• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
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.ObjectInputStream;
29 import java.util.ArrayList;
30 import java.util.Arrays;
31 import java.util.Collections;
32 import java.util.HashMap;
33 import java.util.HashSet;
34 import java.util.Iterator;
35 import java.util.List;
36 import java.util.Map;
37 import java.util.Set;
38 import java.util.logging.Level;
39 import java.util.logging.Logger;
40 import java.util.regex.Matcher;
41 import java.util.regex.Pattern;
42 
43 /**
44  * Utility for international phone numbers. Functionality includes formatting, parsing and
45  * validation.
46  *
47  * <p>If you use this library, and want to be notified about important changes, please sign up to
48  * our <a href="http://groups.google.com/group/libphonenumber-discuss/about">mailing list</a>.
49  *
50  * NOTE: A lot of methods in this class require Region Code strings. These must be provided using
51  * ISO 3166-1 two-letter country-code format. These should be in upper-case. The list of the codes
52  * can be found here: http://www.iso.org/iso/english_country_names_and_code_elements
53  *
54  * @author Shaopeng Jia
55  * @author Lara Rennie
56  */
57 public class PhoneNumberUtil {
58   /** Flags to use when compiling regular expressions for phone numbers. */
59   static final int REGEX_FLAGS = Pattern.UNICODE_CASE | Pattern.CASE_INSENSITIVE;
60   // The minimum and maximum length of the national significant number.
61   private static final int MIN_LENGTH_FOR_NSN = 3;
62   // The ITU says the maximum length should be 15, but we have found longer numbers in Germany.
63   static final int MAX_LENGTH_FOR_NSN = 16;
64   // The maximum length of the country calling code.
65   static final int MAX_LENGTH_COUNTRY_CODE = 3;
66   // We don't allow input strings for parsing to be longer than 250 chars. This prevents malicious
67   // input from overflowing the regular-expression engine.
68   private static final int MAX_INPUT_STRING_LENGTH = 250;
69   static final String META_DATA_FILE_PREFIX =
70       "/com/android/i18n/phonenumbers/data/PhoneNumberMetadataProto";
71   private String currentFilePrefix = META_DATA_FILE_PREFIX;
72   private static final Logger LOGGER = Logger.getLogger(PhoneNumberUtil.class.getName());
73 
74   // A mapping from a country calling code to the region codes which denote the region represented
75   // by that country calling code. In the case of multiple regions sharing a calling code, such as
76   // the NANPA regions, the one indicated with "isMainCountryForCode" in the metadata should be
77   // first.
78   private Map<Integer, List<String>> countryCallingCodeToRegionCodeMap = null;
79 
80   // The set of regions the library supports.
81   // There are roughly 240 of them and we set the initial capacity of the HashSet to 320 to offer a
82   // load factor of roughly 0.75.
83   private final Set<String> supportedRegions = new HashSet<String>(320);
84 
85   // Region-code for the unknown region.
86   private static final String UNKNOWN_REGION = "ZZ";
87 
88   // The set of regions that share country calling code 1.
89   // There are roughly 26 regions and we set the initial capacity of the HashSet to 35 to offer a
90   // load factor of roughly 0.75.
91   private final Set<String> nanpaRegions = new HashSet<String>(35);
92   private static final int NANPA_COUNTRY_CODE = 1;
93 
94   // The prefix that needs to be inserted in front of a Colombian landline number when dialed from
95   // a mobile phone in Colombia.
96   private static final String COLOMBIA_MOBILE_TO_FIXED_LINE_PREFIX = "3";
97 
98   // The PLUS_SIGN signifies the international prefix.
99   static final char PLUS_SIGN = '+';
100 
101   private static final char STAR_SIGN = '*';
102 
103   private static final String RFC3966_EXTN_PREFIX = ";ext=";
104 
105   // A map that contains characters that are essential when dialling. That means any of the
106   // characters in this map must not be removed from a number when dialing, otherwise the call will
107   // not reach the intended destination.
