Home | History | Annotate | Download | only in phonenumbers
      1 /*
      2  * Copyright (C) 2009 The Libphonenumber Authors
      3  *
      4  * Licensed under the Apache License, Version 2.0 (the "License");
      5  * you may not use this file except in compliance with the License.
      6  * You may obtain a copy of the License at
      7  *
      8  * http://www.apache.org/licenses/LICENSE-2.0
      9  *
     10  * Unless required by applicable law or agreed to in writing, software
     11  * distributed under the License is distributed on an "AS IS" BASIS,
     12  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
     13  * See the License for the specific language governing permissions and
     14  * limitations under the License.
     15  */
     16 
     17 package com.android.i18n.phonenumbers;
     18 
     19 import com.android.i18n.phonenumbers.Phonemetadata.NumberFormat;
     20 import com.android.i18n.phonenumbers.Phonemetadata.PhoneMetadata;
     21 import com.android.i18n.phonenumbers.Phonemetadata.PhoneMetadataCollection;
     22 import com.android.i18n.phonenumbers.Phonemetadata.PhoneNumberDesc;
     23 import com.android.i18n.phonenumbers.Phonenumber.PhoneNumber;
     24 import com.android.i18n.phonenumbers.Phonenumber.PhoneNumber.CountryCodeSource;
     25 
     26 import java.io.IOException;
     27 import java.io.InputStream;
     28 import java.io.ObjectInput;
     29 import java.io.ObjectInputStream;
     30 import java.util.ArrayList;
     31 import java.util.Arrays;
     32 import java.util.Collections;
     33 import java.util.HashMap;
     34 import java.util.HashSet;
     35 import java.util.Iterator;
     36 import java.util.List;
     37 import java.util.Map;
     38 import java.util.Set;
     39 import java.util.logging.Level;
     40 import java.util.logging.Logger;
     41 import java.util.regex.Matcher;
     42 import java.util.regex.Pattern;
     43 
     44 /**
     45  * Utility for international phone numbers. Functionality includes formatting, parsing and
     46  * validation.
     47  *
     48  * <p>If you use this library, and want to be notified about important changes, please sign up to
     49  * our <a href="http://groups.google.com/group/libphonenumber-discuss/about">mailing list</a>.
     50  *
     51  * NOTE: A lot of methods in this class require Region Code strings. These must be provided using
     52  * ISO 3166-1 two-letter country-code format. These should be in upper-case. The list of the codes
     53  * can be found here:
     54  * http://www.iso.org/iso/country_codes/iso_3166_code_lists/country_names_and_code_elements.htm
     55  *
     56  * @author Shaopeng Jia
     57  */
     58 public class PhoneNumberUtil {
     59   // @VisibleForTesting
     60   static final MetadataLoader DEFAULT_METADATA_LOADER = new MetadataLoader() {
     61     public InputStream loadMetadata(String metadataFileName) {
     62       return PhoneNumberUtil.class.getResourceAsStream(metadataFileName);
     63     }
     64   };
     65 
     66   private static final Logger logger = Logger.getLogger(PhoneNumberUtil.class.getName());
     67 
     68   /** Flags to use when compiling regular expressions for phone numbers. */
     69   static final int REGEX_FLAGS = Pattern.UNICODE_CASE | Pattern.CASE_INSENSITIVE;
     70   // The minimum and maximum length of the national significant number.
     71   private static final int MIN_LENGTH_FOR_NSN = 2;
     72   // The ITU says the maximum length should be 15, but we have found longer numbers in Germany.
     73   static final int MAX_LENGTH_FOR_NSN = 17;
     74   // The maximum length of the country calling code.
     75   static final int MAX_LENGTH_COUNTRY_CODE = 3;
     76   // We don't allow input strings for parsing to be longer than 250 chars. This prevents malicious
     77   // input from overflowing the regular-expression engine.
     78   private static final int MAX_INPUT_STRING_LENGTH = 250;
     79 
     80   private static final String META_DATA_FILE_PREFIX =
     81       "/com/android/i18n/phonenumbers/data/PhoneNumberMetadataProto";
     82 
     83   // Region-code for the unknown region.
     84   private static final String UNKNOWN_REGION = "ZZ";
     85 
     86   private static final int NANPA_COUNTRY_CODE = 1;
     87 
     88   // The prefix that needs to be inserted in front of a Colombian landline number when dialed from
     89   // a mobile phone in Colombia.
     90   private static final String COLOMBIA_MOBILE_TO_FIXED_LINE_PREFIX = "3";
     91 
     92   // Map of country calling codes that use a mobile token before the area code. One example of when
     93   // this is relevant is when determining the length of the national destination code, which should
     94   // be the length of the area code plus the length of the mobile token.
     95   private static final Map<Integer, String> MOBILE_TOKEN_MAPPINGS;
     96 
     97   // The PLUS_SIGN signifies the international prefix.
     98   static final char PLUS_SIGN = '+';
     99 
    100   private static final char STAR_SIGN = '*';
    101 
    102   private static final String RFC3966_EXTN_PREFIX = ";ext=";
    103   private static final String RFC3966_PREFIX = "tel:";
    104   private static final String RFC3966_PHONE_CONTEXT = ";phone-context=";
    105   private static final String RFC3966_ISDN_SUBADDRESS = ";isub=";
    106 
    107   // A map that contains characters that are essential when dialling. That means any of the
    108   // characters in this map must not be removed from a number when dialling, otherwise the call
    109   // will not reach the intended destination.
    110   private static final Map<Character, Character> DIALLABLE_CHAR_MAPPINGS;
    111 
    112   // Only upper-case variants of alpha characters are stored.
    113   private static final Map<Character, Character> ALPHA_MAPPINGS;
    114 
    115   // For performance reasons, amalgamate both into one map.
    116   private static final Map<Character, Character> ALPHA_PHONE_MAPPINGS;
    117 
    118   // Separate map of all symbols that we wish to retain when formatting alpha numbers. This
    119   // includes digits, ASCII letters and number grouping symbols such as "-" and " ".
    120   private static final Map<Character, Character> ALL_PLUS_NUMBER_GROUPING_SYMBOLS;
    121 
    122   static {
    123     HashMap<Integer, String> mobileTokenMap = new HashMap<Integer, String>();
    124     mobileTokenMap.put(52, "1");
    125     mobileTokenMap.put(54, "9");
    126     MOBILE_TOKEN_MAPPINGS = Collections.unmodifiableMap(mobileTokenMap);
    127 
    128     // Simple ASCII digits map used to populate ALPHA_PHONE_MAPPINGS and
    129     // ALL_PLUS_NUMBER_GROUPING_SYMBOLS.
    130     HashMap<Character, Character> asciiDigitMappings = new HashMap<Character, Character>();
    131     asciiDigitMappings.put('0', '0');
    132     asciiDigitMappings.put('1', '1');
    133     asciiDigitMappings.put('2', '2');
    134     asciiDigitMappings.put('3', '3');
    135     asciiDigitMappings.put('4', '4');
    136     asciiDigitMappings.put('5', '5');
    137     asciiDigitMappings.put('6', '6');
    138     asciiDigitMappings.put('7', '7');
    139     asciiDigitMappings.put('8', '8');
    140     asciiDigitMappings.put('9', '9');
    141 
    142     HashMap<Character, Character> alphaMap = new HashMap<Character, Character>(40);
    143     alphaMap.put('A', '2');
    144     alphaMap.put('B', '2');
    145     alphaMap.put('C', '2');
    146     alphaMap.put('D', '3');
    147     alphaMap.put('E', '3');
    148     alphaMap.put('F', '3');
    149     alphaMap.put('G', '4');
    150     alphaMap.put('H', '4');
    151     alphaMap.put('I', '4');
    152     alphaMap.put('J', '5');
    153     alphaMap.put('K', '5');
    154     alphaMap.put('L', '5');
    155     alphaMap.put('M', '6');
    156     alphaMap.put('N', '6');
    157     alphaMap.put('O', '6');
    158     alphaMap.put('P', '7');
    159     alphaMap.put('Q', '7');
    160     alphaMap.put('R', '7');
    161     alphaMap.put('S', '7');
    162     alphaMap.put('T', '8');
    163     alphaMap.put('U', '8');
    164     alphaMap.put('V', '8');
    165     alphaMap.put('W', '9');
    166     alphaMap.put('X', '9');
    167     alphaMap.put('Y', '9');
    168     alphaMap.put('Z', '9');
    169     ALPHA_MAPPINGS = Collections.unmodifiableMap(alphaMap);
    170 
    171     HashMap<Character, Character> combinedMap = new HashMap<Character, Character>(100);
    172     combinedMap.putAll(ALPHA_MAPPINGS);
    173     combinedMap.putAll(asciiDigitMappings);
    174     ALPHA_PHONE_MAPPINGS = Collections.unmodifiableMap(combinedMap);
    175 
    176     HashMap<Character, Character> diallableCharMap = new HashMap<Character, Character>();
    177     diallableCharMap.putAll(asciiDigitMappings);
    178     diallableCharMap.put(PLUS_SIGN, PLUS_SIGN);
    179     diallableCharMap.put('*', '*');
    180     DIALLABLE_CHAR_MAPPINGS = Collections.unmodifiableMap(diallableCharMap);
    181 
    182     HashMap<Character, Character> allPlusNumberGroupings = new HashMap<Character, Character>();
    183     // Put (lower letter -> upper letter) and (upper letter -> upper letter) mappings.
    184     for (char c : ALPHA_MAPPINGS.keySet()) {
    185       allPlusNumberGroupings.put(Character.toLowerCase(c), c);
    186       allPlusNumberGroupings.put(c, c);
    187     }
    188     allPlusNumberGroupings.putAll(asciiDigitMappings);
    189     // Put grouping symbols.
    190     allPlusNumberGroupings.put('-', '-');
    191     allPlusNumberGroupings.put('\uFF0D', '-');
    192     allPlusNumberGroupings.put('\u2010', '-');
    193     allPlusNumberGroupings.put('\u2011', '-');
    194     allPlusNumberGroupings.put('\u2012', '-');
    195     allPlusNumberGroupings.put('\u2013', '-');
    196     allPlusNumberGroupings.put('\u2014', '-');
    197     allPlusNumberGroupings.put('\u2015', '-');
    198     allPlusNumberGroupings.put('\u2212', '-');
    199     allPlusNumberGroupings.put('/', '/');
    200     allPlusNumberGroupings.put('\uFF0F', '/');
    201     allPlusNumberGroupings.put(' ', ' ');
    202     allPlusNumberGroupings.put('\u3000', ' ');
    203     allPlusNumberGroupings.put('\u2060', ' ');
    204     allPlusNumberGroupings.put('.', '.');
    205     allPlusNumberGroupings.put('\uFF0E', '.');
    206     ALL_PLUS_NUMBER_GROUPING_SYMBOLS = Collections.unmodifiableMap(allPlusNumberGroupings);
    207   }
    208 
    209   // Pattern that makes it easy to distinguish whether a region has a unique international dialing
    210   // prefix or not. If a region has a unique international prefix (e.g. 011 in USA), it will be
    211   // represented as a string that contains a sequence of ASCII digits. If there are multiple
    212   // available international prefixes in a region, they will be represented as a regex string that
    213   // always contains character(s) other than ASCII digits.
    214   // Note this regex also includes tilde, which signals waiting for the tone.
    215   private static final Pattern UNIQUE_INTERNATIONAL_PREFIX =
    216       Pattern.compile("[\\d]+(?:[~\u2053\u223C\uFF5E][\\d]+)?");
    217 
    218   // Regular expression of acceptable punctuation found in phone numbers. This excludes punctuation
    219   // found as a leading character only.
    220   // This consists of dash characters, white space characters, full stops, slashes,
    221   // square brackets, parentheses and tildes. It also includes the letter 'x' as that is found as a
    222   // placeholder for carrier information in some phone numbers. Full-width variants are also
    223   // present.
    224   static final String VALID_PUNCTUATION = "-x\u2010-\u2015\u2212\u30FC\uFF0D-\uFF0F " +
    225       "\u00A0\u00AD\u200B\u2060\u3000()\uFF08\uFF09\uFF3B\uFF3D.\\[\\]/~\u2053\u223C\uFF5E";
    226 
    227   private static final String DIGITS = "\\p{Nd}";
    228   // We accept alpha characters in phone numbers, ASCII only, upper and lower case.
    229   private static final String VALID_ALPHA =
    230       Arrays.toString(ALPHA_MAPPINGS.keySet().toArray()).replaceAll("[, \\[\\]]", "") +
    231       Arrays.toString(ALPHA_MAPPINGS.keySet().toArray()).toLowerCase().replaceAll("[, \\[\\]]", "");
    232   static final String PLUS_CHARS = "+\uFF0B";
    233   static final Pattern PLUS_CHARS_PATTERN = Pattern.compile("[" + PLUS_CHARS + "]+");
    234   private static final Pattern SEPARATOR_PATTERN = Pattern.compile("[" + VALID_PUNCTUATION + "]+");
    235   private static final Pattern CAPTURING_DIGIT_PATTERN = Pattern.compile("(" + DIGITS + ")");
    236 
    237   // Regular expression of acceptable characters that may start a phone number for the purposes of
    238   // parsing. This allows us to strip away meaningless prefixes to phone numbers that may be
    239   // mistakenly given to us. This consists of digits, the plus symbol and arabic-indic digits. This
    240   // does not contain alpha characters, although they may be used later in the number. It also does
    241   // not include other punctuation, as this will be stripped later during parsing and is of no
    242   // information value when parsing a number.
    243   private static final String VALID_START_CHAR = "[" + PLUS_CHARS + DIGITS + "]";
    244   private static final Pattern VALID_START_CHAR_PATTERN = Pattern.compile(VALID_START_CHAR);
    245 
    246   // Regular expression of characters typically used to start a second phone number for the purposes
    247   // of parsing. This allows us to strip off parts of the number that are actually the start of
    248   // another number, such as for: (530) 583-6985 x302/x2303 -> the second extension here makes this
    249   // actually two phone numbers, (530) 583-6985 x302 and (530) 583-6985 x2303. We remove the second
    250   // extension so that the first number is parsed correctly.
    251   private static final String SECOND_NUMBER_START = "[\\\\/] *x";
    252   static final Pattern SECOND_NUMBER_START_PATTERN = Pattern.compile(SECOND_NUMBER_START);
    253 
    254   // Regular expression of trailing characters that we want to remove. We remove all characters that
    255   // are not alpha or numerical characters. The hash character is retained here, as it may signify
    256   // the previous block was an extension.
    257   private static final String UNWANTED_END_CHARS = "[[\\P{N}&&\\P{L}]&&[^#]]+$";
    258   static final Pattern UNWANTED_END_CHAR_PATTERN = Pattern.compile(UNWANTED_END_CHARS);
    259 
    260   // We use this pattern to check if the phone number has at least three letters in it - if so, then
    261   // we treat it as a number where some phone-number digits are represented by letters.
    262   private static final Pattern VALID_ALPHA_PHONE_PATTERN = Pattern.compile("(?:.*?[A-Za-z]){3}.*");
    263 
    264   // Regular expression of viable phone numbers. This is location independent. Checks we have at
    265   // least three leading digits, and only valid punctuation, alpha characters and
    266   // digits in the phone number. Does not include extension data.
    267   // The symbol 'x' is allowed here as valid punctuation since it is often used as a placeholder for
    268   // carrier codes, for example in Brazilian phone numbers. We also allow multiple "+" characters at
    269   // the start.
    270   // Corresponds to the following:
    271   // [digits]{minLengthNsn}|
    272   // plus_sign*(([punctuation]|[star])*[digits]){3,}([punctuation]|[star]|[digits]|[alpha])*
    273   //
    274   // The first reg-ex is to allow short numbers (two digits long) to be parsed if they are entered
    275   // as "15" etc, but only if there is no punctuation in them. The second expression restricts the
    276   // number of digits to three or more, but then allows them to be in international form, and to
    277   // have alpha-characters and punctuation.
    278   //
    279   // Note VALID_PUNCTUATION starts with a -, so must be the first in the range.
    280   private static final String VALID_PHONE_NUMBER =
    281       DIGITS + "{" + MIN_LENGTH_FOR_NSN + "}" + "|" +
    282       "[" + PLUS_CHARS + "]*+(?:[" + VALID_PUNCTUATION + STAR_SIGN + "]*" + DIGITS + "){3,}[" +
    283       VALID_PUNCTUATION + STAR_SIGN + VALID_ALPHA + DIGITS + "]*";
    284 
    285   // Default extension prefix to use when formatting. This will be put in front of any extension
    286   // component of the number, after the main national number is formatted. For example, if you wish
    287   // the default extension formatting to be " extn: 3456", then you should specify " extn: " here
    288   // as the default extension prefix. This can be overridden by region-specific preferences.
    289   private static final String DEFAULT_EXTN_PREFIX = " ext. ";
    290 
    291   // Pattern to capture digits used in an extension. Places a maximum length of "7" for an
    292   // extension.
    293   private static final String CAPTURING_EXTN_DIGITS = "(" + DIGITS + "{1,7})";
    294   // Regexp of all possible ways to write extensions, for use when parsing. This will be run as a
    295   // case-insensitive regexp match. Wide character versions are also provided after each ASCII
    296   // version.
    297   private static final String EXTN_PATTERNS_FOR_PARSING;
    298   static final String EXTN_PATTERNS_FOR_MATCHING;
    299   static {
    300     // One-character symbols that can be used to indicate an extension.
    301     String singleExtnSymbolsForMatching = "x\uFF58#\uFF03~\uFF5E";
    302     // For parsing, we are slightly more lenient in our interpretation than for matching. Here we
    303     // allow a "comma" as a possible extension indicator. When matching, this is hardly ever used to
    304     // indicate this.
    305     String singleExtnSymbolsForParsing = "," + singleExtnSymbolsForMatching;
    306 
    307     EXTN_PATTERNS_FOR_PARSING = createExtnPattern(singleExtnSymbolsForParsing);
    308     EXTN_PATTERNS_FOR_MATCHING = createExtnPattern(singleExtnSymbolsForMatching);
    309   }
    310 
    311   /**
    312    * Helper initialiser method to create the regular-expression pattern to match extensions,
    313    * allowing the one-char extension symbols provided by {@code singleExtnSymbols}.
    314    */
    315   private static String createExtnPattern(String singleExtnSymbols) {
    316     // There are three regular expressions here. The first covers RFC 3966 format, where the
    317     // extension is added using ";ext=". The second more generic one starts with optional white
    318     // space and ends with an optional full stop (.), followed by zero or more spaces/tabs and then
    319     // the numbers themselves. The other one covers the special case of American numbers where the
    320     // extension is written with a hash at the end, such as "- 503#".
    321     // Note that the only capturing groups should be around the digits that you want to capture as
    322     // part of the extension, or else parsing will fail!
    323     // Canonical-equivalence doesn't seem to be an option with Android java, so we allow two options
    324     // for representing the accented o - the character itself, and one in the unicode decomposed
    325     // form with the combining acute accent.
    326     return (RFC3966_EXTN_PREFIX + CAPTURING_EXTN_DIGITS + "|" + "[ \u00A0\\t,]*" +
    327             "(?:e?xt(?:ensi(?:o\u0301?|\u00F3))?n?|\uFF45?\uFF58\uFF54\uFF4E?|" +
    328             "[" + singleExtnSymbols + "]|int|anexo|\uFF49\uFF4E\uFF54)" +
    329             "[:\\.\uFF0E]?[ \u00A0\\t,-]*" + CAPTURING_EXTN_DIGITS + "#?|" +
    330             "[- ]+(" + DIGITS + "{1,5})#");
    331   }
    332 
    333   // Regexp of all known extension prefixes used by different regions followed by 1 or more valid
    334   // digits, for use when parsing.
    335   private static final Pattern EXTN_PATTERN =
    336       Pattern.compile("(?:" + EXTN_PATTERNS_FOR_PARSING + ")$", REGEX_FLAGS);
    337 
    338   // We append optionally the extension pattern to the end here, as a valid phone number may
    339   // have an extension prefix appended, followed by 1 or more digits.
    340   private static final Pattern VALID_PHONE_NUMBER_PATTERN =
    341       Pattern.compile(VALID_PHONE_NUMBER + "(?:" + EXTN_PATTERNS_FOR_PARSING + ")?", REGEX_FLAGS);
    342 
    343   static final Pattern NON_DIGITS_PATTERN = Pattern.compile("(\\D+)");
    344 
    345   // The FIRST_GROUP_PATTERN was originally set to $1 but there are some countries for which the
    346   // first group is not used in the national pattern (e.g. Argentina) so the $1 group does not match
    347   // correctly.  Therefore, we use \d, so that the first group actually used in the pattern will be
    348   // matched.
    349   private static final Pattern FIRST_GROUP_PATTERN = Pattern.compile("(\\$\\d)");
    350   private static final Pattern NP_PATTERN = Pattern.compile("\\$NP");
    351   private static final Pattern FG_PATTERN = Pattern.compile("\\$FG");
    352   private static final Pattern CC_PATTERN = Pattern.compile("\\$CC");
    353 
    354   // A pattern that is used to determine if the national prefix formatting rule has the first group
    355   // only, i.e., does not start with the national prefix. Note that the pattern explicitly allows
    356   // for unbalanced parentheses.
