Develop Reference

Get short weekdays names for Portuguese language in Android

I need to return the short version of the weekdays names using getShortWeekdays() in Android. I tested with different device language settings. Although it seems to be working with various language settings, unfortunately when the selected language is Portuguese (Portugal), it returns the full day names. Any advice?
DateFormatSymbols symbols = new DateFormatSymbols(Locale.getDefault());
String[] dayNames = symbols.getShortWeekdays();
Output if language is English (United States):
["", "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat"]
Output if language is Portuguese (Brasil):
["", "dom.", "seg.", "ter.", "qua.", "qui.", "sex.", "sáb."]
Output if language is Portuguese (Portugal):
["", "domingo", "segunda", "terça", "quarta", "quinta", "sexta", "sábado"]

At first, I wondered if using a different package for the method's class would produce different results, but it didn't.
final java.util.Locale pt_BR = new java.util.Locale("pt", "BR");
android.icu.text.DateFormatSymbols symbolsPtBr1 = new android.icu.text.DateFormatSymbols(pt_BR);
String[] dayNamesBtBr1 = symbolsPtBr1.getShortWeekdays();
Log.e("LOCALE_ANDROID", String.format("%s(%s): (android.icu.text): %s", pt_BR.getDisplayVariant(), pt_BR.getDisplayName(), Arrays.toString(dayNamesBtBr1)));
java.text.DateFormatSymbols symbolsPtBr2 = new java.text.DateFormatSymbols(pt_BR);
String[] dayNamesBtBr2 = symbolsPtBr2.getShortWeekdays();
Log.e("LOCALE_ANDROID", String.format("%s(%s): (java.text): %s", pt_BR.getDisplayVariant(), pt_BR.getDisplayName(), Arrays.toString(dayNamesBtBr2)));
final java.util.Locale pt_PT = new java.util.Locale("pt", "PT");
android.icu.text.DateFormatSymbols symbolsPtPt1 = new android.icu.text.DateFormatSymbols(pt_PT);
String[] dayNamesBtPt1 = symbolsPtPt1.getShortWeekdays();
Log.e("LOCALE_ANDROID", String.format("%s(%s): (android.icu.text): %s", pt_PT.getDisplayVariant(), pt_PT.getDisplayName(), Arrays.toString(dayNamesBtPt1)));
java.text.DateFormatSymbols symbolsPtPt2 = new java.text.DateFormatSymbols(pt_PT);
String[] dayNamesBtPt2 = symbolsPtPt2.getShortWeekdays();
Log.e("LOCALE_ANDROID", String.format("%s(%s): (java.text): %s", pt_PT.getDisplayVariant(), pt_PT.getDisplayName(), Arrays.toString(dayNamesBtPt2)));
The log output (tested on Android 7.0):
LOCALE_ANDROID: (Portuguese (Brazil)): (android.icu.text): [, Dom, Seg, Ter, Qua, Qui, Sex, Sáb]
LOCALE_ANDROID: (Portuguese (Brazil)): (java.text): [, Dom, Seg, Ter, Qua, Qui, Sex, Sáb]
LOCALE_ANDROID: (Portuguese (Portugal)): (android.icu.text): [, domingo, segunda, terça, quarta, quinta, sexta, sábado]
LOCALE_ANDROID: (Portuguese (Portugal)): (java.text): [, domingo, segunda, terça, quarta, quinta, sexta, sábado]
Therefore, you can just truncate each weekday to a two- or three-letter symbol length with substring(0, len) and capitalize it with StringUtils.capitalize() method as this is the approach used for the issue in OsmAnd https://github.com/osmandapp/Osmand/issues/5115

If the language is english then use an another link than if it's german

how does the if condition look like?
if(Locale.getDefault().equals("English")){
Log.i("Language","Englisch");
} else if(Locale.getDefault().equals("Deutsch")){
Log.i("Language","Deutsch");
}
This won't work

Locale.getDefault() will return a static Locale Object, not a String. So calling Locale.getDefault().equals("English") will not work.
Try this:
String language = Locale.getDefault().getLanguage();
if (language.equals("en"))
{
// Use English link
}
else if (language.equals("de"))
{
// Use German link
}
public String getLanguage ()
Added in API level 1
Returns the language code for this Locale or the
empty string if no language was set.
Or:
String language = Locale.getDefault().getDisplayLanguage();
if (language.equals("English"))
{
// Use English link
}
else if (...) { ....
public final String getDisplayLanguage ()
Added in API level 1 Equivalent to
getDisplayLanguage(Locale.getDefault()).
Here are other possibilities:
Locale.getDefault().getLanguage() ---> en
Locale.getDefault().getISO3Language() ---> eng
Locale.getDefault().getCountry() ---> US
Locale.getDefault().getISO3Country() ---> USA
Locale.getDefault().getDisplayCountry() ---> United States
Locale.getDefault().getDisplayName() ---> English (United States)
Locale.getDefault().toString() ---> en_US
Locale.getDefault().getDisplayLanguage()---> English
Documentation here.