108   private static final Map<Character, Character> DIALLABLE_CHAR_MAPPINGS;
109 
110   // Only upper-case variants of alpha characters are stored.
111   private static final Map<Character, Character> ALPHA_MAPPINGS;
112 
113   // For performance reasons, amalgamate both into one map.
114   private static final Map<Character, Character> ALPHA_PHONE_MAPPINGS;
115 
116   // Separate map of all symbols that we wish to retain when formatting alpha numbers. This
117   // includes digits, ASCII letters and number grouping symbols such as "-" and " ".
118   private static final Map<Character, Character> ALL_PLUS_NUMBER_GROUPING_SYMBOLS;
119 
120   static {
121     // Simple ASCII digits map used to populate ALPHA_PHONE_MAPPINGS and
122     // ALL_PLUS_NUMBER_GROUPING_SYMBOLS.
123     HashMap<Character, Character> asciiDigitMappings = new HashMap<Character, Character>();
124     asciiDigitMappings.put('0', '0');
125     asciiDigitMappings.put('1', '1');
126     asciiDigitMappings.put('2', '2');
127     asciiDigitMappings.put('3', '3');
128     asciiDigitMappings.put('4', '4');
129     asciiDigitMappings.put('5', '5');
130     asciiDigitMappings.put('6', '6');
131     asciiDigitMappings.put('7', '7');
132     asciiDigitMappings.put('8', '8');
133     asciiDigitMappings.put('9', '9');
134 
135     HashMap<Character, Character> alphaMap = new HashMap<Character, Character>(40);
136     alphaMap.put('A', '2');
137     alphaMap.put('B', '2');
138     alphaMap.put('C', '2');
139     alphaMap.put('D', '3');
140     alphaMap.put('E', '3');
141     alphaMap.put('F', '3');
142     alphaMap.put('G', '4');
143     alphaMap.put('H', '4');
144     alphaMap.put('I', '4');
145     alphaMap.put('J', '5');
146     alphaMap.put('K', '5');
147     alphaMap.put('L', '5');
148     alphaMap.put('M', '6');
149     alphaMap.put('N', '6');
150     alphaMap.put('O', '6');
151     alphaMap.put('P', '7');
152     alphaMap.put('Q', '7');
153     alphaMap.put('R', '7');
154     alphaMap.put('S', '7');
155     alphaMap.put('T', '8');
156     alphaMap.put('U', '8');
157     alphaMap.put('V', '8');
158     alphaMap.put('W', '9');
159     alphaMap.put('X', '9');
160     alphaMap.put('Y', '9');
161     alphaMap.put('Z', '9');
162     ALPHA_MAPPINGS = Collections.unmodifiableMap(alphaMap);
163 
164     HashMap<Character, Character> combinedMap = new HashMap<Character, Character>(100);
165     combinedMap.putAll(ALPHA_MAPPINGS);
166     combinedMap.putAll(asciiDigitMappings);
167     ALPHA_PHONE_MAPPINGS = Collections.unmodifiableMap(combinedMap);
168 
169     HashMap<Character, Character> diallableCharMap = new HashMap<Character, Character>();
170     diallableCharMap.putAll(asciiDigitMappings);
171     diallableCharMap.put('+', '+');
172     diallableCharMap.put('*', '*');
173     DIALLABLE_CHAR_MAPPINGS = Collections.unmodifiableMap(diallableCharMap);
174 
175     HashMap<Character, Character> allPlusNumberGroupings = new HashMap<Character, Character>();
176     // Put (lower letter -> upper letter) and (upper letter -> upper letter) mappings.
177     for (char c : ALPHA_MAPPINGS.keySet()) {
Character.toLowerCase(c)178       allPlusNumberGroupings.put(Character.toLowerCase(c), c);
allPlusNumberGroupings.put(c, c)179       allPlusNumberGroupings.put(c, c);
180     }
181     allPlusNumberGroupings.putAll(asciiDigitMappings);
182     // Put grouping symbols.