    357   private static final Pattern FIRST_GROUP_ONLY_PREFIX_PATTERN = Pattern.compile("\\(?\\$1\\)?");
    358 
    359   private static PhoneNumberUtil instance = null;
    360 
    361   public static final String REGION_CODE_FOR_NON_GEO_ENTITY = "001";
    362 
    363   /**
    364    * INTERNATIONAL and NATIONAL formats are consistent with the definition in ITU-T Recommendation
    365    * E123. For example, the number of the Google Switzerland office will be written as
    366    * "+41 44 668 1800" in INTERNATIONAL format, and as "044 668 1800" in NATIONAL format.
    367    * E164 format is as per INTERNATIONAL format but with no formatting applied, e.g.
    368    * "+41446681800". RFC3966 is as per INTERNATIONAL format, but with all spaces and other
    369    * separating symbols replaced with a hyphen, and with any phone number extension appended with
    370    * ";ext=". It also will have a prefix of "tel:" added, e.g. "tel:+41-44-668-1800".
    371    *
    372    * Note: If you are considering storing the number in a neutral format, you are highly advised to
    373    * use the PhoneNumber class.
    374    */
    375   public enum PhoneNumberFormat {
    376     E164,
    377     INTERNATIONAL,
    378     NATIONAL,
    379     RFC3966
    380   }
    381 
    382   /**
    383    * Type of phone numbers.
    384    */
    385   public enum PhoneNumberType {
    386     FIXED_LINE,
    387     MOBILE,
    388     // In some regions (e.g. the USA), it is impossible to distinguish between fixed-line and
    389     // mobile numbers by looking at the phone number itself.
    390     FIXED_LINE_OR_MOBILE,
    391     // Freephone lines
    392     TOLL_FREE,
    393     PREMIUM_RATE,
    394     // The cost of this call is shared between the caller and the recipient, and is hence typically
    395     // less than PREMIUM_RATE calls. See // http://en.wikipedia.org/wiki/Shared_Cost_Service for
    396     // more information.
    397     SHARED_COST,
    398     // Voice over IP numbers. This includes TSoIP (Telephony Service over IP).
    399     VOIP,
    400     // A personal number is associated with a particular person, and may be routed to either a
    401     // MOBILE or FIXED_LINE number. Some more information can be found here:
    402     // http://en.wikipedia.org/wiki/Personal_Numbers
    403     PERSONAL_NUMBER,
    404     PAGER,
    405     // Used for "Universal Access Numbers" or "Company Numbers". They may be further routed to
    406     // specific offices, but allow one number to be used for a company.
    407     UAN,
    408     // Used for "Voice Mail Access Numbers".
    409     VOICEMAIL,
    410     // A phone number is of type UNKNOWN when it does not fit any of the known patterns for a
    411     // specific region.
    412     UNKNOWN
    413   }
    414 
    415   /**
    416    * Types of phone number matches. See detailed description beside the isNumberMatch() method.
    417    */
    418   public enum MatchType {
    419     NOT_A_NUMBER,
    420     NO_MATCH,
    421     SHORT_NSN_MATCH,
    422     NSN_MATCH,
    423     EXACT_MATCH,
    424   }
    425 
    426   /**
    427    * Possible outcomes when testing if a PhoneNumber is possible.
    428    */
    429   public enum ValidationResult {
    430     IS_POSSIBLE,
    431     INVALID_COUNTRY_CODE,
    432     TOO_SHORT,
    433     TOO_LONG,
    434   }
    435 
    436   /**
    437    * Leniency when {@linkplain PhoneNumberUtil#findNumbers finding} potential phone numbers in text
    438    * segments. The levels here are ordered in increasing strictness.
    439    */
    440   public enum Leniency {
    441     /**
    442      * Phone numbers accepted are {@linkplain PhoneNumberUtil#isPossibleNumber(PhoneNumber)
    443      * possible}, but not necessarily {@linkplain PhoneNumberUtil#isValidNumber(PhoneNumber) valid}.
    444      */
    445     POSSIBLE {
    446       @Override
    447       boolean verify(PhoneNumber number, String candidate, PhoneNumberUtil util) {
    448         return util.isPossibleNumber(number);
    449       }
    450     },
    451     /**
    452      * Phone numbers accepted are {@linkplain PhoneNumberUtil#isPossibleNumber(PhoneNumber)
    453      * possible} and {@linkplain PhoneNumberUtil#isValidNumber(PhoneNumber) valid}. Numbers written
    454      * in national format must have their national-prefix present if it is usually written for a
    455      * number of this type.
    456      */
    457     VALID {
    458       @Override
    459       boolean verify(PhoneNumber number, String candidate, PhoneNumberUtil util) {
    460         if (!util.isValidNumber(number) ||
    461             !PhoneNumberMatcher.containsOnlyValidXChars(number, candidate, util)) {
    462           return false;
    463         }
    464         return PhoneNumberMatcher.isNationalPrefixPresentIfRequired(number, util);
    465       }
    466     },
    467     /**
    468      * Phone numbers accepted are {@linkplain PhoneNumberUtil#isValidNumber(PhoneNumber) valid} and
    469      * are grouped in a possible way for this locale. For example, a US number written as
    470      * "65 02 53 00 00" and "650253 0000" are not accepted at this leniency level, whereas
    471      * "650 253 0000", "650 2530000" or "6502530000" are.
    472      * Numbers with more than one '/' symbol in the national significant number are also dropped at
    473      * this level.
    474      * <p>
    475      * Warning: This level might result in lower coverage especially for regions outside of country
    476      * code "+1". If you are not sure about which level to use, email the discussion group
    477      * libphonenumber-discuss (at) googlegroups.com.
    478      */
    479     STRICT_GROUPING {
    480       @Override
    481       boolean verify(PhoneNumber number, String candidate, PhoneNumberUtil util) {
    482         if (!util.isValidNumber(number) ||
    483             !PhoneNumberMatcher.containsOnlyValidXChars(number, candidate, util) ||
    484             PhoneNumberMatcher.containsMoreThanOneSlashInNationalNumber(number, candidate) ||
    485             !PhoneNumberMatcher.isNationalPrefixPresentIfRequired(number, util)) {
    486           return false;
    487         }
    488         return PhoneNumberMatcher.checkNumberGroupingIsValid(
    489             number, candidate, util, new PhoneNumberMatcher.NumberGroupingChecker() {
    490               public boolean checkGroups(PhoneNumberUtil util, PhoneNumber number,
    491                                          StringBuilder normalizedCandidate,
    492                                          String[] expectedNumberGroups) {
    493                 return PhoneNumberMatcher.allNumberGroupsRemainGrouped(
    494                     util, number, normalizedCandidate, expectedNumberGroups);
    495               }
    496             });
    497       }
    498     },
    499     /**
    500      * Phone numbers accepted are {@linkplain PhoneNumberUtil#isValidNumber(PhoneNumber) valid} and
    501      * are grouped in the same way that we would have formatted it, or as a single block. For
    502      * example, a US number written as "650 2530000" is not accepted at this leniency level, whereas
    503      * "650 253 0000" or "6502530000" are.
    504      * Numbers with more than one '/' symbol are also dropped at this level.
    505      * <p>
    506      * Warning: This level might result in lower coverage especially for regions outside of country
    507      * code "+1". If you are not sure about which level to use, email the discussion group
    508      * libphonenumber-discuss (at) googlegroups.com.
    509      */
    510     EXACT_GROUPING {
    511       @Override
    512       boolean verify(PhoneNumber number, String candidate, PhoneNumberUtil util) {
    513         if (!util.isValidNumber(number) ||
    514             !PhoneNumberMatcher.containsOnlyValidXChars(number, candidate, util) ||
    515             PhoneNumberMatcher.containsMoreThanOneSlashInNationalNumber(number, candidate) ||
    516             !PhoneNumberMatcher.isNationalPrefixPresentIfRequired(number, util)) {
    517           return false;
    518         }
    519         return PhoneNumberMatcher.checkNumberGroupingIsValid(
    520             number, candidate, util, new PhoneNumberMatcher.NumberGroupingChecker() {
    521               public boolean checkGroups(PhoneNumberUtil util, PhoneNumber number,
    522                                          StringBuilder normalizedCandidate,
    523                                          String[] expectedNumberGroups) {
    524                 return PhoneNumberMatcher.allNumberGroupsAreExactlyPresent(
    525                     util, number, normalizedCandidate, expectedNumberGroups);
    526               }
    527             });
    528       }
    529     };
    530 
    531     /** Returns true if {@code number} is a verified number according to this leniency. */
    532     abstract boolean verify(PhoneNumber number, String candidate, PhoneNumberUtil util);
    533   }
    534 
    535   // A mapping from a country calling code to the region codes which denote the region represented
    536   // by that country calling code. In the case of multiple regions sharing a calling code, such as
    537   // the NANPA regions, the one indicated with "isMainCountryForCode" in the metadata should be
    538   // first.
    539   private final Map<Integer, List<String>> countryCallingCodeToRegionCodeMap;
    540 
    541   // The set of regions that share country calling code 1.
    542   // There are roughly 26 regions and we set the initial capacity of the HashSet to 35 to offer a
    543   // load factor of roughly 0.75.
    544   private final Set<String> nanpaRegions = new HashSet<String>(35);
    545 
    546   // A mapping from a region code to the PhoneMetadata for that region.
    547   // Note: Synchronization, though only needed for the Android version of the library, is used in
    548   // all versions for consistency.
    549   private final Map<String, PhoneMetadata> regionToMetadataMap =
    550       Collections.synchronizedMap(new HashMap<String, PhoneMetadata>());
    551 
    552   // A mapping from a country calling code for a non-geographical entity to the PhoneMetadata for
    553   // that country calling code. Examples of the country calling codes include 800 (International
    554   // Toll Free Service) and 808 (International Shared Cost Service).
    555   // Note: Synchronization, though only needed for the Android version of the library, is used in
    556   // all versions for consistency.
    557   private final Map<Integer, PhoneMetadata> countryCodeToNonGeographicalMetadataMap =
    558       Collections.synchronizedMap(new HashMap<Integer, PhoneMetadata>());
    559 
    560   // A cache for frequently used region-specific regular expressions.
    561   // The initial capacity is set to 100 as this seems to be an optimal value for Android, based on
    562   // performance measurements.
    563   private final RegexCache regexCache = new RegexCache(100);
    564 
    565   // The set of regions the library supports.
    566   // There are roughly 240 of them and we set the initial capacity of the HashSet to 320 to offer a
    567   // load factor of roughly 0.75.
    568   private final Set<String> supportedRegions = new HashSet<String>(320);
    569 
    570   // The set of county calling codes that map to the non-geo entity region ("001"). This set
    571   // currently contains < 12 elements so the default capacity of 16 (load factor=0.75) is fine.
    572   private final Set<Integer> countryCodesForNonGeographicalRegion = new HashSet<Integer>();
    573 
    574   // The prefix of the metadata files from which region data is loaded.
    575   private final String currentFilePrefix;
    576   // The metadata loader used to inject alternative metadata sources.
    577   private final MetadataLoader metadataLoader;
    578 
    579   /**
    580    * This class implements a singleton, so the only constructor is private.
    581    */
    582   // @VisibleForTesting
    583   PhoneNumberUtil(String filePrefix, MetadataLoader metadataLoader,
    584       Map<Integer, List<String>> countryCallingCodeToRegionCodeMap) {
    585     this.currentFilePrefix = filePrefix;
    586     this.metadataLoader = metadataLoader;
    587     this.countryCallingCodeToRegionCodeMap = countryCallingCodeToRegionCodeMap;
    588     for (Map.Entry<Integer, List<String>> entry : countryCallingCodeToRegionCodeMap.entrySet()) {
    589       List<String> regionCodes = entry.getValue();
    590       // We can assume that if the county calling code maps to the non-geo entity region code then
    591       // that's the only region code it maps to.
    592       if (regionCodes.size() == 1 && REGION_CODE_FOR_NON_GEO_ENTITY.equals(regionCodes.get(0))) {
    593         // This is the subset of all country codes that map to the non-geo entity region code.
    594         countryCodesForNonGeographicalRegion.add(entry.getKey());
    595       } else {
    596         // The supported regions set does not include the "001" non-geo entity region code.
    597         supportedRegions.addAll(regionCodes);
    598       }
    599     }
    600     // If the non-geo entity still got added to the set of supported regions it must be because
    601     // there are entries that list the non-geo entity alongside normal regions (which is wrong).
    602     // If we discover this, remove the non-geo entity from the set of supported regions and log.
    603     if (supportedRegions.remove(REGION_CODE_FOR_NON_GEO_ENTITY)) {
    604       logger.log(Level.WARNING, "invalid metadata " +
    605           "(country calling code was mapped to the non-geo entity as well as specific region(s))");
    606     }
    607     nanpaRegions.addAll(countryCallingCodeToRegionCodeMap.get(NANPA_COUNTRY_CODE));
    608   }
    609 
    610   // @VisibleForTesting
    611   void loadMetadataFromFile(String filePrefix, String regionCode, int countryCallingCode,
    612       MetadataLoader metadataLoader) {
    613     boolean isNonGeoRegion = REGION_CODE_FOR_NON_GEO_ENTITY.equals(regionCode);
    614     String fileName = filePrefix + "_" +
    615         (isNonGeoRegion ? String.valueOf(countryCallingCode) : regionCode);
    616     InputStream source = metadataLoader.loadMetadata(fileName);
    617     if (source == null) {
    618       logger.log(Level.SEVERE, "missing metadata: " + fileName);
    619       throw new IllegalStateException("missing metadata: " + fileName);
    620     }
    621     ObjectInputStream in = null;
    622     try {
    623       in = new ObjectInputStream(source);
    624       PhoneMetadataCollection metadataCollection = loadMetadataAndCloseInput(in);
    625       List<PhoneMetadata> metadataList = metadataCollection.getMetadataList();
    626       if (metadataList.isEmpty()) {
    627         logger.log(Level.SEVERE, "empty metadata: " + fileName);
    628         throw new IllegalStateException("empty metadata: " + fileName);
    629       }
    630       if (metadataList.size() > 1) {
    631         logger.log(Level.WARNING, "invalid metadata (too many entries): " + fileName);
    632       }
    633       PhoneMetadata metadata = metadataList.get(0);
    634       if (isNonGeoRegion) {
    635         countryCodeToNonGeographicalMetadataMap.put(countryCallingCode, metadata);
    636       } else {
    637         regionToMetadataMap.put(regionCode, metadata);
    638       }
    639     } catch (IOException e) {
    640       logger.log(Level.SEVERE, "cannot load/parse metadata: " + fileName, e);
    641       throw new RuntimeException("cannot load/parse metadata: " + fileName, e);
    642     }
    643   }
    644 
    645   /**
    646    * Loads the metadata protocol buffer from the given stream and closes the stream afterwards. Any
    647    * exceptions that occur while reading the stream are propagated (though exceptions that occur
    648    * when the stream is closed will be ignored).
    649    *
    650    * @param source  the non-null stream from which metadata is to be read.
    651    * @return        the loaded metadata protocol buffer.
    652    */
    653   private static PhoneMetadataCollection loadMetadataAndCloseInput(ObjectInputStream source) {
    654     PhoneMetadataCollection metadataCollection = new PhoneMetadataCollection();
    655     try {
    656       metadataCollection.readExternal(source);
    657     } catch (IOException e) {
    658       logger.log(Level.WARNING, "error reading input (ignored)", e);
    659     } finally {
    660       try {
    661         source.close();
    662       } catch (IOException e) {
    663         logger.log(Level.WARNING, "error closing input stream (ignored)", e);
    664       } finally {
    665         return metadataCollection;
    666       }
    667     }
    668   }
    669 
    670   /**
    671    * Attempts to extract a possible number from the string passed in. This currently strips all
    672    * leading characters that cannot be used to start a phone number. Characters that can be used to
    673    * start a phone number are defined in the VALID_START_CHAR_PATTERN. If none of these characters
    674    * are found in the number passed in, an empty string is returned. This function also attempts to
    675    * strip off any alternative extensions or endings if two or more are present, such as in the case
    676    * of: (530) 583-6985 x302/x2303. The second extension here makes this actually two phone numbers,
    677    * (530) 583-6985 x302 and (530) 583-6985 x2303. We remove the second extension so that the first
    678    * number is parsed correctly.
    679    *
    680    * @param number  the string that might contain a phone number
    681    * @return        the number, stripped of any non-phone-number prefix (such as "Tel:") or an empty
    682    *                string if no character used to start phone numbers (such as + or any digit) is
    683    *                found in the number
    684    */
    685   static String extractPossibleNumber(String number) {
    686     Matcher m = VALID_START_CHAR_PATTERN.matcher(number);
    687     if (m.find()) {
    688       number = number.substring(m.start());
    689       // Remove trailing non-alpha non-numerical characters.
    690       Matcher trailingCharsMatcher = UNWANTED_END_CHAR_PATTERN.matcher(number);
    691       if (trailingCharsMatcher.find()) {
    692         number = number.substring(0, trailingCharsMatcher.start());
    693         logger.log(Level.FINER, "Stripped trailing characters: " + number);
    694       }
    695       // Check for extra numbers at the end.
    696       Matcher secondNumber = SECOND_NUMBER_START_PATTERN.matcher(number);
    697       if (secondNumber.find()) {
    698         number = number.substring(0, secondNumber.start());
    699       }
    700       return number;
    701     } else {
    702       return "";
    703     }
    704   }
    705 
    706   /**
    707    * Checks to see if the string of characters could possibly be a phone number at all. At the
    708    * moment, checks to see that the string begins with at least 2 digits, ignoring any punctuation
    709    * commonly found in phone numbers.
    710    * This method does not require the number to be normalized in advance - but does assume that
    711    * leading non-number symbols have been removed, such as by the method extractPossibleNumber.
    712    *
    713    * @param number  string to be checked for viability as a phone number
    714    * @return        true if the number could be a phone number of some sort, otherwise false
    715    */
    716   // @VisibleForTesting
    717   static boolean isViablePhoneNumber(String number) {
    718     if (number.length() < MIN_LENGTH_FOR_NSN) {
    719       return false;
    720     }
    721     Matcher m = VALID_PHONE_NUMBER_PATTERN.matcher(number);
    722     return m.matches();
    723   }
    724 
    725   /**
    726    * Normalizes a string of characters representing a phone number. This performs the following
    727    * conversions:
    728    *   Punctuation is stripped.
    729    *   For ALPHA/VANITY numbers:
    730    *   Letters are converted to their numeric representation on a telephone keypad. The keypad
    731    *       used here is the one defined in ITU Recommendation E.161. This is only done if there are
    732    *       3 or more letters in the number, to lessen the risk that such letters are typos.
    733    *   For other numbers:
    734    *   Wide-ascii digits are converted to normal ASCII (European) digits.
    735    *   Arabic-Indic numerals are converted to European numerals.
    736    *   Spurious alpha characters are stripped.
    737    *
    738    * @param number  a string of characters representing a phone number
    739    * @return        the normalized string version of the phone number
    740    */
    741   static String normalize(String number) {
    742     Matcher m = VALID_ALPHA_PHONE_PATTERN.matcher(number);
    743     if (m.matches()) {
    744       return normalizeHelper(number, ALPHA_PHONE_MAPPINGS, true);
    745     } else {
    746       return normalizeDigitsOnly(number);
    747     }
    748   }
    749 
    750   /**
    751    * Normalizes a string of characters representing a phone number. This is a wrapper for
    752    * normalize(String number) but does in-place normalization of the StringBuilder provided.
    753    *
    754    * @param number  a StringBuilder of characters representing a phone number that will be
    755    *     normalized in place
    756    */
    757   static void normalize(StringBuilder number) {
    758     String normalizedNumber = normalize(number.toString());
    759     number.replace(0, number.length(), normalizedNumber);
    760   }
    761 
    762   /**
    763    * Normalizes a string of characters representing a phone number. This converts wide-ascii and
    764    * arabic-indic numerals to European numerals, and strips punctuation and alpha characters.
    765    *
    766    * @param number  a string of characters representing a phone number
    767    * @return        the normalized string version of the phone number
    768    */
    769   public static String normalizeDigitsOnly(String number) {
    770     return normalizeDigits(number, false /* strip non-digits */).toString();
    771   }
    772 
    773   static StringBuilder normalizeDigits(String number, boolean keepNonDigits) {
    774     StringBuilder normalizedDigits = new StringBuilder(number.length());
    775     for (char c : number.toCharArray()) {
    776       int digit = Character.digit(c, 10);
    777       if (digit != -1) {
    778         normalizedDigits.append(digit);
    779       } else if (keepNonDigits) {
    780         normalizedDigits.append(c);
    781       }
    782     }
    783     return normalizedDigits;
    784   }
    785 
    786   /**
    787    * Normalizes a string of characters representing a phone number. This strips all characters which
    788    * are not diallable on a mobile phone keypad (including all non-ASCII digits).
    789    *
    790    * @param number  a string of characters representing a phone number
    791    * @return        the normalized string version of the phone number
    792    */
    793   static String normalizeDiallableCharsOnly(String number) {
    794     return normalizeHelper(number, DIALLABLE_CHAR_MAPPINGS, true /* remove non matches */);
    795   }
    796 
    797   /**
    798    * Converts all alpha characters in a number to their respective digits on a keypad, but retains
    799    * existing formatting.