Locale.getDefault returns a Locale, not a String. If you want to know more about that locale you call getLanguage and getCountry on the Locale. Look up ISO code for language and country.
Usually, in android you solve something like this by different resources. Under res/values-en/ and res/values-de/ you can define language specific strings.

Localization of code generated natural language strings

I'm writing an app in which I want to display a string, describing an object in my model, to the user. I've started thinking about localization and I'm a bit stumped as to the best approach to deal with the difference in grammar between languages.
Imagine if in my model I have a verb an object and a quantity (pseudocode):
int _quantity = 6;
String _object = "#string/object";
String _verb = "#string/verb";
This should render in English as _verb + _quantity + _object (i.e. "eat 6 eggs", whereas in German it should render as _quantity + _object + _verb (i.e. "6 eggs eat" (but, obviously, using German word tokens rather than English :) ))
Is there a standard approach to this type of problem in Android?
Thanks.

OK, I've come up with a solution which, based on what I know, is the best I can find. If anyone has a better solution, please post!
1) In strings.xml, define any bits of text needed (for me, these equated to the localised names for enumeration members and for the string used to build natural language strings (more on this in a bit))
Example:
<resources>
...
<string name="verb_ate">ate</string>
<string name="verb_threw">threw</string>
...
<string name="modelobject_naturallanguagedescription">I {0} {1} {2}</string>
...
</resources>
2) In any enumerations, decorate them with the resource ids of the strings they are to use for display
Example:
public enum VerbType
Ate(R.string.verb_ate),
Threw(R.string.verb_threw);
private final int _stringID;
private VerbType(int stringID) {
_stringID = stringID;
}
public int getStringID() {
return _stringID;
}
}
3) Make a helper function which will take the model object and the context and will piece together the required natural language string (using java.text.MessageFormat), plus a helper function for each enumeration:
Example (assume that ModelObject has _verb, _quantity, _object and relevant getters):
public static String getVerb(VerbType v, Context c) {
return c.getResources().getString(v.getStringID());
}
public static String getNaturalLanguageString(ModelObject o, Context c) {
MessageFormat mf = new MessageFormat(c.getResources().getString(R.string.modelobject_naturallanguagedescription);
return mf.format(new Object[] {getVerb(o.getVerb(),c), o.getQuantity(), o.getObject()});
}
So, having done all of this, it's easy enough to define another strings file for a different language (German for example (I don't know German, so the language won't be correct)):
<resources>
...
<string name="verb_ate">gegessen</string>
<string name="verb_threw">gethrown</string>
...
<string name="modelobject_naturallanguagedescription">Ich haben {1} {2} {0}</string>
...
</resources>
So, in an english locale getNaturalLanguageString might return:
I ate 6 eggs
whereas in a German locale it might return:
Ich haben 6 eggs gegessen

android tts, change languages

I’m developing an application that uses TTS for Android.
I don’t seem to get something in setting the language for TTS.
I have been able to set my application to speak in Italian, but not in Finnish or Russian. However, both those are shown when I output this:
Log.i("-------------",Arrays.toString(Locale.getAvailableLocales()));
Which gives:
[ar, ar_EG, bg, bg_BG, ca, ca_ES, cs, cs_CZ, da, da_DK, de, de_AT, de_BE, de_CH, de_DE, de_LI, de_LU, el, el_CY, el_GR, en, en_AU, en_BE, en_BW, en_BZ, en_CA, en_GB, en_HK, en_IE, en_IN, en_JM, en_MH, en_MT, en_NA, en_NZ, en_PH, en_PK, en_RH, en_SG, en_TT, en_US, en_US_POSIX, en_VI, en_ZA, en_ZW, es, es_AR, es_BO, es_CL, es_CO, es_CR, es_DO, es_EC, es_ES, es_GT, es_HN, es_MX, es_NI, es_PA, es_PE, es_PR, es_PY, es_SV, es_US, es_UY, es_VE, et, et_EE, eu, eu_ES, fa, fa_IR, **fi, fi_FI**, fr, fr_BE, fr_CA, fr_CH, fr_FR, fr_LU, fr_MC, gl, gl_ES, hr, hr_HR, hu, hu_HU, in, in_ID, is, is_IS, it, it_CH, it_IT, iw, iw_IL, ja, ja_JP, kk, kk_KZ, ko, ko_KR, lt, lt_LT, lv, lv_LV, mk, mk_MK, ms, ms_BN, ms_MY, nl, nl_BE, nl_NL, no, no_NO, no_NO_NY, pl, pl_PL, pt, pt_BR, pt_PT, ro, ro_RO, **ru, ru_RU**, ru_UA, sh, sh_BA, sh_CS, sh_YU, sk, sk_SK, sl, sl_SI, sq, sq_AL, sr, sr_BA, sr_ME, sr_RS, sv, sv_FI, sv_SE, th, th_TH, tr, tr_TR, uk, uk_UA, vi, vi_VN, zh, zh_CN, zh_HK, zh_HANS_SG, zh_HANT_MO, zh_MO, zh_TW]
So, what’s the problem?