183     allPlusNumberGroupings.put('-', '-');
184     allPlusNumberGroupings.put('\uFF0D', '-');
185     allPlusNumberGroupings.put('\u2010', '-');
186     allPlusNumberGroupings.put('\u2011', '-');
187     allPlusNumberGroupings.put('\u2012', '-');
188     allPlusNumberGroupings.put('\u2013', '-');
189     allPlusNumberGroupings.put('\u2014', '-');
190     allPlusNumberGroupings.put('\u2015', '-');
191     allPlusNumberGroupings.put('\u2212', '-');
192     allPlusNumberGroupings.put('/', '/');
193     allPlusNumberGroupings.put('\uFF0F', '/');
194     allPlusNumberGroupings.put(' ', ' ');
195     allPlusNumberGroupings.put('\u3000', ' ');
196     allPlusNumberGroupings.put('\u2060', ' ');
197     allPlusNumberGroupings.put('.', '.');
198     allPlusNumberGroupings.put('\uFF0E', '.');
199     ALL_PLUS_NUMBER_GROUPING_SYMBOLS = Collections.unmodifiableMap(allPlusNumberGroupings);
200   }
201 
202   // Pattern that makes it easy to distinguish whether a region has a unique international dialing
203   // prefix or not. If a region has a unique international prefix (e.g. 011 in USA), it will be
204   // represented as a string that contains a sequence of ASCII digits. If there are multiple
205   // available international prefixes in a region, they will be represented as a regex string that
206   // always contains character(s) other than ASCII digits.
207   // Note this regex also includes tilde, which signals waiting for the tone.
208   private static final Pattern UNIQUE_INTERNATIONAL_PREFIX =
209       Pattern.compile("[\\d]+(?:[~\u2053\u223C\uFF5E][\\d]+)?");
210 
211   // Regular expression of acceptable punctuation found in phone numbers. This excludes punctuation
212   // found as a leading character only.
213   // This consists of dash characters, white space characters, full stops, slashes,
214   // square brackets, parentheses and tildes. It also includes the letter 'x' as that is found as a
215   // placeholder for carrier information in some phone numbers. Full-width variants are also
216   // present.
217   static final String VALID_PUNCTUATION = "-x\u2010-\u2015\u2212\u30FC\uFF0D-\uFF0F " +
218       "\u00A0\u200B\u2060\u3000()\uFF08\uFF09\uFF3B\uFF3D.\\[\\]/~\u2053\u223C\uFF5E";
219 
220   private static final String DIGITS = "\\p{Nd}";
221   // We accept alpha characters in phone numbers, ASCII only, upper and lower case.
222   private static final String VALID_ALPHA =
223       Arrays.toString(ALPHA_MAPPINGS.keySet().toArray()).replaceAll("[, \\[\\]]", "") +
224       Arrays.toString(ALPHA_MAPPINGS.keySet().toArray()).toLowerCase().replaceAll("[, \\[\\]]", "");
225   static final String PLUS_CHARS = "+\uFF0B";
226   static final Pattern PLUS_CHARS_PATTERN = Pattern.compile("[" + PLUS_CHARS + "]+");
227   private static final Pattern SEPARATOR_PATTERN = Pattern.compile("[" + VALID_PUNCTUATION + "]+");
228   private static final Pattern CAPTURING_DIGIT_PATTERN = Pattern.compile("(" + DIGITS + ")");
229 
230   // Regular expression of acceptable characters that may start a phone number for the purposes of
231   // parsing. This allows us to strip away meaningless prefixes to phone numbers that may be
232   // mistakenly given to us. This consists of digits, the plus symbol and arabic-indic digits. This
233   // does not contain alpha characters, although they may be used later in the number. It also does
234   // not include other punctuation, as this will be stripped later during parsing and is of no
235   // information value when parsing a number.