    800    */
    801   public static String convertAlphaCharactersInNumber(String number) {
    802     return normalizeHelper(number, ALPHA_PHONE_MAPPINGS, false);
    803   }
    804 
    805   /**
    806    * Gets the length of the geographical area code from the {@code nationalNumber_} field of the
    807    * PhoneNumber object passed in, so that clients could use it to split a national significant
    808    * number into geographical area code and subscriber number. It works in such a way that the
    809    * resultant subscriber number should be diallable, at least on some devices. An example of how
    810    * this could be used:
    811    *
    812    * <pre>
    813    * PhoneNumberUtil phoneUtil = PhoneNumberUtil.getInstance();
    814    * PhoneNumber number = phoneUtil.parse("16502530000", "US");
    815    * String nationalSignificantNumber = phoneUtil.getNationalSignificantNumber(number);
    816    * String areaCode;
    817    * String subscriberNumber;
    818    *
    819    * int areaCodeLength = phoneUtil.getLengthOfGeographicalAreaCode(number);
    820    * if (areaCodeLength > 0) {
    821    *   areaCode = nationalSignificantNumber.substring(0, areaCodeLength);
    822    *   subscriberNumber = nationalSignificantNumber.substring(areaCodeLength);
    823    * } else {
    824    *   areaCode = "";
    825    *   subscriberNumber = nationalSignificantNumber;
    826    * }
    827    * </pre>
    828    *
    829    * N.B.: area code is a very ambiguous concept, so the I18N team generally recommends against
    830    * using it for most purposes, but recommends using the more general {@code national_number}
    831    * instead. Read the following carefully before deciding to use this method:
    832    * <ul>
    833    *  <li> geographical area codes change over time, and this method honors those changes;
    834    *    therefore, it doesn't guarantee the stability of the result it produces.
    835    *  <li> subscriber numbers may not be diallable from all devices (notably mobile devices, which
    836    *    typically requires the full national_number to be dialled in most regions).
    837    *  <li> most non-geographical numbers have no area codes, including numbers from non-geographical
    838    *    entities
    839    *  <li> some geographical numbers have no area codes.
    840    * </ul>
    841    * @param number  the PhoneNumber object for which clients want to know the length of the area
    842    *     code.
    843    * @return  the length of area code of the PhoneNumber object passed in.
    844    */
    845   public int getLengthOfGeographicalAreaCode(PhoneNumber number) {
    846     PhoneMetadata metadata = getMetadataForRegion(getRegionCodeForNumber(number));
    847     if (metadata == null) {
    848       return 0;
    849     }
    850     // If a country doesn't use a national prefix, and this number doesn't have an Italian leading
    851     // zero, we assume it is a closed dialling plan with no area codes.
    852     if (!metadata.hasNationalPrefix() && !number.isItalianLeadingZero()) {
    853       return 0;
    854     }
    855 
    856     if (!isNumberGeographical(number)) {
    857       return 0;
    858     }
    859 
    860     return getLengthOfNationalDestinationCode(number);
    861   }
    862 
    863   /**
    864    * Gets the length of the national destination code (NDC) from the PhoneNumber object passed in,
    865    * so that clients could use it to split a national significant number into NDC and subscriber
    866    * number. The NDC of a phone number is normally the first group of digit(s) right after the
    867    * country calling code when the number is formatted in the international format, if there is a
    868    * subscriber number part that follows. An example of how this could be used:
    869    *
    870    * <pre>
    871    * PhoneNumberUtil phoneUtil = PhoneNumberUtil.getInstance();
    872    * PhoneNumber number = phoneUtil.parse("18002530000", "US");
    873    * String nationalSignificantNumber = phoneUtil.getNationalSignificantNumber(number);
    874    * String nationalDestinationCode;
    875    * String subscriberNumber;
    876    *
    877    * int nationalDestinationCodeLength = phoneUtil.getLengthOfNationalDestinationCode(number);
    878    * if (nationalDestinationCodeLength > 0) {
    879    *   nationalDestinationCode = nationalSignificantNumber.substring(0,
    880    *       nationalDestinationCodeLength);
    881    *   subscriberNumber = nationalSignificantNumber.substring(nationalDestinationCodeLength);
    882    * } else {
    883    *   nationalDestinationCode = "";
    884    *   subscriberNumber = nationalSignificantNumber;
    885    * }
    886    * </pre>
    887    *
    888    * Refer to the unittests to see the difference between this function and
    889    * {@link #getLengthOfGeographicalAreaCode}.
    890    *
    891    * @param number  the PhoneNumber object for which clients want to know the length of the NDC.
    892    * @return  the length of NDC of the PhoneNumber object passed in.
    893    */
    894   public int getLengthOfNationalDestinationCode(PhoneNumber number) {
    895     PhoneNumber copiedProto;
    896     if (number.hasExtension()) {
    897       // We don't want to alter the proto given to us, but we don't want to include the extension
    898       // when we format it, so we copy it and clear the extension here.
    899       copiedProto = new PhoneNumber();
    900       copiedProto.mergeFrom(number);
    901       copiedProto.clearExtension();
    902     } else {
    903       copiedProto = number;
    904     }
    905 
    906     String nationalSignificantNumber = format(copiedProto,
    907                                               PhoneNumberUtil.PhoneNumberFormat.INTERNATIONAL);
    908     String[] numberGroups = NON_DIGITS_PATTERN.split(nationalSignificantNumber);
    909     // The pattern will start with "+COUNTRY_CODE " so the first group will always be the empty
    910     // string (before the + symbol) and the second group will be the country calling code. The third
    911     // group will be area code if it is not the last group.
    912     if (numberGroups.length <= 3) {
    913       return 0;
    914     }
    915 
    916     if (getNumberType(number) == PhoneNumberType.MOBILE) {
    917       // For example Argentinian mobile numbers, when formatted in the international format, are in
    918       // the form of +54 9 NDC XXXX.... As a result, we take the length of the third group (NDC) and
    919       // add the length of the second group (which is the mobile token), which also forms part of
    920       // the national significant number. This assumes that the mobile token is always formatted
    921       // separately from the rest of the phone number.
    922       String mobileToken = getCountryMobileToken(number.getCountryCode());
    923       if (!mobileToken.equals("")) {
    924         return numberGroups[2].length() + numberGroups[3].length();
    925       }
    926     }
    927     return numberGroups[2].length();
    928   }
    929 
    930   /**
    931    * Returns the mobile token for the provided country calling code if it has one, otherwise
    932    * returns an empty string. A mobile token is a number inserted before the area code when dialing
    933    * a mobile number from that country from abroad.
    934    *
    935    * @param countryCallingCode  the country calling code for which we want the mobile token
    936    * @return  the mobile token, as a string, for the given country calling code
    937    */
    938   public static String getCountryMobileToken(int countryCallingCode) {
    939     if (MOBILE_TOKEN_MAPPINGS.containsKey(countryCallingCode)) {
    940       return MOBILE_TOKEN_MAPPINGS.get(countryCallingCode);
    941     }
    942     return "";
    943   }
    944 
    945   /**
    946    * Normalizes a string of characters representing a phone number by replacing all characters found
    947    * in the accompanying map with the values therein, and stripping all other characters if
    948    * removeNonMatches is true.
    949    *
    950    * @param number                     a string of characters representing a phone number
    951    * @param normalizationReplacements  a mapping of characters to what they should be replaced by in
    952    *                                   the normalized version of the phone number
    953    * @param removeNonMatches           indicates whether characters that are not able to be replaced
    954    *                                   should be stripped from the number. If this is false, they
    955    *                                   will be left unchanged in the number.
    956    * @return  the normalized string version of the phone number
    957    */
    958   private static String normalizeHelper(String number,
    959                                         Map<Character, Character> normalizationReplacements,
    960                                         boolean removeNonMatches) {
    961     StringBuilder normalizedNumber = new StringBuilder(number.length());
    962     for (int i = 0; i < number.length(); i++) {
    963       char character = number.charAt(i);
    964       Character newDigit = normalizationReplacements.get(Character.toUpperCase(character));
    965       if (newDigit != null) {
    966         normalizedNumber.append(newDigit);
    967       } else if (!removeNonMatches) {
    968         normalizedNumber.append(character);
    969       }
    970       // If neither of the above are true, we remove this character.
    971     }
    972     return normalizedNumber.toString();
    973   }
    974 
    975   /**
    976    * Sets or resets the PhoneNumberUtil singleton instance. If set to null, the next call to
    977    * {@code getInstance()} will load (and return) the default instance.
    978    */
    979   // @VisibleForTesting
    980   static synchronized void setInstance(PhoneNumberUtil util) {
    981     instance = util;
    982   }
    983 
    984   /**
    985    * Convenience method to get a list of what regions the library has metadata for.
    986    */
    987   public Set<String> getSupportedRegions() {
    988     return Collections.unmodifiableSet(supportedRegions);
    989   }
    990 
    991   /**
    992    * Convenience method to get a list of what global network calling codes the library has metadata
    993    * for.
    994    */
    995   public Set<Integer> getSupportedGlobalNetworkCallingCodes() {
    996     return Collections.unmodifiableSet(countryCodesForNonGeographicalRegion);
    997   }
    998 
    999   /**
   1000    * Gets a {@link PhoneNumberUtil} instance to carry out international phone number formatting,
   1001    * parsing, or validation. The instance is loaded with phone number metadata for a number of most
   1002    * commonly used regions.
   1003    *
   1004    * <p>The {@link PhoneNumberUtil} is implemented as a singleton. Therefore, calling getInstance
   1005    * multiple times will only result in one instance being created.
   1006    *
   1007    * @return a PhoneNumberUtil instance
   1008    */
   1009   public static synchronized PhoneNumberUtil getInstance() {
   1010     if (instance == null) {
   1011       setInstance(createInstance(DEFAULT_METADATA_LOADER));
   1012     }
   1013     return instance;
   1014   }
   1015 
   1016   /**
   1017    * Create a new {@link PhoneNumberUtil} instance to carry out international phone number
   1018    * formatting, parsing, or validation. The instance is loaded with all metadata by
   1019    * using the metadataLoader specified.
   1020    *
   1021    * This method should only be used in the rare case in which you want to manage your own
   1022    * metadata loading. Calling this method multiple times is very expensive, as each time
   1023    * a new instance is created from scratch. When in doubt, use {@link #getInstance}.
   1024    *
   1025    * @param metadataLoader Customized metadata loader. If null, default metadata loader will
   1026    *     be used. This should not be null.
   1027    * @return a PhoneNumberUtil instance
   1028    */
   1029   public static PhoneNumberUtil createInstance(MetadataLoader metadataLoader) {
   1030     if (metadataLoader == null) {
   1031       throw new IllegalArgumentException("metadataLoader could not be null.");
   1032     }
   1033     return new PhoneNumberUtil(META_DATA_FILE_PREFIX, metadataLoader,
   1034         CountryCodeToRegionCodeMap.getCountryCodeToRegionCodeMap());
   1035   }
   1036 
   1037   /**
   1038    * Helper function to check if the national prefix formatting rule has the first group only, i.e.,
   1039    * does not start with the national prefix.
   1040    */
   1041   static boolean formattingRuleHasFirstGroupOnly(String nationalPrefixFormattingRule) {
   1042     return nationalPrefixFormattingRule.length() == 0 ||
   1043         FIRST_GROUP_ONLY_PREFIX_PATTERN.matcher(nationalPrefixFormattingRule).matches();
   1044   }
   1045 
   1046   /**
   1047    * Tests whether a phone number has a geographical association. It checks if the number is
   1048    * associated to a certain region in the country where it belongs to. Note that this doesn't
   1049    * verify if the number is actually in use.
   1050    */
   1051   boolean isNumberGeographical(PhoneNumber phoneNumber) {
   1052     PhoneNumberType numberType = getNumberType(phoneNumber);
   1053     // TODO: Include mobile phone numbers from countries like Indonesia, which has some
   1054     // mobile numbers that are geographical.
   1055     return numberType == PhoneNumberType.FIXED_LINE ||
   1056         numberType == PhoneNumberType.FIXED_LINE_OR_MOBILE;
   1057   }
   1058 
   1059   /**
   1060    * Helper function to check region code is not unknown or null.
   1061    */
   1062   private boolean isValidRegionCode(String regionCode) {
   1063     return regionCode != null && supportedRegions.contains(regionCode);
   1064   }
   1065 
   1066   /**
   1067    * Helper function to check the country calling code is valid.
   1068    */
   1069   private boolean hasValidCountryCallingCode(int countryCallingCode) {
   1070     return countryCallingCodeToRegionCodeMap.containsKey(countryCallingCode);
   1071   }
   1072 
   1073   /**
   1074    * Formats a phone number in the specified format using default rules. Note that this does not
   1075    * promise to produce a phone number that the user can dial from where they are - although we do
   1076    * format in either 'national' or 'international' format depending on what the client asks for, we
   1077    * do not currently support a more abbreviated format, such as for users in the same "area" who
   1078    * could potentially dial the number without area code. Note that if the phone number has a
   1079    * country calling code of 0 or an otherwise invalid country calling code, we cannot work out
   1080    * which formatting rules to apply so we return the national significant number with no formatting
   1081    * applied.
   1082    *
   1083    * @param number         the phone number to be formatted
   1084    * @param numberFormat   the format the phone number should be formatted into
   1085    * @return  the formatted phone number
   1086    */
   1087   public String format(PhoneNumber number, PhoneNumberFormat numberFormat) {
   1088     if (number.getNationalNumber() == 0 && number.hasRawInput()) {
   1089       // Unparseable numbers that kept their raw input just use that.
   1090       // This is the only case where a number can be formatted as E164 without a
   1091       // leading '+' symbol (but the original number wasn't parseable anyway).
   1092       // TODO: Consider removing the 'if' above so that unparseable
   1093       // strings without raw input format to the empty string instead of "+00"
   1094       String rawInput = number.getRawInput();
   1095       if (rawInput.length() > 0) {
   1096         return rawInput;
   1097       }
   1098     }
   1099     StringBuilder formattedNumber = new StringBuilder(20);
   1100     format(number, numberFormat, formattedNumber);
   1101     return formattedNumber.toString();
   1102   }
   1103 
   1104   /**
   1105    * Same as {@link #format(PhoneNumber, PhoneNumberFormat)}, but accepts a mutable StringBuilder as
   1106    * a parameter to decrease object creation when invoked many times.
   1107    */
   1108   public void format(PhoneNumber number, PhoneNumberFormat numberFormat,
   1109                      StringBuilder formattedNumber) {
   1110     // Clear the StringBuilder first.
   1111     formattedNumber.setLength(0);
   1112     int countryCallingCode = number.getCountryCode();
   1113     String nationalSignificantNumber = getNationalSignificantNumber(number);
   1114     if (numberFormat == PhoneNumberFormat.E164) {
   1115       // Early exit for E164 case (even if the country calling code is invalid) since no formatting
   1116       // of the national number needs to be applied. Extensions are not formatted.
   1117       formattedNumber.append(nationalSignificantNumber);
   1118       prefixNumberWithCountryCallingCode(countryCallingCode, PhoneNumberFormat.E164,
   1119                                          formattedNumber);
   1120       return;
   1121     }
   1122     if (!hasValidCountryCallingCode(countryCallingCode)) {
   1123       formattedNumber.append(nationalSignificantNumber);
   1124       return;
   1125     }
   1126     // Note getRegionCodeForCountryCode() is used because formatting information for regions which
   1127     // share a country calling code is contained by only one region for performance reasons. For
   1128     // example, for NANPA regions it will be contained in the metadata for US.
   1129     String regionCode = getRegionCodeForCountryCode(countryCallingCode);
   1130     // Metadata cannot be null because the country calling code is valid (which means that the
   1131     // region code cannot be ZZ and must be one of our supported region codes).
   1132     PhoneMetadata metadata =
   1133         getMetadataForRegionOrCallingCode(countryCallingCode, regionCode);
   1134     formattedNumber.append(formatNsn(nationalSignificantNumber, metadata, numberFormat));
   1135     maybeAppendFormattedExtension(number, metadata, numberFormat, formattedNumber);
   1136     prefixNumberWithCountryCallingCode(countryCallingCode, numberFormat, formattedNumber);
   1137   }
   1138 
   1139   /**
   1140    * Formats a phone number in the specified format using client-defined formatting rules. Note that
   1141    * if the phone number has a country calling code of zero or an otherwise invalid country calling
   1142    * code, we cannot work out things like whether there should be a national prefix applied, or how
   1143    * to format extensions, so we return the national significant number with no formatting applied.
   1144    *
   1145    * @param number                        the phone number to be formatted
   1146    * @param numberFormat                  the format the phone number should be formatted into
   1147    * @param userDefinedFormats            formatting rules specified by clients
   1148    * @return  the formatted phone number
   1149    */
   1150   public String formatByPattern(PhoneNumber number,
   1151                                 PhoneNumberFormat numberFormat,
   1152                                 List<NumberFormat> userDefinedFormats) {
   1153     int countryCallingCode = number.getCountryCode();
   1154     String nationalSignificantNumber = getNationalSignificantNumber(number);
   1155     if (!hasValidCountryCallingCode(countryCallingCode)) {
   1156       return nationalSignificantNumber;
   1157     }
   1158     // Note getRegionCodeForCountryCode() is used because formatting information for regions which
   1159     // share a country calling code is contained by only one region for performance reasons. For
   1160     // example, for NANPA regions it will be contained in the metadata for US.
   1161     String regionCode = getRegionCodeForCountryCode(countryCallingCode);
   1162     // Metadata cannot be null because the country calling code is valid
   1163     PhoneMetadata metadata =
   1164         getMetadataForRegionOrCallingCode(countryCallingCode, regionCode);
   1165 
   1166     StringBuilder formattedNumber = new StringBuilder(20);
   1167 
   1168     NumberFormat formattingPattern =
   1169         chooseFormattingPatternForNumber(userDefinedFormats, nationalSignificantNumber);
   1170     if (formattingPattern == null) {
   1171       // If no pattern above is matched, we format the number as a whole.
   1172       formattedNumber.append(nationalSignificantNumber);
   1173     } else {
   1174       NumberFormat numFormatCopy = new NumberFormat();
   1175       // Before we do a replacement of the national prefix pattern $NP with the national prefix, we
   1176       // need to copy the rule so that subsequent replacements for different numbers have the
   1177       // appropriate national prefix.
   1178       numFormatCopy.mergeFrom(formattingPattern);
   1179       String nationalPrefixFormattingRule = formattingPattern.getNationalPrefixFormattingRule();
   1180       if (nationalPrefixFormattingRule.length() > 0) {
   1181         String nationalPrefix = metadata.getNationalPrefix();
   1182         if (nationalPrefix.length() > 0) {
   1183           // Replace $NP with national prefix and $FG with the first group ($1).
   1184           nationalPrefixFormattingRule =
   1185               NP_PATTERN.matcher(nationalPrefixFormattingRule).replaceFirst(nationalPrefix);
   1186           nationalPrefixFormattingRule =
   1187               FG_PATTERN.matcher(nationalPrefixFormattingRule).replaceFirst("\\$1");
   1188           numFormatCopy.setNationalPrefixFormattingRule(nationalPrefixFormattingRule);
   1189         } else {
   1190           // We don't want to have a rule for how to format the national prefix if there isn't one.
   1191           numFormatCopy.clearNationalPrefixFormattingRule();
   1192         }
   1193       }
   1194       formattedNumber.append(
   1195           formatNsnUsingPattern(nationalSignificantNumber, numFormatCopy, numberFormat));
   1196     }
   1197     maybeAppendFormattedExtension(number, metadata, numberFormat, formattedNumber);
   1198     prefixNumberWithCountryCallingCode(countryCallingCode, numberFormat, formattedNumber);
   1199     return formattedNumber.toString();
   1200   }
   1201 
   1202   /**
   1203    * Formats a phone number in national format for dialing using the carrier as specified in the
   1204    * {@code carrierCode}. The {@code carrierCode} will always be used regardless of whether the
   1205    * phone number already has a preferred domestic carrier code stored. If {@code carrierCode}
   1206    * contains an empty string, returns the number in national format without any carrier code.
   1207    *
   1208    * @param number  the phone number to be formatted
   1209    * @param carrierCode  the carrier selection code to be used
   1210    * @return  the formatted phone number in national format for dialing using the carrier as
   1211    *          specified in the {@code carrierCode}
   1212    */
   1213   public String formatNationalNumberWithCarrierCode(PhoneNumber number, String carrierCode) {
   1214     int countryCallingCode = number.getCountryCode();
   1215     String nationalSignificantNumber = getNationalSignificantNumber(number);
   1216     if (!hasValidCountryCallingCode(countryCallingCode)) {
   1217       return nationalSignificantNumber;
   1218     }
   1219 
   1220     // Note getRegionCodeForCountryCode() is used because formatting information for regions which
   1221     // share a country calling code is contained by only one region for performance reasons. For
   1222     // example, for NANPA regions it will be contained in the metadata for US.
   1223     String regionCode = getRegionCodeForCountryCode(countryCallingCode);
   1224     // Metadata cannot be null because the country calling code is valid.