Locale.getAvailableLocales() returns the locale codes understood by the device, which has no relation to the locales supported by the current text-to-speech engine.
After creating a connection to a TTS engine, you can call TextToSpeech.isLanguageAvailable(Locale) to determine if a language is supported by the engine and whether data for that particular language is installed.

Get available locales for text to speech (TTS)

I'm working on a text-to-speech implementation of a flashcard program. Text in different languages should be read out. In order to do this properly the user has to select the language of the text to read (will be stored and used later without question).
Is there a possibility of getting the available TTS languages on an Android system? If not, is there a possibility of getting all availably locales on the system?
I guess, I got it: getAvailableLocales() and tts.isLocaleAvailable(locale)

Someone else has done the hard work, at http://kaviddiss.com/2012/08/12/android-text-to-speech-languages/
To save you time, here's their code extract
TextToSpeech tts = ...
// let's assume tts is already inited at this point:
Locale[] locales = Locale.getAvailableLocales();
List<Locale> localeList = new ArrayList<Locale>();
for (Locale locale : locales) {
int res = tts.isLanguageAvailable(locale);
if (res == TextToSpeech.LANG_COUNTRY_AVAILABLE) {
localeList.add(locale);
}
}
// at this point the localeList object will contain
// all available languages for Text to Speech
The results depend on which TTS engine has been selected. For instance, one of my phones includes both the Pico-TTS and Google-text-to-speech engines.
Q-Smart (Vietnamese Phone with Google TTS as selected engine)
D/SpeakRepeatedly( 3979): Engine Google Text-to-speech Engine:com.google.android.tts
D/SpeakRepeatedly( 3979): Engine Pico TTS:com.svox.pico
D/SpeakRepeatedly( 3979): German (Germany):German:de_DE
D/SpeakRepeatedly( 3979): English (United Kingdom):English:en_GB
D/SpeakRepeatedly( 3979): English (United States):English:en_US
D/SpeakRepeatedly( 3979): English (United States,Computer):English:en_US_POSIX
D/SpeakRepeatedly( 3979): Spanish (Spain):Spanish:es_ES
D/SpeakRepeatedly( 3979): French (France):French:fr_FR
D/SpeakRepeatedly( 3979): Italian (Italy):Italian:it_IT
D/SpeakRepeatedly( 3979): Portuguese (Brazil):Portuguese:pt_BR
D/SpeakRepeatedly( 3979): Portuguese (Portugal):Portuguese:pt_PT
And with Pico selected
D/SpeakRepeatedly( 4837): Engine Google Text-to-speech Engine:com.google.android.tts
D/SpeakRepeatedly( 4837): Engine Pico TTS:com.svox.pico
D/SpeakRepeatedly( 4837): German (Germany):German:de_DE
D/SpeakRepeatedly( 4837): English (United Kingdom):English:en_GB
D/SpeakRepeatedly( 4837): English (United States):English:en_US
D/SpeakRepeatedly( 4837): English (United States,Computer):English:en_US_POSIX
D/SpeakRepeatedly( 4837): Spanish (Spain):Spanish:es_ES
D/SpeakRepeatedly( 4837): French (France):French:fr_FR
D/SpeakRepeatedly( 4837): Italian (Italy):Italian:it_IT
Note:
Portuguese isn’t listed in the TTS Settings UI. When I select Portuguese programmatically in my app it speaks with a Portuguese accent! FWIW here's my code to select Portuguese (it accepts both Brazilian and Portuguese locales).
if (locale.getDisplayName().startsWith("Portuguese")) {
Log.i(SPEAK_REPEATEDLY, "Setting Locale to: " + locale.toString());
tts.setLanguage(locale);
}
}