236   private static final String VALID_START_CHAR = "[" + PLUS_CHARS + DIGITS + "]";
237   private static final Pattern VALID_START_CHAR_PATTERN = Pattern.compile(VALID_START_CHAR);
238 
239   // Regular expression of characters typically used to start a second phone number for the purposes
240   // of parsing. This allows us to strip off parts of the number that are actually the start of
241   // another number, such as for: (530) 583-6985 x302/x2303 -> the second extension here makes this
242   // actually two phone numbers, (530) 583-6985 x302 and (530) 583-6985 x2303. We remove the second
243   // extension so that the first number is parsed correctly.
244   private static final String SECOND_NUMBER_START = "[\\\\/] *x";
245   static final Pattern SECOND_NUMBER_START_PATTERN = Pattern.compile(SECOND_NUMBER_START);
246 
247   // Regular expression of trailing characters that we want to remove. We remove all characters that
248   // are not alpha or numerical characters. The hash character is retained here, as it may signify
249   // the previous block was an extension.
250   private static final String UNWANTED_END_CHARS = "[[\\P{N}&&\\P{L}]&&[^#]]+$";
251   static final Pattern UNWANTED_END_CHAR_PATTERN = Pattern.compile(UNWANTED_END_CHARS);
252 
253   // We use this pattern to check if the phone number has at least three letters in it - if so, then
254   // we treat it as a number where some phone-number digits are represented by letters.
255   private static final Pattern VALID_ALPHA_PHONE_PATTERN = Pattern.compile("(?:.*?[A-Za-z]){3}.*");
256 
257   // Regular expression of viable phone numbers. This is location independent. Checks we have at
258   // least three leading digits, and only valid punctuation, alpha characters and
259   // digits in the phone number. Does not include extension data.
260   // The symbol 'x' is allowed here as valid punctuation since it is often used as a placeholder for
261   // carrier codes, for example in Brazilian phone numbers. We also allow multiple "+" characters at
262   // the start.
263   // Corresponds to the following:
264   // plus_sign*(([punctuation]|[star])*[digits]){3,}([punctuation]|[star]|[digits]|[alpha])*
265   // Note VALID_PUNCTUATION starts with a -, so must be the first in the range.
266   private static final String VALID_PHONE_NUMBER =
267       "[" + PLUS_CHARS + "]*+(?:[" + VALID_PUNCTUATION + STAR_SIGN + "]*" + DIGITS + "){3,}[" +
268       VALID_PUNCTUATION + STAR_SIGN + VALID_ALPHA + DIGITS + "]*";
269 
270   // Default extension prefix to use when formatting. This will be put in front of any extension
271   // component of the number, after the main national number is formatted. For example, if you wish
272   // the default extension formatting to be " extn: 3456", then you should specify " extn: " here
273   // as the default extension prefix. This can be overridden by region-specific preferences.
274   private static final String DEFAULT_EXTN_PREFIX = " ext. ";
275 
276   // Pattern to capture digits used in an extension. Places a maximum length of "7" for an
277   // extension.
278   private static final String CAPTURING_EXTN_DIGITS = "(" + DIGITS + "{1,7})";
279   // Regexp of all possible ways to write extensions, for use when parsing. This will be run as a
280   // case-insensitive regexp match. Wide character versions are also provided after each ASCII
281   // version.
282   private static final String EXTN_PATTERNS_FOR_PARSING;
283   static final String EXTN_PATTERNS_FOR_MATCHING;
284   static {
285     // One-character symbols that can be used to indicate an extension.
286     String singleExtnSymbolsForMatching = "x\uFF58#\uFF03~\uFF5E";
287     // For parsing, we are slightly more lenient in our interpretation than for matching. Here we
288     // allow a "comma" as a possible extension indicator. When matching, this is hardly ever used to
289     // indicate this.
290     String singleExtnSymbolsForParsing = "," + singleExtnSymbolsForMatching;
291 
292     EXTN_PATTERNS_FOR_PARSING = createExtnPattern(singleExtnSymbolsForParsing);
293     EXTN_PATTERNS_FOR_MATCHING = createExtnPattern(singleExtnSymbolsForMatching);
294   }
295 
296   /**
297    * Helper initialiser method to create the regular-expression pattern to match extensions,
298    * allowing the one-char extension symbols provided by {@code singleExtnSymbols}.