   1225     PhoneMetadata metadata = getMetadataForRegionOrCallingCode(countryCallingCode, regionCode);
   1226 
   1227     StringBuilder formattedNumber = new StringBuilder(20);
   1228     formattedNumber.append(formatNsn(nationalSignificantNumber, metadata,
   1229                                      PhoneNumberFormat.NATIONAL, carrierCode));
   1230     maybeAppendFormattedExtension(number, metadata, PhoneNumberFormat.NATIONAL, formattedNumber);
   1231     prefixNumberWithCountryCallingCode(countryCallingCode, PhoneNumberFormat.NATIONAL,
   1232                                        formattedNumber);
   1233     return formattedNumber.toString();
   1234   }
   1235 
   1236   private PhoneMetadata getMetadataForRegionOrCallingCode(
   1237       int countryCallingCode, String regionCode) {
   1238     return REGION_CODE_FOR_NON_GEO_ENTITY.equals(regionCode)
   1239         ? getMetadataForNonGeographicalRegion(countryCallingCode)
   1240         : getMetadataForRegion(regionCode);
   1241   }
   1242 
   1243   /**
   1244    * Formats a phone number in national format for dialing using the carrier as specified in the
   1245    * preferredDomesticCarrierCode field of the PhoneNumber object passed in. If that is missing,
   1246    * use the {@code fallbackCarrierCode} passed in instead. If there is no
   1247    * {@code preferredDomesticCarrierCode}, and the {@code fallbackCarrierCode} contains an empty
   1248    * string, return the number in national format without any carrier code.
   1249    *
   1250    * <p>Use {@link #formatNationalNumberWithCarrierCode} instead if the carrier code passed in
   1251    * should take precedence over the number's {@code preferredDomesticCarrierCode} when formatting.
   1252    *
   1253    * @param number  the phone number to be formatted
   1254    * @param fallbackCarrierCode  the carrier selection code to be used, if none is found in the
   1255    *     phone number itself
   1256    * @return  the formatted phone number in national format for dialing using the number's
   1257    *     {@code preferredDomesticCarrierCode}, or the {@code fallbackCarrierCode} passed in if
   1258    *     none is found
   1259    */
   1260   public String formatNationalNumberWithPreferredCarrierCode(PhoneNumber number,
   1261                                                              String fallbackCarrierCode) {
   1262     return formatNationalNumberWithCarrierCode(number, number.hasPreferredDomesticCarrierCode()
   1263                                                        ? number.getPreferredDomesticCarrierCode()
   1264                                                        : fallbackCarrierCode);
   1265   }
   1266 
   1267   /**
   1268    * Returns a number formatted in such a way that it can be dialed from a mobile phone in a
   1269    * specific region. If the number cannot be reached from the region (e.g. some countries block
   1270    * toll-free numbers from being called outside of the country), the method returns an empty
   1271    * string.
   1272    *
   1273    * @param number  the phone number to be formatted
   1274    * @param regionCallingFrom  the region where the call is being placed
   1275    * @param withFormatting  whether the number should be returned with formatting symbols, such as
   1276    *     spaces and dashes.
   1277    * @return  the formatted phone number
   1278    */
   1279   public String formatNumberForMobileDialing(PhoneNumber number, String regionCallingFrom,
   1280                                              boolean withFormatting) {
   1281     int countryCallingCode = number.getCountryCode();
   1282     if (!hasValidCountryCallingCode(countryCallingCode)) {
   1283       return number.hasRawInput() ? number.getRawInput() : "";
   1284     }
   1285 
   1286     String formattedNumber = "";
   1287     // Clear the extension, as that part cannot normally be dialed together with the main number.
   1288     PhoneNumber numberNoExt = new PhoneNumber().mergeFrom(number).clearExtension();
   1289     String regionCode = getRegionCodeForCountryCode(countryCallingCode);
   1290     PhoneNumberType numberType = getNumberType(numberNoExt);
   1291     boolean isValidNumber = (numberType != PhoneNumberType.UNKNOWN);
   1292     if (regionCallingFrom.equals(regionCode)) {
   1293       boolean isFixedLineOrMobile =
   1294           (numberType == PhoneNumberType.FIXED_LINE) || (numberType == PhoneNumberType.MOBILE) ||
   1295           (numberType == PhoneNumberType.FIXED_LINE_OR_MOBILE);
   1296       // Carrier codes may be needed in some countries. We handle this here.
   1297       if (regionCode.equals("CO") && numberType == PhoneNumberType.FIXED_LINE) {
   1298         formattedNumber =
   1299             formatNationalNumberWithCarrierCode(numberNoExt, COLOMBIA_MOBILE_TO_FIXED_LINE_PREFIX);
   1300       } else if (regionCode.equals("BR") && isFixedLineOrMobile) {
   1301         formattedNumber = numberNoExt.hasPreferredDomesticCarrierCode()
   1302             ? formattedNumber = formatNationalNumberWithPreferredCarrierCode(numberNoExt, "")
   1303             // Brazilian fixed line and mobile numbers need to be dialed with a carrier code when
   1304             // called within Brazil. Without that, most of the carriers won't connect the call.
   1305             // Because of that, we return an empty string here.
   1306             : "";
   1307       } else if (isValidNumber && regionCode.equals("HU")) {
   1308         // The national format for HU numbers doesn't contain the national prefix, because that is
   1309         // how numbers are normally written down. However, the national prefix is obligatory when
   1310         // dialing from a mobile phone, except for short numbers. As a result, we add it back here
   1311         // if it is a valid regular length phone number.
   1312         formattedNumber =
   1313             getNddPrefixForRegion(regionCode, true /* strip non-digits */) +
   1314             " " + format(numberNoExt, PhoneNumberFormat.NATIONAL);
   1315       } else if (countryCallingCode == NANPA_COUNTRY_CODE) {
   1316         // For NANPA countries, we output international format for numbers that can be dialed
   1317         // internationally, since that always works, except for numbers which might potentially be
   1318         // short numbers, which are always dialled in national format.
   1319         PhoneMetadata regionMetadata = getMetadataForRegion(regionCallingFrom);
   1320         if (canBeInternationallyDialled(numberNoExt) &&
   1321             !isShorterThanPossibleNormalNumber(regionMetadata,
   1322                 getNationalSignificantNumber(numberNoExt))) {
   1323           formattedNumber = format(numberNoExt, PhoneNumberFormat.INTERNATIONAL);
   1324         } else {
   1325           formattedNumber = format(numberNoExt, PhoneNumberFormat.NATIONAL);
   1326         }
   1327       } else {
   1328         // For non-geographical countries, and Mexican and Chilean fixed line and mobile numbers, we
   1329         // output international format for numbers that can be dialed internationally as that always
   1330         // works.
   1331         if ((regionCode.equals(REGION_CODE_FOR_NON_GEO_ENTITY) ||
   1332             // MX fixed line and mobile numbers should always be formatted in international format,
   1333             // even when dialed within MX. For national format to work, a carrier code needs to be
   1334             // used, and the correct carrier code depends on if the caller and callee are from the
   1335             // same local area. It is trickier to get that to work correctly than using
   1336             // international format, which is tested to work fine on all carriers.
   1337             // CL fixed line numbers need the national prefix when dialing in the national format,
   1338             // but don't have it when used for display. The reverse is true for mobile numbers.
   1339             // As a result, we output them in the international format to make it work.
   1340             ((regionCode.equals("MX") || regionCode.equals("CL")) && isFixedLineOrMobile)) &&
   1341             canBeInternationallyDialled(numberNoExt)) {
   1342           formattedNumber = format(numberNoExt, PhoneNumberFormat.INTERNATIONAL);
   1343         } else {
   1344           formattedNumber = format(numberNoExt, PhoneNumberFormat.NATIONAL);
   1345         }
   1346       }
   1347     } else if (isValidNumber && canBeInternationallyDialled(numberNoExt)) {
   1348       // We assume that short numbers are not diallable from outside their region, so if a number
   1349       // is not a valid regular length phone number, we treat it as if it cannot be internationally
   1350       // dialled.
   1351       return withFormatting ? format(numberNoExt, PhoneNumberFormat.INTERNATIONAL)
   1352                             : format(numberNoExt, PhoneNumberFormat.E164);
   1353     }
   1354     return withFormatting ? formattedNumber
   1355                           : normalizeDiallableCharsOnly(formattedNumber);
   1356   }
   1357 
   1358   /**
   1359    * Formats a phone number for out-of-country dialing purposes. If no regionCallingFrom is
   1360    * supplied, we format the number in its INTERNATIONAL format. If the country calling code is the
   1361    * same as that of the region where the number is from, then NATIONAL formatting will be applied.
   1362    *
   1363    * <p>If the number itself has a country calling code of zero or an otherwise invalid country
   1364    * calling code, then we return the number with no formatting applied.
   1365    *
   1366    * <p>Note this function takes care of the case for calling inside of NANPA and between Russia and
   1367    * Kazakhstan (who share the same country calling code). In those cases, no international prefix
   1368    * is used. For regions which have multiple international prefixes, the number in its
   1369    * INTERNATIONAL format will be returned instead.
   1370    *
   1371    * @param number               the phone number to be formatted
   1372    * @param regionCallingFrom    the region where the call is being placed
   1373    * @return  the formatted phone number
   1374    */
   1375   public String formatOutOfCountryCallingNumber(PhoneNumber number,
   1376                                                 String regionCallingFrom) {
   1377     if (!isValidRegionCode(regionCallingFrom)) {
   1378       logger.log(Level.WARNING,
   1379                  "Trying to format number from invalid region "
   1380                  + regionCallingFrom
   1381                  + ". International formatting applied.");
   1382       return format(number, PhoneNumberFormat.INTERNATIONAL);
   1383     }
   1384     int countryCallingCode = number.getCountryCode();
   1385     String nationalSignificantNumber = getNationalSignificantNumber(number);
   1386     if (!hasValidCountryCallingCode(countryCallingCode)) {
   1387       return nationalSignificantNumber;
   1388     }
   1389     if (countryCallingCode == NANPA_COUNTRY_CODE) {
   1390       if (isNANPACountry(regionCallingFrom)) {
   1391         // For NANPA regions, return the national format for these regions but prefix it with the
   1392         // country calling code.
   1393         return countryCallingCode + " " + format(number, PhoneNumberFormat.NATIONAL);
   1394       }
   1395     } else if (countryCallingCode == getCountryCodeForValidRegion(regionCallingFrom)) {
   1396       // If regions share a country calling code, the country calling code need not be dialled.
   1397       // This also applies when dialling within a region, so this if clause covers both these cases.
   1398       // Technically this is the case for dialling from La Reunion to other overseas departments of
   1399       // France (French Guiana, Martinique, Guadeloupe), but not vice versa - so we don't cover this
   1400       // edge case for now and for those cases return the version including country calling code.
   1401       // Details here: http://www.petitfute.com/voyage/225-info-pratiques-reunion
   1402       return format(number, PhoneNumberFormat.NATIONAL);
   1403     }
   1404     // Metadata cannot be null because we checked 'isValidRegionCode()' above.
   1405     PhoneMetadata metadataForRegionCallingFrom = getMetadataForRegion(regionCallingFrom);
   1406     String internationalPrefix = metadataForRegionCallingFrom.getInternationalPrefix();
   1407 
   1408     // For regions that have multiple international prefixes, the international format of the
   1409     // number is returned, unless there is a preferred international prefix.
   1410     String internationalPrefixForFormatting = "";
   1411     if (UNIQUE_INTERNATIONAL_PREFIX.matcher(internationalPrefix).matches()) {
   1412       internationalPrefixForFormatting = internationalPrefix;
   1413     } else if (metadataForRegionCallingFrom.hasPreferredInternationalPrefix()) {
   1414       internationalPrefixForFormatting =
   1415           metadataForRegionCallingFrom.getPreferredInternationalPrefix();
   1416     }
   1417 
   1418     String regionCode = getRegionCodeForCountryCode(countryCallingCode);
   1419     // Metadata cannot be null because the country calling code is valid.
   1420     PhoneMetadata metadataForRegion =
   1421         getMetadataForRegionOrCallingCode(countryCallingCode, regionCode);
   1422     String formattedNationalNumber =
   1423         formatNsn(nationalSignificantNumber, metadataForRegion, PhoneNumberFormat.INTERNATIONAL);
   1424     StringBuilder formattedNumber = new StringBuilder(formattedNationalNumber);
   1425     maybeAppendFormattedExtension(number, metadataForRegion, PhoneNumberFormat.INTERNATIONAL,
   1426                                   formattedNumber);
   1427     if (internationalPrefixForFormatting.length() > 0) {
   1428       formattedNumber.insert(0, " ").insert(0, countryCallingCode).insert(0, " ")
   1429           .insert(0, internationalPrefixForFormatting);
   1430     } else {
   1431       prefixNumberWithCountryCallingCode(countryCallingCode,
   1432                                          PhoneNumberFormat.INTERNATIONAL,
   1433                                          formattedNumber);
   1434     }
   1435     return formattedNumber.toString();
   1436   }
   1437 
   1438   /**
   1439    * Formats a phone number using the original phone number format that the number is parsed from.
   1440    * The original format is embedded in the country_code_source field of the PhoneNumber object
   1441    * passed in. If such information is missing, the number will be formatted into the NATIONAL
   1442    * format by default. When the number contains a leading zero and this is unexpected for this
   1443    * country, or we don't have a formatting pattern for the number, the method returns the raw input
   1444    * when it is available.
   1445    *
   1446    * Note this method guarantees no digit will be inserted, removed or modified as a result of
   1447    * formatting.
   1448    *
   1449    * @param number  the phone number that needs to be formatted in its original number format
   1450    * @param regionCallingFrom  the region whose IDD needs to be prefixed if the original number
   1451    *     has one
   1452    * @return  the formatted phone number in its original number format
   1453    */
   1454   public String formatInOriginalFormat(PhoneNumber number, String regionCallingFrom) {
   1455     if (number.hasRawInput() &&
   1456         (hasUnexpectedItalianLeadingZero(number) || !hasFormattingPatternForNumber(number))) {
   1457       // We check if we have the formatting pattern because without that, we might format the number
   1458       // as a group without national prefix.
   1459       return number.getRawInput();
   1460     }
   1461     if (!number.hasCountryCodeSource()) {
   1462       return format(number, PhoneNumberFormat.NATIONAL);
   1463     }
   1464     String formattedNumber;
   1465     switch (number.getCountryCodeSource()) {
   1466       case FROM_NUMBER_WITH_PLUS_SIGN:
   1467         formattedNumber = format(number, PhoneNumberFormat.INTERNATIONAL);
   1468         break;
   1469       case FROM_NUMBER_WITH_IDD:
   1470         formattedNumber = formatOutOfCountryCallingNumber(number, regionCallingFrom);
   1471         break;
   1472       case FROM_NUMBER_WITHOUT_PLUS_SIGN:
   1473         formattedNumber = format(number, PhoneNumberFormat.INTERNATIONAL).substring(1);
   1474         break;
   1475       case FROM_DEFAULT_COUNTRY:
   1476         // Fall-through to default case.
   1477       default:
   1478         String regionCode = getRegionCodeForCountryCode(number.getCountryCode());
   1479         // We strip non-digits from the NDD here, and from the raw input later, so that we can
   1480         // compare them easily.
   1481         String nationalPrefix = getNddPrefixForRegion(regionCode, true /* strip non-digits */);
   1482         String nationalFormat = format(number, PhoneNumberFormat.NATIONAL);
   1483         if (nationalPrefix == null || nationalPrefix.length() == 0) {
   1484           // If the region doesn't have a national prefix at all, we can safely return the national
   1485           // format without worrying about a national prefix being added.
   1486           formattedNumber = nationalFormat;
   1487           break;
   1488         }
   1489         // Otherwise, we check if the original number was entered with a national prefix.
   1490         if (rawInputContainsNationalPrefix(
   1491             number.getRawInput(), nationalPrefix, regionCode)) {
   1492           // If so, we can safely return the national format.
   1493           formattedNumber = nationalFormat;
   1494           break;
   1495         }
   1496         // Metadata cannot be null here because getNddPrefixForRegion() (above) returns null if
   1497         // there is no metadata for the region.
   1498         PhoneMetadata metadata = getMetadataForRegion(regionCode);
   1499         String nationalNumber = getNationalSignificantNumber(number);
   1500         NumberFormat formatRule =
   1501             chooseFormattingPatternForNumber(metadata.numberFormats(), nationalNumber);
   1502         // The format rule could still be null here if the national number was 0 and there was no
   1503         // raw input (this should not be possible for numbers generated by the phonenumber library
   1504         // as they would also not have a country calling code and we would have exited earlier).
   1505         if (formatRule == null) {
   1506           formattedNumber = nationalFormat;
   1507           break;
   1508         }
   1509         // When the format we apply to this number doesn't contain national prefix, we can just
   1510         // return the national format.
   1511         // TODO: Refactor the code below with the code in isNationalPrefixPresentIfRequired.
   1512         String candidateNationalPrefixRule = formatRule.getNationalPrefixFormattingRule();
   1513         // We assume that the first-group symbol will never be _before_ the national prefix.
   1514         int indexOfFirstGroup = candidateNationalPrefixRule.indexOf("$1");
   1515         if (indexOfFirstGroup <= 0) {
   1516           formattedNumber = nationalFormat;
   1517           break;
   1518         }
   1519         candidateNationalPrefixRule =
   1520             candidateNationalPrefixRule.substring(0, indexOfFirstGroup);
   1521         candidateNationalPrefixRule = normalizeDigitsOnly(candidateNationalPrefixRule);
   1522         if (candidateNationalPrefixRule.length() == 0) {
   1523           // National prefix not used when formatting this number.
   1524           formattedNumber = nationalFormat;
   1525           break;
   1526         }
   1527         // Otherwise, we need to remove the national prefix from our output.
   1528         NumberFormat numFormatCopy = new NumberFormat();
   1529         numFormatCopy.mergeFrom(formatRule);
   1530         numFormatCopy.clearNationalPrefixFormattingRule();
   1531         List<NumberFormat> numberFormats = new ArrayList<NumberFormat>(1);
   1532         numberFormats.add(numFormatCopy);
   1533         formattedNumber = formatByPattern(number, PhoneNumberFormat.NATIONAL, numberFormats);
   1534         break;
   1535     }
   1536     String rawInput = number.getRawInput();
   1537     // If no digit is inserted/removed/modified as a result of our formatting, we return the
   1538     // formatted phone number; otherwise we return the raw input the user entered.
   1539     if (formattedNumber != null && rawInput.length() > 0) {
   1540       String normalizedFormattedNumber = normalizeDiallableCharsOnly(formattedNumber);
   1541       String normalizedRawInput = normalizeDiallableCharsOnly(rawInput);
   1542       if (!normalizedFormattedNumber.equals(normalizedRawInput)) {
   1543         formattedNumber = rawInput;
   1544       }
   1545     }
   1546     return formattedNumber;
   1547   }
   1548 
   1549   // Check if rawInput, which is assumed to be in the national format, has a national prefix. The
   1550   // national prefix is assumed to be in digits-only form.
   1551   private boolean rawInputContainsNationalPrefix(String rawInput, String nationalPrefix,
   1552       String regionCode) {
   1553     String normalizedNationalNumber = normalizeDigitsOnly(rawInput);
   1554     if (normalizedNationalNumber.startsWith(nationalPrefix)) {
   1555       try {
   1556         // Some Japanese numbers (e.g. 00777123) might be mistaken to contain the national prefix
   1557         // when written without it (e.g. 0777123) if we just do prefix matching. To tackle that, we
   1558         // check the validity of the number if the assumed national prefix is removed (777123 won't
   1559         // be valid in Japan).
   1560         return isValidNumber(
   1561             parse(normalizedNationalNumber.substring(nationalPrefix.length()), regionCode));
   1562       } catch (NumberParseException e) {
   1563         return false;
   1564       }
   1565     }
   1566     return false;
   1567   }
   1568 
   1569   /**
   1570    * Returns true if a number is from a region whose national significant number couldn't contain a
   1571    * leading zero, but has the italian_leading_zero field set to true.
   1572    */
   1573   private boolean hasUnexpectedItalianLeadingZero(PhoneNumber number) {
   1574     return number.isItalianLeadingZero() && !isLeadingZeroPossible(number.getCountryCode());
   1575   }
   1576 
   1577   private boolean hasFormattingPatternForNumber(PhoneNumber number) {
   1578     int countryCallingCode = number.getCountryCode();
   1579     String phoneNumberRegion = getRegionCodeForCountryCode(countryCallingCode);
   1580     PhoneMetadata metadata =
   1581         getMetadataForRegionOrCallingCode(countryCallingCode, phoneNumberRegion);
   1582     if (metadata == null) {
   1583       return false;
   1584     }
   1585     String nationalNumber = getNationalSignificantNumber(number);
   1586     NumberFormat formatRule =
   1587         chooseFormattingPatternForNumber(metadata.numberFormats(), nationalNumber);
   1588     return formatRule != null;
   1589   }
   1590 
   1591   /**
   1592    * Formats a phone number for out-of-country dialing purposes.
   1593    *
   1594    * Note that in this version, if the number was entered originally using alpha characters and
   1595    * this version of the number is stored in raw_input, this representation of the number will be
   1596    * used rather than the digit representation. Grouping information, as specified by characters
   1597    * such as "-" and " ", will be retained.
   1598    *
   1599    * <p><b>Caveats:</b></p>
   1600    * <ul>
   1601    *  <li> This will not produce good results if the country calling code is both present in the raw
   1602    *       input _and_ is the start of the national number. This is not a problem in the regions
   1603    *       which typically use alpha numbers.
   1604    *  <li> This will also not produce good results if the raw input has any grouping information
   1605    *       within the first three digits of the national number, and if the function needs to strip
   1606    *       preceding digits/words in the raw input before these digits. Normally people group the
   1607    *       first three digits together so this is not a huge problem - and will be fixed if it
   1608    *       proves to be so.