Since different TTS engines return different results for isLanguageAvailable, I found that the following solution works best on several common TTS engines.
Please also note that starting with Android Lollipop, there is a simple method in TextToSpeech called getAvailableLanguages that does that easily for you (if the device is running API 21 or later).
You need to call the following methods in the onInit method of your OnInitListener assigned to the TextToSpeech object.
ArrayList<Locale> languages;
TextToSpeech initTTS;
private void initSupportedLanguagesLollipop()
{
Set<Locale> availableLocales = initTTS.getAvailableLanguages();
for (Locale locale : availableLocales)
{
languages.add(locale);
}
}
private void initSupportedLanguagesLegacy()
{
Locale[] allLocales = Locale.getAvailableLocales();
for (Locale locale : allLocales)
{
try
{
int res = initTTS.isLanguageAvailable(locale);
boolean hasVariant = (null != locale.getVariant() && locale.getVariant().length() > 0);
boolean hasCountry = (null != locale.getCountry() && locale.getCountry().length() > 0);
boolean isLocaleSupported =
false == hasVariant && false == hasCountry && res == TextToSpeech.LANG_AVAILABLE ||
false == hasVariant && true == hasCountry && res == TextToSpeech.LANG_COUNTRY_AVAILABLE ||
res == TextToSpeech.LANG_COUNTRY_VAR_AVAILABLE;
Log.d(TAG, "TextToSpeech Engine isLanguageAvailable " + locale + " (supported=" + isLocaleSupported + ",res=" + res + ", country=" + locale.getCountry() + ", variant=" + locale.getVariant() + ")");
if (true == isLocaleSupported)
{
languages.add(locale);
}
}
catch (Exception ex)
{
Log.e(TAG, "Error checking if language is available for TTS (locale=" + locale +"): " + ex.getClass().getSimpleName() + "-" + ex.getMessage());
}
}
}

Find all available TTS Locale on the device using following function.
Locale.getAvailableLocales()
Output of: Arrays.toString(Locale.getAvailableLocales())
[ar, ar_EG, bg, bg_BG, ca, ca_ES, cs, cs_CZ, da, da_DK, de, de_AT, de_BE, de_CH, de_DE, de_LI, de_LU, el, el_CY, el_GR, en, en
_AU, en_BE, en_BW, en_BZ, en_CA, en_GB, en_HK, en_IE, en_IN, en_JM, en_MH, en_MT, en_NA, en_NZ, en_PH, en_PK, en_RH, en_SG, en_TT, en_US, en_US_POSIX,
en_VI, en_ZA, en_ZW, es, es_AR, es_BO, es_CL, es_CO, es_CR, es_DO, es_EC, es_ES, es_GT, es_HN, es_MX, es_NI, es_PA, es_PE, es_PR, es_PY, es_SV, es_US
, es_UY, es_VE, et, et_EE, eu, eu_ES, fa, fa_IR, fi, fi_FI, fr, fr_BE, fr_CA, fr_CH, fr_FR, fr_LU, fr_MC, gl, gl_ES, hr, hr_HR, hu, hu_HU, in, in_ID,
is, is_IS, it, it_CH, it_IT, iw, iw_IL, ja, ja_JP, kk, kk_KZ, ko, ko_KR, lt, lt_LT, lv, lv_LV, mk, mk_MK, ms, ms_BN, ms_MY, nl, nl_BE, nl_NL, no, no_N
O, no_NO_NY, pl, pl_PL, pt, pt_BR, pt_PT, ro, ro_RO, ru, ru_RU, ru_UA, sh, sh_BA, sh_CS, sh_YU, sk, sk_SK, sl, sl_SI, sq, sq_AL, sr, sr_BA, sr_ME, sr_
RS, sv, sv_FI, sv_SE, th, th_TH, tr, tr_TR, uk, uk_UA, vi, vi_VN, zh, zh_CN, zh_HK, zh_HANS_SG, zh_HANT_MO, zh_MO, zh_TW]

Starting from Android 5.0 (API level 21), TextToSpeech.getAvailableLanguages has been added to fetch a set of all locales supported by the TTS engine.
TextToSpeech tts; // assume this is initialized
tts.getAvailableLanguages(); // returns a set of available locales
I have also noticed that the set of locales returned by TextToSpeech.getAvailableLanguages might not be a strict subset of Locale.getAvailableLocales, i.e. there might a locale supported by the TTS engine that isn't supported by the system.