber, test);
		}
		}
		}