   1609    * </ul>
   1610    *
   1611    * @param number  the phone number that needs to be formatted
   1612    * @param regionCallingFrom  the region where the call is being placed
   1613    * @return  the formatted phone number
   1614    */
   1615   public String formatOutOfCountryKeepingAlphaChars(PhoneNumber number,
   1616                                                     String regionCallingFrom) {
   1617     String rawInput = number.getRawInput();
   1618     // If there is no raw input, then we can't keep alpha characters because there aren't any.
   1619     // In this case, we return formatOutOfCountryCallingNumber.
   1620     if (rawInput.length() == 0) {
   1621       return formatOutOfCountryCallingNumber(number, regionCallingFrom);
   1622     }
   1623     int countryCode = number.getCountryCode();
   1624     if (!hasValidCountryCallingCode(countryCode)) {
   1625       return rawInput;
   1626     }
   1627     // Strip any prefix such as country calling code, IDD, that was present. We do this by comparing
   1628     // the number in raw_input with the parsed number.
   1629     // To do this, first we normalize punctuation. We retain number grouping symbols such as " "
   1630     // only.
   1631     rawInput = normalizeHelper(rawInput, ALL_PLUS_NUMBER_GROUPING_SYMBOLS, true);
   1632     // Now we trim everything before the first three digits in the parsed number. We choose three
   1633     // because all valid alpha numbers have 3 digits at the start - if it does not, then we don't
   1634     // trim anything at all. Similarly, if the national number was less than three digits, we don't
   1635     // trim anything at all.
   1636     String nationalNumber = getNationalSignificantNumber(number);
   1637     if (nationalNumber.length() > 3) {
   1638       int firstNationalNumberDigit = rawInput.indexOf(nationalNumber.substring(0, 3));
   1639       if (firstNationalNumberDigit != -1) {
   1640         rawInput = rawInput.substring(firstNationalNumberDigit);
   1641       }
   1642     }
   1643     PhoneMetadata metadataForRegionCallingFrom = getMetadataForRegion(regionCallingFrom);
   1644     if (countryCode == NANPA_COUNTRY_CODE) {
   1645       if (isNANPACountry(regionCallingFrom)) {
   1646         return countryCode + " " + rawInput;
   1647       }
   1648     } else if (metadataForRegionCallingFrom != null &&
   1649                countryCode == getCountryCodeForValidRegion(regionCallingFrom)) {
   1650       NumberFormat formattingPattern =
   1651           chooseFormattingPatternForNumber(metadataForRegionCallingFrom.numberFormats(),
   1652                                            nationalNumber);
   1653       if (formattingPattern == null) {
   1654         // If no pattern above is matched, we format the original input.
   1655         return rawInput;
   1656       }
   1657       NumberFormat newFormat = new NumberFormat();
   1658       newFormat.mergeFrom(formattingPattern);
   1659       // The first group is the first group of digits that the user wrote together.
   1660       newFormat.setPattern("(\\d+)(.*)");
   1661       // Here we just concatenate them back together after the national prefix has been fixed.
   1662       newFormat.setFormat("$1$2");
   1663       // Now we format using this pattern instead of the default pattern, but with the national
   1664       // prefix prefixed if necessary.
   1665       // This will not work in the cases where the pattern (and not the leading digits) decide
   1666       // whether a national prefix needs to be used, since we have overridden the pattern to match
   1667       // anything, but that is not the case in the metadata to date.
   1668       return formatNsnUsingPattern(rawInput, newFormat, PhoneNumberFormat.NATIONAL);
   1669     }
   1670     String internationalPrefixForFormatting = "";
   1671     // If an unsupported region-calling-from is entered, or a country with multiple international
   1672     // prefixes, the international format of the number is returned, unless there is a preferred
   1673     // international prefix.
   1674     if (metadataForRegionCallingFrom != null) {
   1675       String internationalPrefix = metadataForRegionCallingFrom.getInternationalPrefix();
   1676       internationalPrefixForFormatting =
   1677           UNIQUE_INTERNATIONAL_PREFIX.matcher(internationalPrefix).matches()
   1678           ? internationalPrefix
   1679           : metadataForRegionCallingFrom.getPreferredInternationalPrefix();
   1680     }
   1681     StringBuilder formattedNumber = new StringBuilder(rawInput);
   1682     String regionCode = getRegionCodeForCountryCode(countryCode);
   1683     // Metadata cannot be null because the country calling code is valid.
   1684     PhoneMetadata metadataForRegion = getMetadataForRegionOrCallingCode(countryCode, regionCode);
   1685     maybeAppendFormattedExtension(number, metadataForRegion,
   1686                                   PhoneNumberFormat.INTERNATIONAL, formattedNumber);
   1687     if (internationalPrefixForFormatting.length() > 0) {
   1688       formattedNumber.insert(0, " ").insert(0, countryCode).insert(0, " ")
   1689           .insert(0, internationalPrefixForFormatting);
   1690     } else {
   1691       // Invalid region entered as country-calling-from (so no metadata was found for it) or the
   1692       // region chosen has multiple international dialling prefixes.
   1693       logger.log(Level.WARNING,
   1694                  "Trying to format number from invalid region "
   1695                  + regionCallingFrom
   1696                  + ". International formatting applied.");
   1697       prefixNumberWithCountryCallingCode(countryCode,
   1698                                          PhoneNumberFormat.INTERNATIONAL,
   1699                                          formattedNumber);
   1700     }
   1701     return formattedNumber.toString();
   1702   }
   1703 
   1704   /**
   1705    * Gets the national significant number of the a phone number. Note a national significant number
   1706    * doesn't contain a national prefix or any formatting.
   1707    *
   1708    * @param number  the phone number for which the national significant number is needed
   1709    * @return  the national significant number of the PhoneNumber object passed in
   1710    */
   1711   public String getNationalSignificantNumber(PhoneNumber number) {
   1712     // If leading zero(s) have been set, we prefix this now. Note this is not a national prefix.
   1713     StringBuilder nationalNumber = new StringBuilder();
   1714     if (number.isItalianLeadingZero()) {
   1715       char[] zeros = new char[number.getNumberOfLeadingZeros()];
   1716       Arrays.fill(zeros, '0');
   1717       nationalNumber.append(new String(zeros));
   1718     }
   1719     nationalNumber.append(number.getNationalNumber());
   1720     return nationalNumber.toString();
   1721   }
   1722 
   1723   /**
   1724    * A helper function that is used by format and formatByPattern.
   1725    */
   1726   private void prefixNumberWithCountryCallingCode(int countryCallingCode,
   1727                                                   PhoneNumberFormat numberFormat,
   1728                                                   StringBuilder formattedNumber) {
   1729     switch (numberFormat) {
   1730       case E164:
   1731         formattedNumber.insert(0, countryCallingCode).insert(0, PLUS_SIGN);
   1732         return;
   1733       case INTERNATIONAL:
   1734         formattedNumber.insert(0, " ").insert(0, countryCallingCode).insert(0, PLUS_SIGN);
   1735         return;
   1736       case RFC3966:
   1737         formattedNumber.insert(0, "-").insert(0, countryCallingCode).insert(0, PLUS_SIGN)
   1738             .insert(0, RFC3966_PREFIX);
   1739         return;
   1740       case NATIONAL:
   1741       default:
   1742         return;
   1743     }
   1744   }
   1745 
   1746   // Simple wrapper of formatNsn for the common case of no carrier code.
   1747   private String formatNsn(String number, PhoneMetadata metadata, PhoneNumberFormat numberFormat) {
   1748     return formatNsn(number, metadata, numberFormat, null);
   1749   }
   1750 
   1751   // Note in some regions, the national number can be written in two completely different ways
   1752   // depending on whether it forms part of the NATIONAL format or INTERNATIONAL format. The
   1753   // numberFormat parameter here is used to specify which format to use for those cases. If a
   1754   // carrierCode is specified, this will be inserted into the formatted string to replace $CC.
   1755   private String formatNsn(String number,
   1756                            PhoneMetadata metadata,
   1757                            PhoneNumberFormat numberFormat,
   1758                            String carrierCode) {
   1759     List<NumberFormat> intlNumberFormats = metadata.intlNumberFormats();
   1760     // When the intlNumberFormats exists, we use that to format national number for the
   1761     // INTERNATIONAL format instead of using the numberDesc.numberFormats.
   1762     List<NumberFormat> availableFormats =
   1763         (intlNumberFormats.size() == 0 || numberFormat == PhoneNumberFormat.NATIONAL)
   1764         ? metadata.numberFormats()
   1765         : metadata.intlNumberFormats();
   1766     NumberFormat formattingPattern = chooseFormattingPatternForNumber(availableFormats, number);
   1767     return (formattingPattern == null)
   1768         ? number
   1769         : formatNsnUsingPattern(number, formattingPattern, numberFormat, carrierCode);
   1770   }
   1771 
   1772   NumberFormat chooseFormattingPatternForNumber(List<NumberFormat> availableFormats,
   1773                                                 String nationalNumber) {
   1774     for (NumberFormat numFormat : availableFormats) {
   1775       int size = numFormat.leadingDigitsPatternSize();
   1776       if (size == 0 || regexCache.getPatternForRegex(
   1777               // We always use the last leading_digits_pattern, as it is the most detailed.
   1778               numFormat.getLeadingDigitsPattern(size - 1)).matcher(nationalNumber).lookingAt()) {
   1779         Matcher m = regexCache.getPatternForRegex(numFormat.getPattern()).matcher(nationalNumber);
   1780         if (m.matches()) {
   1781           return numFormat;
   1782         }
   1783       }
   1784     }
   1785     return null;
   1786   }
   1787 
   1788   // Simple wrapper of formatNsnUsingPattern for the common case of no carrier code.
   1789   String formatNsnUsingPattern(String nationalNumber,
   1790                                NumberFormat formattingPattern,
   1791                                PhoneNumberFormat numberFormat) {
   1792     return formatNsnUsingPattern(nationalNumber, formattingPattern, numberFormat, null);
   1793   }
   1794 
   1795   // Note that carrierCode is optional - if null or an empty string, no carrier code replacement
   1796   // will take place.
   1797   private String formatNsnUsingPattern(String nationalNumber,
   1798                                        NumberFormat formattingPattern,
   1799                                        PhoneNumberFormat numberFormat,
   1800                                        String carrierCode) {
   1801     String numberFormatRule = formattingPattern.getFormat();
   1802     Matcher m =
   1803         regexCache.getPatternForRegex(formattingPattern.getPattern()).matcher(nationalNumber);
   1804     String formattedNationalNumber = "";
   1805     if (numberFormat == PhoneNumberFormat.NATIONAL &&
   1806         carrierCode != null && carrierCode.length() > 0 &&
   1807         formattingPattern.getDomesticCarrierCodeFormattingRule().length() > 0) {
   1808       // Replace the $CC in the formatting rule with the desired carrier code.
   1809       String carrierCodeFormattingRule = formattingPattern.getDomesticCarrierCodeFormattingRule();
   1810       carrierCodeFormattingRule =
   1811           CC_PATTERN.matcher(carrierCodeFormattingRule).replaceFirst(carrierCode);
   1812       // Now replace the $FG in the formatting rule with the first group and the carrier code
   1813       // combined in the appropriate way.
   1814       numberFormatRule = FIRST_GROUP_PATTERN.matcher(numberFormatRule)
   1815           .replaceFirst(carrierCodeFormattingRule);
   1816       formattedNationalNumber = m.replaceAll(numberFormatRule);
   1817     } else {
   1818       // Use the national prefix formatting rule instead.
   1819       String nationalPrefixFormattingRule = formattingPattern.getNationalPrefixFormattingRule();
   1820       if (numberFormat == PhoneNumberFormat.NATIONAL &&
   1821           nationalPrefixFormattingRule != null &&
   1822           nationalPrefixFormattingRule.length() > 0) {
   1823         Matcher firstGroupMatcher = FIRST_GROUP_PATTERN.matcher(numberFormatRule);
   1824         formattedNationalNumber =
   1825             m.replaceAll(firstGroupMatcher.replaceFirst(nationalPrefixFormattingRule));
   1826       } else {
   1827         formattedNationalNumber = m.replaceAll(numberFormatRule);
   1828       }
   1829     }
   1830     if (numberFormat == PhoneNumberFormat.RFC3966) {
   1831       // Strip any leading punctuation.
   1832       Matcher matcher = SEPARATOR_PATTERN.matcher(formattedNationalNumber);
   1833       if (matcher.lookingAt()) {
   1834         formattedNationalNumber = matcher.replaceFirst("");
   1835       }
   1836       // Replace the rest with a dash between each number group.
   1837       formattedNationalNumber = matcher.reset(formattedNationalNumber).replaceAll("-");
   1838     }
   1839     return formattedNationalNumber;
   1840   }
   1841 
   1842   /**
   1843    * Gets a valid number for the specified region.
   1844    *
   1845    * @param regionCode  the region for which an example number is needed
   1846    * @return  a valid fixed-line number for the specified region. Returns null when the metadata
   1847    *    does not contain such information, or the region 001 is passed in. For 001 (representing
   1848    *    non-geographical numbers), call {@link #getExampleNumberForNonGeoEntity} instead.
   1849    */
   1850   public PhoneNumber getExampleNumber(String regionCode) {
   1851     return getExampleNumberForType(regionCode, PhoneNumberType.FIXED_LINE);
   1852   }
   1853 
   1854   /**
   1855    * Gets a valid number for the specified region and number type.
   1856    *
   1857    * @param regionCode  the region for which an example number is needed
   1858    * @param type  the type of number that is needed
   1859    * @return  a valid number for the specified region and type. Returns null when the metadata
   1860    *     does not contain such information or if an invalid region or region 001 was entered.
   1861    *     For 001 (representing non-geographical numbers), call
   1862    *     {@link #getExampleNumberForNonGeoEntity} instead.
   1863    */
   1864   public PhoneNumber getExampleNumberForType(String regionCode, PhoneNumberType type) {
   1865     // Check the region code is valid.
   1866     if (!isValidRegionCode(regionCode)) {
   1867       logger.log(Level.WARNING, "Invalid or unknown region code provided: " + regionCode);
   1868       return null;
   1869     }
   1870     PhoneNumberDesc desc = getNumberDescByType(getMetadataForRegion(regionCode), type);
   1871     try {
   1872       if (desc.hasExampleNumber()) {
   1873         return parse(desc.getExampleNumber(), regionCode);
   1874       }
   1875     } catch (NumberParseException e) {
   1876       logger.log(Level.SEVERE, e.toString());
   1877     }
   1878     return null;
   1879   }
   1880 
   1881   /**
   1882    * Gets a valid number for the specified country calling code for a non-geographical entity.
   1883    *
   1884    * @param countryCallingCode  the country calling code for a non-geographical entity
   1885    * @return  a valid number for the non-geographical entity. Returns null when the metadata
   1886    *    does not contain such information, or the country calling code passed in does not belong
   1887    *    to a non-geographical entity.
   1888    */
   1889   public PhoneNumber getExampleNumberForNonGeoEntity(int countryCallingCode) {
   1890     PhoneMetadata metadata = getMetadataForNonGeographicalRegion(countryCallingCode);
   1891     if (metadata != null) {
   1892       PhoneNumberDesc desc = metadata.getGeneralDesc();
   1893       try {
   1894         if (desc.hasExampleNumber()) {
   1895           return parse("+" + countryCallingCode + desc.getExampleNumber(), "ZZ");
   1896         }
   1897       } catch (NumberParseException e) {
   1898         logger.log(Level.SEVERE, e.toString());
   1899       }
   1900     } else {
   1901       logger.log(Level.WARNING,
   1902                  "Invalid or unknown country calling code provided: " + countryCallingCode);
   1903     }
   1904     return null;
   1905   }
   1906 
   1907   /**
   1908    * Appends the formatted extension of a phone number to formattedNumber, if the phone number had
   1909    * an extension specified.
   1910    */
   1911   private void maybeAppendFormattedExtension(PhoneNumber number, PhoneMetadata metadata,
   1912                                              PhoneNumberFormat numberFormat,
   1913                                              StringBuilder formattedNumber) {
   1914     if (number.hasExtension() && number.getExtension().length() > 0) {
   1915       if (numberFormat == PhoneNumberFormat.RFC3966) {
   1916         formattedNumber.append(RFC3966_EXTN_PREFIX).append(number.getExtension());
   1917       } else {
   1918         if (metadata.hasPreferredExtnPrefix()) {
   1919           formattedNumber.append(metadata.getPreferredExtnPrefix()).append(number.getExtension());
   1920         } else {
   1921           formattedNumber.append(DEFAULT_EXTN_PREFIX).append(number.getExtension());
   1922         }
   1923       }
   1924     }
   1925   }
   1926 
   1927   PhoneNumberDesc getNumberDescByType(PhoneMetadata metadata, PhoneNumberType type) {
   1928     switch (type) {
   1929       case PREMIUM_RATE:
   1930         return metadata.getPremiumRate();
   1931       case TOLL_FREE:
   1932         return metadata.getTollFree();
   1933       case MOBILE:
   1934         return metadata.getMobile();
   1935       case FIXED_LINE:
   1936       case FIXED_LINE_OR_MOBILE:
   1937         return metadata.getFixedLine();
   1938       case SHARED_COST:
   1939         return metadata.getSharedCost();
   1940       case VOIP:
   1941         return metadata.getVoip();
   1942       case PERSONAL_NUMBER:
   1943         return metadata.getPersonalNumber();
   1944       case PAGER:
   1945         return metadata.getPager();
   1946       case UAN:
   1947         return metadata.getUan();
   1948       case VOICEMAIL:
   1949         return metadata.getVoicemail();
   1950       default:
   1951         return metadata.getGeneralDesc();
   1952     }
   1953   }
   1954 
   1955   /**
   1956    * Gets the type of a phone number.
   1957    *
   1958    * @param number  the phone number that we want to know the type
   1959    * @return  the type of the phone number
   1960    */
   1961   public PhoneNumberType getNumberType(PhoneNumber number) {
   1962     String regionCode = getRegionCodeForNumber(number);
   1963     PhoneMetadata metadata = getMetadataForRegionOrCallingCode(number.getCountryCode(), regionCode);
   1964     if (metadata == null) {
   1965       return PhoneNumberType.UNKNOWN;
   1966     }
   1967     String nationalSignificantNumber = getNationalSignificantNumber(number);
   1968     return getNumberTypeHelper(nationalSignificantNumber, metadata);
   1969   }
   1970 
   1971   private PhoneNumberType getNumberTypeHelper(String nationalNumber, PhoneMetadata metadata) {
   1972     PhoneNumberDesc generalNumberDesc = metadata.getGeneralDesc();
   1973     if (!generalNumberDesc.hasNationalNumberPattern() ||
   1974         !isNumberMatchingDesc(nationalNumber, generalNumberDesc)) {
   1975       return PhoneNumberType.UNKNOWN;
   1976     }
   1977 
   1978     if (isNumberMatchingDesc(nationalNumber, metadata.getPremiumRate())) {
   1979       return PhoneNumberType.PREMIUM_RATE;
   1980     }
   1981     if (isNumberMatchingDesc(nationalNumber, metadata.getTollFree())) {
   1982       return PhoneNumberType.TOLL_FREE;
   1983     }
   1984     if (isNumberMatchingDesc(nationalNumber, metadata.getSharedCost())) {
   1985       return PhoneNumberType.SHARED_COST;
   1986     }
   1987     if (isNumberMatchingDesc(nationalNumber, metadata.getVoip())) {
   1988       return PhoneNumberType.VOIP;
   1989     }
   1990     if (isNumberMatchingDesc(nationalNumber, metadata.getPersonalNumber())) {
   1991       return PhoneNumberType.PERSONAL_NUMBER;
   1992     }
   1993     if (isNumberMatchingDesc(nationalNumber, metadata.getPager())) {
   1994       return PhoneNumberType.PAGER;
   1995     }
   1996     if (isNumberMatchingDesc(nationalNumber, metadata.getUan())) {
   1997       return PhoneNumberType.UAN;
   1998     }
   1999     if (isNumberMatchingDesc(nationalNumber, metadata.getVoicemail())) {
   2000       return PhoneNumberType.VOICEMAIL;
   2001     }
   2002 
   2003     boolean isFixedLine = isNumberMatchingDesc(nationalNumber, metadata.getFixedLine());
   2004     if (isFixedLine) {
   2005       if (metadata.isSameMobileAndFixedLinePattern()) {
   2006         return PhoneNumberType.FIXED_LINE_OR_MOBILE;
   2007       } else if (isNumberMatchingDesc(nationalNumber, metadata.getMobile())) {
   2008         return PhoneNumberType.FIXED_LINE_OR_MOBILE;
   2009       }
   2010       return PhoneNumberType.FIXED_LINE;
   2011     }
   2012     // Otherwise, test to see if the number is mobile. Only do this if certain that the patterns for
   2013     // mobile and fixed line aren't the same.
   2014     if (!metadata.isSameMobileAndFixedLinePattern() &&
   2015         isNumberMatchingDesc(nationalNumber, metadata.getMobile())) {
   2016       return PhoneNumberType.MOBILE;
   2017     }
   2018     return PhoneNumberType.UNKNOWN;
   2019   }
   2020 
   2021   /**
   2022    * Returns the metadata for the given region code or {@code null} if the region code is invalid
   2023    * or unknown.
   2024    */
   2025   PhoneMetadata getMetadataForRegion(String regionCode) {
   2026     if (!isValidRegionCode(regionCode)) {
   2027       return null;
   2028     }
   2029     synchronized (regionToMetadataMap) {
   2030       if (!regionToMetadataMap.containsKey(regionCode)) {
   2031         // The regionCode here will be valid and won't be '001', so we don't need to worry about
   2032         // what to pass in for the country calling code.
   2033         loadMetadataFromFile(currentFilePrefix, regionCode, 0, metadataLoader);
   2034       }
   2035     }
   2036     return regionToMetadataMap.get(regionCode);
   2037   }
   2038 
   2039   PhoneMetadata getMetadataForNonGeographicalRegion(int countryCallingCode) {
   2040     synchronized (countryCodeToNonGeographicalMetadataMap) {
   2041       if (!countryCallingCodeToRegionCodeMap.containsKey(countryCallingCode)) {
   2042         return null;
   2043       }
   2044       if (!countryCodeToNonGeographicalMetadataMap.containsKey(countryCallingCode)) {
   2045         loadMetadataFromFile(
   2046             currentFilePrefix, REGION_CODE_FOR_NON_GEO_ENTITY, countryCallingCode, metadataLoader);
   2047       }
   2048     }
   2049     return countryCodeToNonGeographicalMetadataMap.get(countryCallingCode);
   2050   }
   2051 
   2052   boolean isNumberPossibleForDesc(String nationalNumber, PhoneNumberDesc numberDesc) {
   2053     Matcher possibleNumberPatternMatcher =
   2054         regexCache.getPatternForRegex(numberDesc.getPossibleNumberPattern())
   2055             .matcher(nationalNumber);
   2056     return possibleNumberPatternMatcher.matches();
   2057   }
   2058 
   2059   boolean isNumberMatchingDesc(String nationalNumber, PhoneNumberDesc numberDesc) {
   2060     Matcher nationalNumberPatternMatcher =
   2061         regexCache.getPatternForRegex(numberDesc.getNationalNumberPattern())
   2062             .matcher(nationalNumber);
   2063     return isNumberPossibleForDesc(nationalNumber, numberDesc) &&
   2064         nationalNumberPatternMatcher.matches();
   2065   }
   2066 
   2067   /**
   2068    * Tests whether a phone number matches a valid pattern. Note this doesn't verify the number
   2069    * is actually in use, which is impossible to tell by just looking at a number itself.
   2070    *
   2071    * @param number       the phone number that we want to validate
   2072    * @return  a boolean that indicates whether the number is of a valid pattern
   2073    */
   2074   public boolean isValidNumber(PhoneNumber number) {
   2075     String regionCode = getRegionCodeForNumber(number);
   2076     return isValidNumberForRegion(number, regionCode);
   2077   }
   2078 
   2079   /**
   2080    * Tests whether a phone number is valid for a certain region. Note this doesn't verify the number
   2081    * is actually in use, which is impossible to tell by just looking at a number itself. If the
   2082    * country calling code is not the same as the country calling code for the region, this
   2083    * immediately exits with false. After this, the specific number pattern rules for the region are
   2084    * examined. This is useful for determining for example whether a particular number is valid for
   2085    * Canada, rather than just a valid NANPA number.
   2086    * Warning: In most cases, you want to use {@link #isValidNumber} instead. For example, this
   2087    * method will mark numbers from British Crown dependencies such as the Isle of Man as invalid for
   2088    * the region "GB" (United Kingdom), since it has its own region code, "IM", which may be
   2089    * undesirable.
   2090    *
   2091    * @param number       the phone number that we want to validate
   2092    * @param regionCode   the region that we want to validate the phone number for
   2093    * @return  a boolean that indicates whether the number is of a valid pattern
   2094    */
   2095   public boolean isValidNumberForRegion(PhoneNumber number, String regionCode) {
   2096     int countryCode = number.getCountryCode();
   2097     PhoneMetadata metadata = getMetadataForRegionOrCallingCode(countryCode, regionCode);
   2098     if ((metadata == null) ||
   2099         (!REGION_CODE_FOR_NON_GEO_ENTITY.equals(regionCode) &&
   2100          countryCode != getCountryCodeForValidRegion(regionCode))) {
   2101       // Either the region code was invalid, or the country calling code for this number does not
   2102       // match that of the region code.
   2103       return false;
   2104     }
   2105     PhoneNumberDesc generalNumDesc = metadata.getGeneralDesc();
   2106     String nationalSignificantNumber = getNationalSignificantNumber(number);
   2107 
   2108     // For regions where we don't have metadata for PhoneNumberDesc, we treat any number passed in
   2109     // as a valid number if its national significant number is between the minimum and maximum
   2110     // lengths defined by ITU for a national significant number.
   2111     if (!generalNumDesc.hasNationalNumberPattern()) {
   2112       int numberLength = nationalSignificantNumber.length();
   2113       return numberLength > MIN_LENGTH_FOR_NSN && numberLength <= MAX_LENGTH_FOR_NSN;
   2114     }
   2115     return getNumberTypeHelper(nationalSignificantNumber, metadata) != PhoneNumberType.UNKNOWN;
   2116   }
   2117 
   2118   /**
   2119    * Returns the region where a phone number is from. This could be used for geocoding at the region
   2120    * level.
   2121    *
   2122    * @param number  the phone number whose origin we want to know
   2123    * @return  the region where the phone number is from, or null if no region matches this calling
   2124    *     code
   2125    */
   2126   public String getRegionCodeForNumber(PhoneNumber number) {
   2127     int countryCode = number.getCountryCode();
   2128     List<String> regions = countryCallingCodeToRegionCodeMap.get(countryCode);
   2129     if (regions == null) {
   2130       String numberString = getNationalSignificantNumber(number);
   2131       logger.log(Level.WARNING,
   2132                  "Missing/invalid country_code (" + countryCode + ") for number " + numberString);
   2133       return null;
   2134     }
   2135     if (regions.size() == 1) {
   2136       return regions.get(0);
   2137     } else {
   2138       return getRegionCodeForNumberFromRegionList(number, regions);
   2139     }
   2140   }
   2141 
   2142   private String getRegionCodeForNumberFromRegionList(PhoneNumber number,
   2143                                                       List<String> regionCodes) {
   2144     String nationalNumber = getNationalSignificantNumber(number);
   2145     for (String regionCode : regionCodes) {
   2146       // If leadingDigits is present, use this. Otherwise, do full validation.
   2147       // Metadata cannot be null because the region codes come from the country calling code map.
   2148       PhoneMetadata metadata = getMetadataForRegion(regionCode);
   2149       if (metadata.hasLeadingDigits()) {
   2150         if (regexCache.getPatternForRegex(metadata.getLeadingDigits())
   2151                 .matcher(nationalNumber).lookingAt()) {
   2152           return regionCode;
   2153         }
   2154       } else if (getNumberTypeHelper(nationalNumber, metadata) != PhoneNumberType.UNKNOWN) {
   2155         return regionCode;
   2156       }
   2157     }
   2158     return null;
   2159   }
   2160 
   2161   /**
   2162    * Returns the region code that matches the specific country calling code. In the case of no
   2163    * region code being found, ZZ will be returned. In the case of multiple regions, the one
   2164    * designated in the metadata as the "main" region for this calling code will be returned. If the
   2165    * countryCallingCode entered is valid but doesn't match a specific region (such as in the case of
   2166    * non-geographical calling codes like 800) the value "001" will be returned (corresponding to
   2167    * the value for World in the UN M.49 schema).
   2168    */
   2169   public String getRegionCodeForCountryCode(int countryCallingCode) {
   2170     List<String> regionCodes = countryCallingCodeToRegionCodeMap.get(countryCallingCode);
   2171     return regionCodes == null ? UNKNOWN_REGION : regionCodes.get(0);
   2172   }
   2173 
   2174   /**
   2175    * Returns a list with the region codes that match the specific country calling code. For
   2176    * non-geographical country calling codes, the region code 001 is returned. Also, in the case
   2177    * of no region code being found, an empty list is returned.
   2178    */
   2179   public List<String> getRegionCodesForCountryCode(int countryCallingCode) {
   2180     List<String> regionCodes = countryCallingCodeToRegionCodeMap.get(countryCallingCode);
   2181     return Collections.unmodifiableList(regionCodes == null ? new ArrayList<String>(0)
   2182                                                             : regionCodes);
   2183   }
   2184 
   2185   /**
   2186    * Returns the country calling code for a specific region. For example, this would be 1 for the
   2187    * United States, and 64 for New Zealand.
   2188    *
   2189    * @param regionCode  the region that we want to get the country calling code for
   2190    * @return  the country calling code for the region denoted by regionCode
   2191    */
   2192   public int getCountryCodeForRegion(String regionCode) {
   2193     if (!isValidRegionCode(regionCode)) {
   2194       logger.log(Level.WARNING,
   2195                  "Invalid or missing region code ("
   2196                   + ((regionCode == null) ? "null" : regionCode)
   2197                   + ") provided.");
   2198       return 0;
   2199     }
   2200     return getCountryCodeForValidRegion(regionCode);
   2201   }
   2202 
   2203   /**
   2204    * Returns the country calling code for a specific region. For example, this would be 1 for the
   2205    * United States, and 64 for New Zealand. Assumes the region is already valid.
   2206    *
   2207    * @param regionCode  the region that we want to get the country calling code for
   2208    * @return  the country calling code for the region denoted by regionCode
   2209    * @throws IllegalArgumentException if the region is invalid
   2210    */
   2211   private int getCountryCodeForValidRegion(String regionCode) {
   2212     PhoneMetadata metadata = getMetadataForRegion(regionCode);
   2213     if (metadata == null) {
   2214       throw new IllegalArgumentException("Invalid region code: " + regionCode);
   2215     }
   2216     return metadata.getCountryCode();
   2217   }
   2218 
   2219   /**
   2220    * Returns the national dialling prefix for a specific region. For example, this would be 1 for
   2221    * the United States, and 0 for New Zealand. Set stripNonDigits to true to strip symbols like "~"
   2222    * (which indicates a wait for a dialling tone) from the prefix returned. If no national prefix is
   2223    * present, we return null.
   2224    *
   2225    * <p>Warning: Do not use this method for do-your-own formatting - for some regions, the
   2226    * national dialling prefix is used only for certain types of numbers. Use the library's
   2227    * formatting functions to prefix the national prefix when required.
   2228    *
   2229    * @param regionCode  the region that we want to get the dialling prefix for
   2230    * @param stripNonDigits  true to strip non-digits from the national dialling prefix
   2231    * @return  the dialling prefix for the region denoted by regionCode
   2232    */
   2233   public String getNddPrefixForRegion(String regionCode, boolean stripNonDigits) {
   2234     PhoneMetadata metadata = getMetadataForRegion(regionCode);
   2235     if (metadata == null) {
   2236       logger.log(Level.WARNING,
   2237                  "Invalid or missing region code ("
   2238                   + ((regionCode == null) ? "null" : regionCode)
   2239                   + ") provided.");
   2240       return null;
   2241     }
   2242     String nationalPrefix = metadata.getNationalPrefix();
   2243     // If no national prefix was found, we return null.
   2244     if (nationalPrefix.length() == 0) {
   2245       return null;
   2246     }
   2247     if (stripNonDigits) {
   2248       // Note: if any other non-numeric symbols are ever used in national prefixes, these would have
   2249       // to be removed here as well.
   2250       nationalPrefix = nationalPrefix.replace("~", "");
   2251     }
   2252     return nationalPrefix;
   2253   }
   2254 
   2255   /**
   2256    * Checks if this is a region under the North American Numbering Plan Administration (NANPA).
   2257    *
   2258    * @return  true if regionCode is one of the regions under NANPA
   2259    */
   2260   public boolean isNANPACountry(String regionCode) {
   2261     return nanpaRegions.contains(regionCode);
   2262   }
   2263 
   2264   /**
   2265    * Checks whether the country calling code is from a region whose national significant number
   2266    * could contain a leading zero. An example of such a region is Italy. Returns false if no
   2267    * metadata for the country is found.
   2268    */
   2269   boolean isLeadingZeroPossible(int countryCallingCode) {
   2270     PhoneMetadata mainMetadataForCallingCode =
   2271         getMetadataForRegionOrCallingCode(countryCallingCode,
   2272                                           getRegionCodeForCountryCode(countryCallingCode));
   2273     if (mainMetadataForCallingCode == null) {
   2274       return false;
   2275     }
   2276     return mainMetadataForCallingCode.isLeadingZeroPossible();
   2277   }
   2278 
   2279   /**
   2280    * Checks if the number is a valid vanity (alpha) number such as 800 MICROSOFT. A valid vanity
   2281    * number will start with at least 3 digits and will have three or more alpha characters. This
   2282    * does not do region-specific checks - to work out if this number is actually valid for a region,
   2283    * it should be parsed and methods such as {@link #isPossibleNumberWithReason} and
   2284    * {@link #isValidNumber} should be used.
   2285    *
   2286    * @param number  the number that needs to be checked
   2287    * @return  true if the number is a valid vanity number
   2288    */
   2289   public boolean isAlphaNumber(String number) {
   2290     if (!isViablePhoneNumber(number)) {
   2291       // Number is too short, or doesn't match the basic phone number pattern.
   2292       return false;
   2293     }
   2294     StringBuilder strippedNumber = new StringBuilder(number);
   2295     maybeStripExtension(strippedNumber);
   2296     return VALID_ALPHA_PHONE_PATTERN.matcher(strippedNumber).matches();
   2297   }
   2298 
   2299   /**
   2300    * Convenience wrapper around {@link #isPossibleNumberWithReason}. Instead of returning the reason
   2301    * for failure, this method returns a boolean value.
   2302    * @param number  the number that needs to be checked
   2303    * @return  true if the number is possible
   2304    */
   2305   public boolean isPossibleNumber(PhoneNumber number) {
   2306     return isPossibleNumberWithReason(number) == ValidationResult.IS_POSSIBLE;
   2307   }
   2308 
   2309   /**
   2310    * Helper method to check a number against a particular pattern and determine whether it matches,
   2311    * or is too short or too long. Currently, if a number pattern suggests that numbers of length 7
   2312    * and 10 are possible, and a number in between these possible lengths is entered, such as of
   2313    * length 8, this will return TOO_LONG.
   2314    */
   2315   private ValidationResult testNumberLengthAgainstPattern(Pattern numberPattern, String number) {
   2316     Matcher numberMatcher = numberPattern.matcher(number);
   2317     if (numberMatcher.matches()) {
   2318       return ValidationResult.IS_POSSIBLE;
   2319     }
   2320     if (numberMatcher.lookingAt()) {
   2321       return ValidationResult.TOO_LONG;
   2322     } else {
   2323       return ValidationResult.TOO_SHORT;
   2324     }
   2325   }
   2326 
   2327   /**
   2328    * Helper method to check whether a number is too short to be a regular length phone number in a
   2329    * region.
   2330    */
   2331   private boolean isShorterThanPossibleNormalNumber(PhoneMetadata regionMetadata, String number) {
   2332     Pattern possibleNumberPattern = regexCache.getPatternForRegex(
   2333         regionMetadata.getGeneralDesc().getPossibleNumberPattern());
   2334     return testNumberLengthAgainstPattern(possibleNumberPattern, number) ==
   2335         ValidationResult.TOO_SHORT;
   2336   }
   2337 
   2338   /**
   2339    * Check whether a phone number is a possible number. It provides a more lenient check than
   2340    * {@link #isValidNumber} in the following sense:
   2341    *<ol>
   2342    * <li> It only checks the length of phone numbers. In particular, it doesn't check starting
   2343    *      digits of the number.
   2344    * <li> It doesn't attempt to figure out the type of the number, but uses general rules which
   2345    *      applies to all types of phone numbers in a region. Therefore, it is much faster than
   2346    *      isValidNumber.
   2347    * <li> For fixed line numbers, many regions have the concept of area code, which together with
   2348    *      subscriber number constitute the national significant number. It is sometimes okay to dial
   2349    *      the subscriber number only when dialing in the same area. This function will return
   2350    *      true if the subscriber-number-only version is passed in. On the other hand, because
   2351    *      isValidNumber validates using information on both starting digits (for fixed line
   2352    *      numbers, that would most likely be area codes) and length (obviously includes the
   2353    *      length of area codes for fixed line numbers), it will return false for the
   2354    *      subscriber-number-only version.
   2355    * </ol>
   2356    * @param number  the number that needs to be checked
   2357    * @return  a ValidationResult object which indicates whether the number is possible
   2358    */
   2359   public ValidationResult isPossibleNumberWithReason(PhoneNumber number) {
   2360     String nationalNumber = getNationalSignificantNumber(number);
   2361     int countryCode = number.getCountryCode();
   2362     // Note: For Russian Fed and NANPA numbers, we just use the rules from the default region (US or
   2363     // Russia) since the getRegionCodeForNumber will not work if the number is possible but not
   2364     // valid. This would need to be revisited if the possible number pattern ever differed between
   2365     // various regions within those plans.
   2366     if (!hasValidCountryCallingCode(countryCode)) {
   2367       return ValidationResult.INVALID_COUNTRY_CODE;
   2368     }
   2369     String regionCode = getRegionCodeForCountryCode(countryCode);
   2370     // Metadata cannot be null because the country calling code is valid.
   2371     PhoneMetadata metadata = getMetadataForRegionOrCallingCode(countryCode, regionCode);
   2372     PhoneNumberDesc generalNumDesc = metadata.getGeneralDesc();
   2373     // Handling case of numbers with no metadata.
   2374     if (!generalNumDesc.hasNationalNumberPattern()) {
   2375       logger.log(Level.FINER, "Checking if number is possible with incomplete metadata.");
   2376       int numberLength = nationalNumber.length();
   2377       if (numberLength < MIN_LENGTH_FOR_NSN) {
   2378         return ValidationResult.TOO_SHORT;
   2379       } else if (numberLength > MAX_LENGTH_FOR_NSN) {
   2380         return ValidationResult.TOO_LONG;
   2381       } else {
   2382         return ValidationResult.IS_POSSIBLE;
   2383       }
   2384     }
   2385     Pattern possibleNumberPattern =
   2386         regexCache.getPatternForRegex(generalNumDesc.getPossibleNumberPattern());
   2387     return testNumberLengthAgainstPattern(possibleNumberPattern, nationalNumber);
   2388   }
   2389 
   2390   /**
   2391    * Check whether a phone number is a possible number given a number in the form of a string, and
   2392    * the region where the number could be dialed from. It provides a more lenient check than
   2393    * {@link #isValidNumber}. See {@link #isPossibleNumber(PhoneNumber)} for details.
   2394    *
   2395    * <p>This method first parses the number, then invokes {@link #isPossibleNumber(PhoneNumber)}
   2396    * with the resultant PhoneNumber object.
   2397    *
   2398    * @param number  the number that needs to be checked, in the form of a string
   2399    * @param regionDialingFrom  the region that we are expecting the number to be dialed from.
   2400    *     Note this is different from the region where the number belongs.  For example, the number
   2401    *     +1 650 253 0000 is a number that belongs to US. When written in this form, it can be
   2402    *     dialed from any region. When it is written as 00 1 650 253 0000, it can be dialed from any
   2403    *     region which uses an international dialling prefix of 00. When it is written as
   2404    *     650 253 0000, it can only be dialed from within the US, and when written as 253 0000, it
   2405    *     can only be dialed from within a smaller area in the US (Mountain View, CA, to be more
   2406    *     specific).
   2407    * @return  true if the number is possible
   2408    */
   2409   public boolean isPossibleNumber(String number, String regionDialingFrom) {
   2410     try {
   2411       return isPossibleNumber(parse(number, regionDialingFrom));
   2412     } catch (NumberParseException e) {
   2413       return false;
   2414     }
   2415   }
   2416 
   2417   /**
   2418    * Attempts to extract a valid number from a phone number that is too long to be valid, and resets
   2419    * the PhoneNumber object passed in to that valid version. If no valid number could be extracted,
   2420    * the PhoneNumber object passed in will not be modified.
   2421    * @param number a PhoneNumber object which contains a number that is too long to be valid.
   2422    * @return  true if a valid phone number can be successfully extracted.
   2423    */
   2424   public boolean truncateTooLongNumber(PhoneNumber number) {
   2425     if (isValidNumber(number)) {
   2426       return true;
   2427     }
   2428     PhoneNumber numberCopy = new PhoneNumber();
   2429     numberCopy.mergeFrom(number);
   2430     long nationalNumber = number.getNationalNumber();
   2431     do {
   2432       nationalNumber /= 10;
   2433       numberCopy.setNationalNumber(nationalNumber);
   2434       if (isPossibleNumberWithReason(numberCopy) == ValidationResult.TOO_SHORT ||
   2435           nationalNumber == 0) {
   2436         return false;
   2437       }
   2438     } while (!isValidNumber(numberCopy));
   2439     number.setNationalNumber(nationalNumber);
   2440     return true;
   2441   }
   2442 
   2443   /**
   2444    * Gets an {@link com.android.i18n.phonenumbers.AsYouTypeFormatter} for the specific region.
   2445    *
   2446    * @param regionCode  the region where the phone number is being entered
   2447    * @return  an {@link com.android.i18n.phonenumbers.AsYouTypeFormatter} object, which can be used
   2448    *     to format phone numbers in the specific region "as you type"
   2449    */
   2450   public AsYouTypeFormatter getAsYouTypeFormatter(String regionCode) {
   2451     return new AsYouTypeFormatter(regionCode);
   2452   }
   2453 
   2454   // Extracts country calling code from fullNumber, returns it and places the remaining number in
   2455   // nationalNumber. It assumes that the leading plus sign or IDD has already been removed. Returns
   2456   // 0 if fullNumber doesn't start with a valid country calling code, and leaves nationalNumber
   2457   // unmodified.
   2458   int extractCountryCode(StringBuilder fullNumber, StringBuilder nationalNumber) {
   2459     if ((fullNumber.length() == 0) || (fullNumber.charAt(0) == '0')) {
   2460       // Country codes do not begin with a '0'.
   2461       return 0;
   2462     }
   2463     int potentialCountryCode;
   2464     int numberLength = fullNumber.length();
   2465     for (int i = 1; i <= MAX_LENGTH_COUNTRY_CODE && i <= numberLength; i++) {
   2466       potentialCountryCode = Integer.parseInt(fullNumber.substring(0, i));
   2467       if (countryCallingCodeToRegionCodeMap.containsKey(potentialCountryCode)) {
   2468         nationalNumber.append(fullNumber.substring(i));
   2469         return potentialCountryCode;
   2470       }
   2471     }
   2472     return 0;
   2473   }
   2474 
   2475   /**
   2476    * Tries to extract a country calling code from a number. This method will return zero if no
   2477    * country calling code is considered to be present. Country calling codes are extracted in the
   2478    * following ways:
   2479    * <ul>
   2480    *  <li> by stripping the international dialing prefix of the region the person is dialing from,
   2481    *       if this is present in the number, and looking at the next digits
   2482    *  <li> by stripping the '+' sign if present and then looking at the next digits
   2483    *  <li> by comparing the start of the number and the country calling code of the default region.
   2484    *       If the number is not considered possible for the numbering plan of the default region
   2485    *       initially, but starts with the country calling code of this region, validation will be
   2486    *       reattempted after stripping this country calling code. If this number is considered a
   2487    *       possible number, then the first digits will be considered the country calling code and
   2488    *       removed as such.
   2489    * </ul>
   2490    * It will throw a NumberParseException if the number starts with a '+' but the country calling
   2491    * code supplied after this does not match that of any known region.
   2492    *
   2493    * @param number  non-normalized telephone number that we wish to extract a country calling
   2494    *     code from - may begin with '+'
   2495    * @param defaultRegionMetadata  metadata about the region this number may be from
   2496    * @param nationalNumber  a string buffer to store the national significant number in, in the case
   2497    *     that a country calling code was extracted. The number is appended to any existing contents.
   2498    *     If no country calling code was extracted, this will be left unchanged.
   2499    * @param keepRawInput  true if the country_code_source and preferred_carrier_code fields of
   2500    *     phoneNumber should be populated.
   2501    * @param phoneNumber  the PhoneNumber object where the country_code and country_code_source need
   2502    *     to be populated. Note the country_code is always populated, whereas country_code_source is
   2503    *     only populated when keepCountryCodeSource is true.
   2504    * @return  the country calling code extracted or 0 if none could be extracted
   2505    */
   2506   // @VisibleForTesting
   2507   int maybeExtractCountryCode(String number, PhoneMetadata defaultRegionMetadata,
   2508                               StringBuilder nationalNumber, boolean keepRawInput,
   2509                               PhoneNumber phoneNumber)
   2510       throws NumberParseException {
   2511     if (number.length() == 0) {
   2512       return 0;
   2513     }
   2514     StringBuilder fullNumber = new StringBuilder(number);
   2515     // Set the default prefix to be something that will never match.
   2516     String possibleCountryIddPrefix = "NonMatch";
   2517     if (defaultRegionMetadata != null) {
   2518       possibleCountryIddPrefix = defaultRegionMetadata.getInternationalPrefix();
   2519     }
   2520 
   2521     CountryCodeSource countryCodeSource =
   2522         maybeStripInternationalPrefixAndNormalize(fullNumber, possibleCountryIddPrefix);
   2523     if (keepRawInput) {
   2524       phoneNumber.setCountryCodeSource(countryCodeSource);
   2525     }
   2526     if (countryCodeSource != CountryCodeSource.FROM_DEFAULT_COUNTRY) {
   2527       if (fullNumber.length() <= MIN_LENGTH_FOR_NSN) {
   2528         throw new NumberParseException(NumberParseException.ErrorType.TOO_SHORT_AFTER_IDD,
   2529                                        "Phone number had an IDD, but after this was not "
   2530                                        + "long enough to be a viable phone number.");
   2531       }
   2532       int potentialCountryCode = extractCountryCode(fullNumber, nationalNumber);
   2533       if (potentialCountryCode != 0) {
   2534         phoneNumber.setCountryCode(potentialCountryCode);
   2535         return potentialCountryCode;
   2536       }
   2537 
   2538       // If this fails, they must be using a strange country calling code that we don't recognize,
   2539       // or that doesn't exist.
   2540       throw new NumberParseException(NumberParseException.ErrorType.INVALID_COUNTRY_CODE,
   2541                                      "Country calling code supplied was not recognised.");
   2542     } else if (defaultRegionMetadata != null) {
   2543       // Check to see if the number starts with the country calling code for the default region. If
   2544       // so, we remove the country calling code, and do some checks on the validity of the number
   2545       // before and after.
   2546       int defaultCountryCode = defaultRegionMetadata.getCountryCode();
   2547       String defaultCountryCodeString = String.valueOf(defaultCountryCode);
   2548       String normalizedNumber = fullNumber.toString();
   2549       if (normalizedNumber.startsWith(defaultCountryCodeString)) {
   2550         StringBuilder potentialNationalNumber =
   2551             new StringBuilder(normalizedNumber.substring(defaultCountryCodeString.length()));
   2552         PhoneNumberDesc generalDesc = defaultRegionMetadata.getGeneralDesc();
   2553         Pattern validNumberPattern =
   2554             regexCache.getPatternForRegex(generalDesc.getNationalNumberPattern());
   2555         maybeStripNationalPrefixAndCarrierCode(
   2556             potentialNationalNumber, defaultRegionMetadata, null /* Don't need the carrier code */);
   2557         Pattern possibleNumberPattern =
   2558             regexCache.getPatternForRegex(generalDesc.getPossibleNumberPattern());
   2559         // If the number was not valid before but is valid now, or if it was too long before, we
   2560         // consider the number with the country calling code stripped to be a better result and
   2561         // keep that instead.
   2562         if ((!validNumberPattern.matcher(fullNumber).matches() &&
   2563              validNumberPattern.matcher(potentialNationalNumber).matches()) ||
   2564              testNumberLengthAgainstPattern(possibleNumberPattern, fullNumber.toString())
   2565                   == ValidationResult.TOO_LONG) {
   2566           nationalNumber.append(potentialNationalNumber);
   2567           if (keepRawInput) {
   2568             phoneNumber.setCountryCodeSource(CountryCodeSource.FROM_NUMBER_WITHOUT_PLUS_SIGN);
   2569           }
   2570           phoneNumber.setCountryCode(defaultCountryCode);
   2571           return defaultCountryCode;
   2572         }
   2573       }
   2574     }
   2575     // No country calling code present.
   2576     phoneNumber.setCountryCode(0);
   2577     return 0;
   2578   }
   2579 
   2580   /**
   2581    * Strips the IDD from the start of the number if present. Helper function used by
   2582    * maybeStripInternationalPrefixAndNormalize.
   2583    */
   2584   private boolean parsePrefixAsIdd(Pattern iddPattern, StringBuilder number) {
   2585     Matcher m = iddPattern.matcher(number);
   2586     if (m.lookingAt()) {
   2587       int matchEnd = m.end();
   2588       // Only strip this if the first digit after the match is not a 0, since country calling codes
   2589       // cannot begin with 0.
   2590       Matcher digitMatcher = CAPTURING_DIGIT_PATTERN.matcher(number.substring(matchEnd));
   2591       if (digitMatcher.find()) {
   2592         String normalizedGroup = normalizeDigitsOnly(digitMatcher.group(1));
   2593         if (normalizedGroup.equals("0")) {
   2594           return false;
   2595         }
   2596       }
   2597       number.delete(0, matchEnd);
   2598       return true;
   2599     }
   2600     return false;
   2601   }
   2602 
   2603   /**
   2604    * Strips any international prefix (such as +, 00, 011) present in the number provided, normalizes
   2605    * the resulting number, and indicates if an international prefix was present.
   2606    *
   2607    * @param number  the non-normalized telephone number that we wish to strip any international
   2608    *     dialing prefix from.
   2609    * @param possibleIddPrefix  the international direct dialing prefix from the region we
   2610    *     think this number may be dialed in
   2611    * @return  the corresponding CountryCodeSource if an international dialing prefix could be
   2612    *     removed from the number, otherwise CountryCodeSource.FROM_DEFAULT_COUNTRY if the number did
   2613    *     not seem to be in international format.
   2614    */
   2615   // @VisibleForTesting
   2616   CountryCodeSource maybeStripInternationalPrefixAndNormalize(
   2617       StringBuilder number,
   2618       String possibleIddPrefix) {
   2619     if (number.length() == 0) {
   2620       return CountryCodeSource.FROM_DEFAULT_COUNTRY;
   2621     }
   2622     // Check to see if the number begins with one or more plus signs.
   2623     Matcher m = PLUS_CHARS_PATTERN.matcher(number);
   2624     if (m.lookingAt()) {
   2625       number.delete(0, m.end());
   2626       // Can now normalize the rest of the number since we've consumed the "+" sign at the start.
   2627       normalize(number);
   2628       return CountryCodeSource.FROM_NUMBER_WITH_PLUS_SIGN;
   2629     }
   2630     // Attempt to parse the first digits as an international prefix.
   2631     Pattern iddPattern = regexCache.getPatternForRegex(possibleIddPrefix);
   2632     normalize(number);
   2633     return parsePrefixAsIdd(iddPattern, number)
   2634            ? CountryCodeSource.FROM_NUMBER_WITH_IDD
   2635            : CountryCodeSource.FROM_DEFAULT_COUNTRY;
   2636   }
   2637 
   2638   /**
   2639    * Strips any national prefix (such as 0, 1) present in the number provided.
   2640    *
   2641    * @param number  the normalized telephone number that we wish to strip any national
   2642    *     dialing prefix from
   2643    * @param metadata  the metadata for the region that we think this number is from
   2644    * @param carrierCode  a place to insert the carrier code if one is extracted
   2645    * @return true if a national prefix or carrier code (or both) could be extracted.
   2646    */
   2647   // @VisibleForTesting
   2648   boolean maybeStripNationalPrefixAndCarrierCode(
   2649       StringBuilder number, PhoneMetadata metadata, StringBuilder carrierCode) {
   2650     int numberLength = number.length();
   2651     String possibleNationalPrefix = metadata.getNationalPrefixForParsing();
   2652     if (numberLength == 0 || possibleNationalPrefix.length() == 0) {
   2653       // Early return for numbers of zero length.
   2654       return false;
   2655     }
   2656     // Attempt to parse the first digits as a national prefix.
   2657     Matcher prefixMatcher = regexCache.getPatternForRegex(possibleNationalPrefix).matcher(number);
   2658     if (prefixMatcher.lookingAt()) {
   2659       Pattern nationalNumberRule =
   2660           regexCache.getPatternForRegex(metadata.getGeneralDesc().getNationalNumberPattern());
   2661       // Check if the original number is viable.
   2662       boolean isViableOriginalNumber = nationalNumberRule.matcher(number).matches();
   2663       // prefixMatcher.group(numOfGroups) == null implies nothing was captured by the capturing
   2664       // groups in possibleNationalPrefix; therefore, no transformation is necessary, and we just
   2665       // remove the national prefix.
   2666       int numOfGroups = prefixMatcher.groupCount();
   2667       String transformRule = metadata.getNationalPrefixTransformRule();
   2668       if (transformRule == null || transformRule.length() == 0 ||
   2669           prefixMatcher.group(numOfGroups) == null) {
   2670         // If the original number was viable, and the resultant number is not, we return.
   2671         if (isViableOriginalNumber &&
   2672             !nationalNumberRule.matcher(number.substring(prefixMatcher.end())).matches()) {
   2673           return false;
   2674         }
   2675         if (carrierCode != null && numOfGroups > 0 && prefixMatcher.group(numOfGroups) != null) {
   2676           carrierCode.append(prefixMatcher.group(1));
   2677         }
   2678         number.delete(0, prefixMatcher.end());
   2679         return true;
   2680       } else {
   2681         // Check that the resultant number is still viable. If not, return. Check this by copying
   2682         // the string buffer and making the transformation on the copy first.
   2683         StringBuilder transformedNumber = new StringBuilder(number);
   2684         transformedNumber.replace(0, numberLength, prefixMatcher.replaceFirst(transformRule));
   2685         if (isViableOriginalNumber &&
   2686             !nationalNumberRule.matcher(transformedNumber.toString()).matches()) {
   2687           return false;
   2688         }
   2689         if (carrierCode != null && numOfGroups > 1) {
   2690           carrierCode.append(prefixMatcher.group(1));
   2691         }
   2692         number.replace(0, number.length(), transformedNumber.toString());
   2693         return true;
   2694       }
   2695     }
   2696     return false;
   2697   }
   2698 
   2699   /**
   2700    * Strips any extension (as in, the part of the number dialled after the call is connected,
   2701    * usually indicated with extn, ext, x or similar) from the end of the number, and returns it.
   2702    *
   2703    * @param number  the non-normalized telephone number that we wish to strip the extension from
   2704    * @return        the phone extension
   2705    */
   2706   // @VisibleForTesting
   2707   String maybeStripExtension(StringBuilder number) {
   2708     Matcher m = EXTN_PATTERN.matcher(number);
   2709     // If we find a potential extension, and the number preceding this is a viable number, we assume
   2710     // it is an extension.
   2711     if (m.find() && isViablePhoneNumber(number.substring(0, m.start()))) {
   2712       // The numbers are captured into groups in the regular expression.
   2713       for (int i = 1, length = m.groupCount(); i <= length; i++) {
   2714         if (m.group(i) != null) {
   2715           // We go through the capturing groups until we find one that captured some digits. If none
   2716           // did, then we will return the empty string.
   2717           String extension = m.group(i);
   2718           number.delete(m.start(), number.length());
   2719           return extension;
   2720         }
   2721       }
   2722     }
   2723     return "";
   2724   }
   2725 
   2726   /**
   2727    * Checks to see that the region code used is valid, or if it is not valid, that the number to
   2728    * parse starts with a + symbol so that we can attempt to infer the region from the number.
   2729    * Returns false if it cannot use the region provided and the region cannot be inferred.
   2730    */
   2731   private boolean checkRegionForParsing(String numberToParse, String defaultRegion) {
   2732     if (!isValidRegionCode(defaultRegion)) {
   2733       // If the number is null or empty, we can't infer the region.
   2734       if (numberToParse == null || numberToParse.length() == 0 ||
   2735           !PLUS_CHARS_PATTERN.matcher(numberToParse).lookingAt()) {
   2736         return false;
   2737       }
   2738     }
   2739     return true;
   2740   }
   2741 
   2742   /**
   2743    * Parses a string and returns it in proto buffer format. This method will throw a
   2744    * {@link com.android.i18n.phonenumbers.NumberParseException} if the number is not considered to be
   2745    * a possible number. Note that validation of whether the number is actually a valid number for a
   2746    * particular region is not performed. This can be done separately with {@link #isValidNumber}.
   2747    *
   2748    * @param numberToParse     number that we are attempting to parse. This can contain formatting
   2749    *                          such as +, ( and -, as well as a phone number extension. It can also
   2750    *                          be provided in RFC3966 format.
   2751    * @param defaultRegion     region that we are expecting the number to be from. This is only used
   2752    *                          if the number being parsed is not written in international format.
   2753    *                          The country_code for the number in this case would be stored as that
   2754    *                          of the default region supplied. If the number is guaranteed to
   2755    *                          start with a '+' followed by the country calling code, then
   2756    *                          "ZZ" or null can be supplied.
   2757    * @return                  a phone number proto buffer filled with the parsed number
   2758    * @throws NumberParseException  if the string is not considered to be a viable phone number or if
   2759    *                               no default region was supplied and the number is not in
   2760    *                               international format (does not start with +)
   2761    */
   2762   public PhoneNumber parse(String numberToParse, String defaultRegion)
   2763       throws NumberParseException {
   2764     PhoneNumber phoneNumber = new PhoneNumber();
   2765     parse(numberToParse, defaultRegion, phoneNumber);
   2766     return phoneNumber;
   2767   }
   2768 
   2769   /**
   2770    * Same as {@link #parse(String, String)}, but accepts mutable PhoneNumber as a parameter to
   2771    * decrease object creation when invoked many times.
   2772    */
   2773   public void parse(String numberToParse, String defaultRegion, PhoneNumber phoneNumber)
   2774       throws NumberParseException {
   2775     parseHelper(numberToParse, defaultRegion, false, true, phoneNumber);
   2776   }
   2777 
   2778   /**
   2779    * Parses a string and returns it in proto buffer format. This method differs from {@link #parse}
   2780    * in that it always populates the raw_input field of the protocol buffer with numberToParse as
   2781    * well as the country_code_source field.
   2782    *
   2783    * @param numberToParse     number that we are attempting to parse. This can contain formatting
   2784    *                          such as +, ( and -, as well as a phone number extension.
   2785    * @param defaultRegion     region that we are expecting the number to be from. This is only used
   2786    *                          if the number being parsed is not written in international format.
   2787    *                          The country calling code for the number in this case would be stored
   2788    *                          as that of the default region supplied.
   2789    * @return                  a phone number proto buffer filled with the parsed number
   2790    * @throws NumberParseException  if the string is not considered to be a viable phone number or if
   2791    *                               no default region was supplied
   2792    */
   2793   public PhoneNumber parseAndKeepRawInput(String numberToParse, String defaultRegion)
   2794       throws NumberParseException {
   2795     PhoneNumber phoneNumber = new PhoneNumber();
   2796     parseAndKeepRawInput(numberToParse, defaultRegion, phoneNumber);
   2797     return phoneNumber;
   2798   }
   2799 
   2800   /**
   2801    * Same as{@link #parseAndKeepRawInput(String, String)}, but accepts a mutable PhoneNumber as
   2802    * a parameter to decrease object creation when invoked many times.
   2803    */
   2804   public void parseAndKeepRawInput(String numberToParse, String defaultRegion,
   2805                                    PhoneNumber phoneNumber)
   2806       throws NumberParseException {
   2807     parseHelper(numberToParse, defaultRegion, true, true, phoneNumber);
   2808   }
   2809 
   2810   /**
   2811    * Returns an iterable over all {@link PhoneNumberMatch PhoneNumberMatches} in {@code text}. This
   2812    * is a shortcut for {@link #findNumbers(CharSequence, String, Leniency, long)
   2813    * getMatcher(text, defaultRegion, Leniency.VALID, Long.MAX_VALUE)}.
   2814    *
   2815    * @param text              the text to search for phone numbers, null for no text
   2816    * @param defaultRegion     region that we are expecting the number to be from. This is only used
   2817    *                          if the number being parsed is not written in international format. The
   2818    *                          country_code for the number in this case would be stored as that of
   2819    *                          the default region supplied. May be null if only international
   2820    *                          numbers are expected.
   2821    */
   2822   public Iterable<PhoneNumberMatch> findNumbers(CharSequence text, String defaultRegion) {
   2823     return findNumbers(text, defaultRegion, Leniency.VALID, Long.MAX_VALUE);
   2824   }
   2825 
   2826   /**
   2827    * Returns an iterable over all {@link PhoneNumberMatch PhoneNumberMatches} in {@code text}.
   2828    *
   2829    * @param text              the text to search for phone numbers, null for no text
   2830    * @param defaultRegion     region that we are expecting the number to be from. This is only used
   2831    *                          if the number being parsed is not written in international format. The
   2832    *                          country_code for the number in this case would be stored as that of
   2833    *                          the default region supplied. May be null if only international
   2834    *                          numbers are expected.
   2835    * @param leniency          the leniency to use when evaluating candidate phone numbers
   2836    * @param maxTries          the maximum number of invalid numbers to try before giving up on the
   2837    *                          text. This is to cover degenerate cases where the text has a lot of
   2838    *                          false positives in it. Must be {@code >= 0}.
   2839    */
   2840   public Iterable<PhoneNumberMatch> findNumbers(
   2841       final CharSequence text, final String defaultRegion, final Leniency leniency,
   2842       final long maxTries) {
   2843 
   2844     return new Iterable<PhoneNumberMatch>() {
   2845       public Iterator<PhoneNumberMatch> iterator() {
   2846         return new PhoneNumberMatcher(
   2847             PhoneNumberUtil.this, text, defaultRegion, leniency, maxTries);
   2848       }
   2849     };
   2850   }
   2851 
   2852   /**
   2853    * A helper function to set the values related to leading zeros in a PhoneNumber.
   2854    */
   2855   static void setItalianLeadingZerosForPhoneNumber(String nationalNumber, PhoneNumber phoneNumber) {
   2856     if (nationalNumber.length() > 1 && nationalNumber.charAt(0) == '0') {
   2857       phoneNumber.setItalianLeadingZero(true);
   2858       int numberOfLeadingZeros = 1;
   2859       // Note that if the national number is all "0"s, the last "0" is not counted as a leading
   2860       // zero.
   2861       while (numberOfLeadingZeros < nationalNumber.length() - 1 &&
   2862              nationalNumber.charAt(numberOfLeadingZeros) == '0') {
   2863         numberOfLeadingZeros++;
   2864       }
   2865       if (numberOfLeadingZeros != 1) {
   2866         phoneNumber.setNumberOfLeadingZeros(numberOfLeadingZeros);
   2867       }
   2868     }
   2869   }
   2870 
   2871   /**
   2872    * Parses a string and fills up the phoneNumber. This method is the same as the public
   2873    * parse() method, with the exception that it allows the default region to be null, for use by
   2874    * isNumberMatch(). checkRegion should be set to false if it is permitted for the default region
   2875    * to be null or unknown ("ZZ").
   2876    */
   2877   private void parseHelper(String numberToParse, String defaultRegion, boolean keepRawInput,
   2878                            boolean checkRegion, PhoneNumber phoneNumber)
   2879       throws NumberParseException {
   2880     if (numberToParse == null) {
   2881       throw new NumberParseException(NumberParseException.ErrorType.NOT_A_NUMBER,
   2882                                      "The phone number supplied was null.");
   2883     } else if (numberToParse.length() > MAX_INPUT_STRING_LENGTH) {
   2884       throw new NumberParseException(NumberParseException.ErrorType.TOO_LONG,
   2885                                      "The string supplied was too long to parse.");
   2886     }
   2887 
   2888     StringBuilder nationalNumber = new StringBuilder();
   2889     buildNationalNumberForParsing(numberToParse, nationalNumber);
   2890 
   2891     if (!isViablePhoneNumber(nationalNumber.toString())) {
   2892       throw new NumberParseException(NumberParseException.ErrorType.NOT_A_NUMBER,
   2893                                      "The string supplied did not seem to be a phone number.");
   2894     }
   2895 
   2896     // Check the region supplied is valid, or that the extracted number starts with some sort of +
   2897     // sign so the number's region can be determined.
   2898     if (checkRegion && !checkRegionForParsing(nationalNumber.toString(), defaultRegion)) {
   2899       throw new NumberParseException(NumberParseException.ErrorType.INVALID_COUNTRY_CODE,
   2900                                      "Missing or invalid default region.");
   2901     }
   2902 
   2903     if (keepRawInput) {
   2904       phoneNumber.setRawInput(numberToParse);
   2905     }
   2906     // Attempt to parse extension first, since it doesn't require region-specific data and we want
   2907     // to have the non-normalised number here.
   2908     String extension = maybeStripExtension(nationalNumber);
   2909     if (extension.length() > 0) {
   2910       phoneNumber.setExtension(extension);
   2911     }
   2912 
   2913     PhoneMetadata regionMetadata = getMetadataForRegion(defaultRegion);
   2914     // Check to see if the number is given in international format so we know whether this number is
   2915     // from the default region or not.
   2916     StringBuilder normalizedNationalNumber = new StringBuilder();
   2917     int countryCode = 0;
   2918     try {
   2919       // TODO: This method should really just take in the string buffer that has already
   2920       // been created, and just remove the prefix, rather than taking in a string and then
   2921       // outputting a string buffer.
   2922       countryCode = maybeExtractCountryCode(nationalNumber.toString(), regionMetadata,
   2923                                             normalizedNationalNumber, keepRawInput, phoneNumber);
   2924     } catch (NumberParseException e) {
   2925       Matcher matcher = PLUS_CHARS_PATTERN.matcher(nationalNumber.toString());
   2926       if (e.getErrorType() == NumberParseException.ErrorType.INVALID_COUNTRY_CODE &&
   2927           matcher.lookingAt()) {
   2928         // Strip the plus-char, and try again.
   2929         countryCode = maybeExtractCountryCode(nationalNumber.substring(matcher.end()),
   2930                                               regionMetadata, normalizedNationalNumber,
   2931                                               keepRawInput, phoneNumber);
   2932         if (countryCode == 0) {
   2933           throw new NumberParseException(NumberParseException.ErrorType.INVALID_COUNTRY_CODE,
   2934                                          "Could not interpret numbers after plus-sign.");
   2935         }
   2936       } else {
   2937         throw new NumberParseException(e.getErrorType(), e.getMessage());
   2938       }
   2939     }
   2940     if (countryCode != 0) {
   2941       String phoneNumberRegion = getRegionCodeForCountryCode(countryCode);
   2942       if (!phoneNumberRegion.equals(defaultRegion)) {
   2943         // Metadata cannot be null because the country calling code is valid.
   2944         regionMetadata = getMetadataForRegionOrCallingCode(countryCode, phoneNumberRegion);
   2945       }
   2946     } else {
   2947       // If no extracted country calling code, use the region supplied instead. The national number
   2948       // is just the normalized version of the number we were given to parse.
   2949       normalize(nationalNumber);
   2950       normalizedNationalNumber.append(nationalNumber);
   2951       if (defaultRegion != null) {
   2952         countryCode = regionMetadata.getCountryCode();
   2953         phoneNumber.setCountryCode(countryCode);
   2954       } else if (keepRawInput) {
   2955         phoneNumber.clearCountryCodeSource();
   2956       }
   2957     }
   2958     if (normalizedNationalNumber.length() < MIN_LENGTH_FOR_NSN) {
   2959       throw new NumberParseException(NumberParseException.ErrorType.TOO_SHORT_NSN,
   2960                                      "The string supplied is too short to be a phone number.");
   2961     }
   2962     if (regionMetadata != null) {
   2963       StringBuilder carrierCode = new StringBuilder();
   2964       StringBuilder potentialNationalNumber = new StringBuilder(normalizedNationalNumber);
   2965       maybeStripNationalPrefixAndCarrierCode(potentialNationalNumber, regionMetadata, carrierCode);
   2966       // We require that the NSN remaining after stripping the national prefix and carrier code be
   2967       // of a possible length for the region. Otherwise, we don't do the stripping, since the
   2968       // original number could be a valid short number.
   2969       if (!isShorterThanPossibleNormalNumber(regionMetadata, potentialNationalNumber.toString())) {
   2970         normalizedNationalNumber = potentialNationalNumber;
   2971         if (keepRawInput) {
   2972           phoneNumber.setPreferredDomesticCarrierCode(carrierCode.toString());
   2973         }
   2974       }
   2975     }
   2976     int lengthOfNationalNumber = normalizedNationalNumber.length();
   2977     if (lengthOfNationalNumber < MIN_LENGTH_FOR_NSN) {
   2978       throw new NumberParseException(NumberParseException.ErrorType.TOO_SHORT_NSN,
   2979                                      "The string supplied is too short to be a phone number.");
   2980     }
   2981     if (lengthOfNationalNumber > MAX_LENGTH_FOR_NSN) {
   2982       throw new NumberParseException(NumberParseException.ErrorType.TOO_LONG,
   2983                                      "The string supplied is too long to be a phone number.");
   2984     }
   2985     setItalianLeadingZerosForPhoneNumber(normalizedNationalNumber.toString(), phoneNumber);
   2986     phoneNumber.setNationalNumber(Long.parseLong(normalizedNationalNumber.toString()));
   2987   }
   2988 
   2989   /**
   2990    * Converts numberToParse to a form that we can parse and write it to nationalNumber if it is
   2991    * written in RFC3966; otherwise extract a possible number out of it and write to nationalNumber.
   2992    */
   2993   private void buildNationalNumberForParsing(String numberToParse, StringBuilder nationalNumber) {
   2994     int indexOfPhoneContext = numberToParse.indexOf(RFC3966_PHONE_CONTEXT);
   2995     if (indexOfPhoneContext > 0) {
   2996       int phoneContextStart = indexOfPhoneContext + RFC3966_PHONE_CONTEXT.length();
   2997       // If the phone context contains a phone number prefix, we need to capture it, whereas domains
   2998       // will be ignored.
   2999       if (numberToParse.charAt(phoneContextStart) == PLUS_SIGN) {
   3000         // Additional parameters might follow the phone context. If so, we will remove them here
   3001         // because the parameters after phone context are not important for parsing the
   3002         // phone number.
   3003         int phoneContextEnd = numberToParse.indexOf(';', phoneContextStart);
   3004         if (phoneContextEnd > 0) {
   3005           nationalNumber.append(numberToParse.substring(phoneContextStart, phoneContextEnd));
   3006         } else {
   3007           nationalNumber.append(numberToParse.substring(phoneContextStart));
   3008         }
   3009       }
   3010 
   3011       // Now append everything between the "tel:" prefix and the phone-context. This should include
   3012       // the national number, an optional extension or isdn-subaddress component.
   3013       nationalNumber.append(numberToParse.substring(
   3014           numberToParse.indexOf(RFC3966_PREFIX) + RFC3966_PREFIX.length(), indexOfPhoneContext));
   3015     } else {
   3016       // Extract a possible number from the string passed in (this strips leading characters that
   3017       // could not be the start of a phone number.)
   3018       nationalNumber.append(extractPossibleNumber(numberToParse));
   3019     }
   3020 
   3021     // Delete the isdn-subaddress and everything after it if it is present. Note extension won't
   3022     // appear at the same time with isdn-subaddress according to paragraph 5.3 of the RFC3966 spec,
   3023     int indexOfIsdn = nationalNumber.indexOf(RFC3966_ISDN_SUBADDRESS);
   3024     if (indexOfIsdn > 0) {
   3025       nationalNumber.delete(indexOfIsdn, nationalNumber.length());
   3026     }
   3027     // If both phone context and isdn-subaddress are absent but other parameters are present, the
   3028     // parameters are left in nationalNumber. This is because we are concerned about deleting
   3029     // content from a potential number string when there is no strong evidence that the number is
   3030     // actually written in RFC3966.
   3031   }
   3032 
   3033   /**
   3034    * Takes two phone numbers and compares them for equality.
   3035    *
   3036    * <p>Returns EXACT_MATCH if the country_code, NSN, presence of a leading zero for Italian numbers
   3037    * and any extension present are the same.
   3038    * Returns NSN_MATCH if either or both has no region specified, and the NSNs and extensions are
   3039    * the same.
   3040    * Returns SHORT_NSN_MATCH if either or both has no region specified, or the region specified is
   3041    * the same, and one NSN could be a shorter version of the other number. This includes the case
   3042    * where one has an extension specified, and the other does not.
   3043    * Returns NO_MATCH otherwise.
   3044    * For example, the numbers +1 345 657 1234 and 657 1234 are a SHORT_NSN_MATCH.
   3045    * The numbers +1 345 657 1234 and 345 657 are a NO_MATCH.
   3046    *
   3047    * @param firstNumberIn  first number to compare
   3048    * @param secondNumberIn  second number to compare
   3049    *
   3050    * @return  NO_MATCH, SHORT_NSN_MATCH, NSN_MATCH or EXACT_MATCH depending on the level of equality
   3051    *     of the two numbers, described in the method definition.
   3052    */
   3053   public MatchType isNumberMatch(PhoneNumber firstNumberIn, PhoneNumber secondNumberIn) {
   3054     // Make copies of the phone number so that the numbers passed in are not edited.
   3055     PhoneNumber firstNumber = new PhoneNumber();
   3056     firstNumber.mergeFrom(firstNumberIn);
   3057     PhoneNumber secondNumber = new PhoneNumber();
   3058     secondNumber.mergeFrom(secondNumberIn);
   3059     // First clear raw_input, country_code_source and preferred_domestic_carrier_code fields and any
   3060     // empty-string extensions so that we can use the proto-buffer equality method.
   3061     firstNumber.clearRawInput();
   3062     firstNumber.clearCountryCodeSource();
   3063     firstNumber.clearPreferredDomesticCarrierCode();
   3064     secondNumber.clearRawInput();
   3065     secondNumber.clearCountryCodeSource();
   3066     secondNumber.clearPreferredDomesticCarrierCode();
   3067     if (firstNumber.hasExtension() &&
   3068         firstNumber.getExtension().length() == 0) {
   3069         firstNumber.clearExtension();
   3070     }
   3071     if (secondNumber.hasExtension() &&
   3072         secondNumber.getExtension().length() == 0) {
   3073         secondNumber.clearExtension();
   3074     }
   3075     // Early exit if both had extensions and these are different.
   3076     if (firstNumber.hasExtension() && secondNumber.hasExtension() &&
   3077         !firstNumber.getExtension().equals(secondNumber.getExtension())) {
   3078       return MatchType.NO_MATCH;
   3079     }
   3080     int firstNumberCountryCode = firstNumber.getCountryCode();
   3081     int secondNumberCountryCode = secondNumber.getCountryCode();
   3082     // Both had country_code specified.
   3083     if (firstNumberCountryCode != 0 && secondNumberCountryCode != 0) {
   3084       if (firstNumber.exactlySameAs(secondNumber)) {
   3085         return MatchType.EXACT_MATCH;
   3086       } else if (firstNumberCountryCode == secondNumberCountryCode &&
   3087                  isNationalNumberSuffixOfTheOther(firstNumber, secondNumber)) {
   3088         // A SHORT_NSN_MATCH occurs if there is a difference because of the presence or absence of
   3089         // an 'Italian leading zero', the presence or absence of an extension, or one NSN being a
   3090         // shorter variant of the other.
   3091         return MatchType.SHORT_NSN_MATCH;
   3092       }
   3093       // This is not a match.
   3094       return MatchType.NO_MATCH;
   3095     }
   3096     // Checks cases where one or both country_code fields were not specified. To make equality
   3097     // checks easier, we first set the country_code fields to be equal.
   3098     firstNumber.setCountryCode(secondNumberCountryCode);
   3099     // If all else was the same, then this is an NSN_MATCH.
   3100     if (firstNumber.exactlySameAs(secondNumber)) {
   3101       return MatchType.NSN_MATCH;
   3102     }
   3103     if (isNationalNumberSuffixOfTheOther(firstNumber, secondNumber)) {
   3104       return MatchType.SHORT_NSN_MATCH;
   3105     }
   3106     return MatchType.NO_MATCH;
   3107   }
   3108 
   3109   // Returns true when one national number is the suffix of the other or both are the same.
   3110   private boolean isNationalNumberSuffixOfTheOther(PhoneNumber firstNumber,
   3111                                                    PhoneNumber secondNumber) {
   3112     String firstNumberNationalNumber = String.valueOf(firstNumber.getNationalNumber());
   3113     String secondNumberNationalNumber = String.valueOf(secondNumber.getNationalNumber());
   3114     // Note that endsWith returns true if the numbers are equal.
   3115     return firstNumberNationalNumber.endsWith(secondNumberNationalNumber) ||
   3116            secondNumberNationalNumber.endsWith(firstNumberNationalNumber);
   3117   }
   3118 
   3119   /**
   3120    * Takes two phone numbers as strings and compares them for equality. This is a convenience
   3121    * wrapper for {@link #isNumberMatch(PhoneNumber, PhoneNumber)}. No default region is known.
   3122    *
   3123    * @param firstNumber  first number to compare. Can contain formatting, and can have country
   3124    *     calling code specified with + at the start.
   3125    * @param secondNumber  second number to compare. Can contain formatting, and can have country
   3126    *     calling code specified with + at the start.
   3127    * @return  NOT_A_NUMBER, NO_MATCH, SHORT_NSN_MATCH, NSN_MATCH, EXACT_MATCH. See
   3128    *     {@link #isNumberMatch(PhoneNumber, PhoneNumber)} for more details.
   3129    */
   3130   public MatchType isNumberMatch(String firstNumber, String secondNumber) {
   3131     try {
   3132       PhoneNumber firstNumberAsProto = parse(firstNumber, UNKNOWN_REGION);
   3133       return isNumberMatch(firstNumberAsProto, secondNumber);
   3134     } catch (NumberParseException e) {
   3135       if (e.getErrorType() == NumberParseException.ErrorType.INVALID_COUNTRY_CODE) {
   3136         try {
   3137           PhoneNumber secondNumberAsProto = parse(secondNumber, UNKNOWN_REGION);
   3138           return isNumberMatch(secondNumberAsProto, firstNumber);
   3139         } catch (NumberParseException e2) {
   3140           if (e2.getErrorType() == NumberParseException.ErrorType.INVALID_COUNTRY_CODE) {
   3141             try {
   3142               PhoneNumber firstNumberProto = new PhoneNumber();
   3143               PhoneNumber secondNumberProto = new PhoneNumber();
   3144               parseHelper(firstNumber, null, false, false, firstNumberProto);
   3145               parseHelper(secondNumber, null, false, false, secondNumberProto);
   3146               return isNumberMatch(firstNumberProto, secondNumberProto);
   3147             } catch (NumberParseException e3) {
   3148               // Fall through and return MatchType.NOT_A_NUMBER.
   3149             }
   3150           }
   3151         }
   3152       }
   3153     }
   3154     // One or more of the phone numbers we are trying to match is not a viable phone number.
   3155     return MatchType.NOT_A_NUMBER;
   3156   }
   3157 
   3158   /**
   3159    * Takes two phone numbers and compares them for equality. This is a convenience wrapper for
   3160    * {@link #isNumberMatch(PhoneNumber, PhoneNumber)}. No default region is known.
   3161    *
   3162    * @param firstNumber  first number to compare in proto buffer format.
   3163    * @param secondNumber  second number to compare. Can contain formatting, and can have country
   3164    *     calling code specified with + at the start.
   3165    * @return  NOT_A_NUMBER, NO_MATCH, SHORT_NSN_MATCH, NSN_MATCH, EXACT_MATCH. See
   3166    *     {@link #isNumberMatch(PhoneNumber, PhoneNumber)} for more details.
   3167    */
   3168   public MatchType isNumberMatch(PhoneNumber firstNumber, String secondNumber) {
   3169     // First see if the second number has an implicit country calling code, by attempting to parse
   3170     // it.
   3171     try {
   3172       PhoneNumber secondNumberAsProto = parse(secondNumber, UNKNOWN_REGION);
   3173       return isNumberMatch(firstNumber, secondNumberAsProto);
   3174     } catch (NumberParseException e) {
   3175       if (e.getErrorType() == NumberParseException.ErrorType.INVALID_COUNTRY_CODE) {
   3176         // The second number has no country calling code. EXACT_MATCH is no longer possible.
   3177         // We parse it as if the region was the same as that for the first number, and if
   3178         // EXACT_MATCH is returned, we replace this with NSN_MATCH.
   3179         String firstNumberRegion = getRegionCodeForCountryCode(firstNumber.getCountryCode());
   3180         try {
   3181           if (!firstNumberRegion.equals(UNKNOWN_REGION)) {
   3182             PhoneNumber secondNumberWithFirstNumberRegion = parse(secondNumber, firstNumberRegion);
   3183             MatchType match = isNumberMatch(firstNumber, secondNumberWithFirstNumberRegion);
   3184             if (match == MatchType.EXACT_MATCH) {
   3185               return MatchType.NSN_MATCH;
   3186             }
   3187             return match;
   3188           } else {
   3189             // If the first number didn't have a valid country calling code, then we parse the
   3190             // second number without one as well.
   3191             PhoneNumber secondNumberProto = new PhoneNumber();
   3192             parseHelper(secondNumber, null, false, false, secondNumberProto);
   3193             return isNumberMatch(firstNumber, secondNumberProto);
   3194           }
   3195         } catch (NumberParseException e2) {
   3196           // Fall-through to return NOT_A_NUMBER.
   3197         }
   3198       }
   3199     }
   3200     // One or more of the phone numbers we are trying to match is not a viable phone number.
   3201     return MatchType.NOT_A_NUMBER;
   3202   }
   3203 
   3204   /**
   3205    * Returns true if the number can be dialled from outside the region, or unknown. If the number
   3206    * can only be dialled from within the region, returns false. Does not check the number is a valid
   3207    * number.
   3208    * TODO: Make this method public when we have enough metadata to make it worthwhile.
   3209    *
   3210    * @param number  the phone-number for which we want to know whether it is diallable from
   3211    *     outside the region
   3212    */
   3213   // @VisibleForTesting
   3214   boolean canBeInternationallyDialled(PhoneNumber number) {
   3215     PhoneMetadata metadata = getMetadataForRegion(getRegionCodeForNumber(number));
   3216     if (metadata == null) {
   3217       // Note numbers belonging to non-geographical entities (e.g. +800 numbers) are always
   3218       // internationally diallable, and will be caught here.
   3219       return true;
   3220     }
   3221     String nationalSignificantNumber = getNationalSignificantNumber(number);
   3222     return !isNumberMatchingDesc(nationalSignificantNumber, metadata.getNoInternationalDialling());
   3223   }
   3224 
   3225   /**
   3226    * Returns true if the supplied region supports mobile number portability. Returns false for
   3227    * invalid, unknown or regions that don't support mobile number portability.
   3228    *
   3229    * @param regionCode  the region for which we want to know whether it supports mobile number
   3230    *                    portability or not.
   3231    */
   3232   public boolean isMobileNumberPortableRegion(String regionCode) {
   3233     PhoneMetadata metadata = getMetadataForRegion(regionCode);
   3234     if (metadata == null) {
   3235       logger.log(Level.WARNING, "Invalid or unknown region code provided: " + regionCode);
   3236       return false;
   3237     }
   3238     return metadata.isMobileNumberPortableRegion();
   3239   }
   3240 }
   3241