299    */
createExtnPattern(String singleExtnSymbols)300   private static String createExtnPattern(String singleExtnSymbols) {
301     // There are three regular expressions here. The first covers RFC 3966 format, where the
302     // extension is added using ";ext=". The second more generic one starts with optional white
303     // space and ends with an optional full stop (.), followed by zero or more spaces/tabs and then
304     // the numbers themselves. The other one covers the special case of American numbers where the
305     // extension is written with a hash at the end, such as "- 503#".
306     // Note that the only capturing groups should be around the digits that you want to capture as
307     // part of the extension, or else parsing will fail!
308     // Canonical-equivalence doesn't seem to be an option with Android java, so we allow two options
309     // for representing the accented o - the character itself, and one in the unicode decomposed
310     // form with the combining acute accent.
311     return (RFC3966_EXTN_PREFIX + CAPTURING_EXTN_DIGITS + "|" + "[ \u00A0\\t,]*" +
312             "(?:e?xt(?:ensi(?:o\u0301?|\u00F3))?n?|\uFF45?\uFF58\uFF54\uFF4E?|" +
313             "[" + singleExtnSymbols + "]|int|anexo|\uFF49\uFF4E\uFF54)" +
314             "[:\\.\uFF0E]?[ \u00A0\\t,-]*" + CAPTURING_EXTN_DIGITS + "#?|" +
315             "[- ]+(" + DIGITS + "{1,5})#");
316   }
317 
318   // Regexp of all known extension prefixes used by different regions followed by 1 or more valid
319   // digits, for use when parsing.
320   private static final Pattern EXTN_PATTERN =
321       Pattern.compile("(?:" + EXTN_PATTERNS_FOR_PARSING + ")$", REGEX_FLAGS);
322 
323   // We append optionally the extension pattern to the end here, as a valid phone number may
324   // have an extension prefix appended, followed by 1 or more digits.
325   private static final Pattern VALID_PHONE_NUMBER_PATTERN =
326       Pattern.compile(VALID_PHONE_NUMBER + "(?:" + EXTN_PATTERNS_FOR_PARSING + ")?", REGEX_FLAGS);
327 
328   private static final Pattern NON_DIGITS_PATTERN = Pattern.compile("(\\D+)");
329 
330   // The FIRST_GROUP_PATTERN was originally set to $1 but there are some countries for which the
331   // first group is not used in the national pattern (e.g. Argentina) so the $1 group does not match
332   // correctly.  Therefore, we use \d, so that the first group actually used in the pattern will be
333   // matched.
334   private static final Pattern FIRST_GROUP_PATTERN = Pattern.compile("(\\$\\d)");
335   private static final Pattern NP_PATTERN = Pattern.compile("\\$NP");
336   private static final Pattern FG_PATTERN = Pattern.compile("\\$FG");
337   private static final Pattern CC_PATTERN = Pattern.compile("\\$CC");
338 
339   private static PhoneNumberUtil instance = null;
340 
341   // A mapping from a region code to the PhoneMetadata for that region.
342   private final Map<String, PhoneMetadata> regionToMetadataMap =
343       Collections.synchronizedMap(new HashMap<String, PhoneMetadata>());
344 
345   // A mapping from a country calling code for a non-geographical entity to the PhoneMetadata for
346   // that country calling code. Examples of the country calling codes include 800 (International
347   // Toll Free Service) and 808 (International Shared Cost Service).
348   private final Map<Integer, PhoneMetadata> countryCodeToNonGeographicalMetadataMap =
349       Collections.synchronizedMap(new HashMap<Integer, PhoneMetadata>());
350 
351   // A cache for frequently used region-specific regular expressions.
352   // As most people use phone numbers primarily from one to two countries, and there are roughly 60
353   // regular expressions needed, the initial capacity of 100 offers a rough load factor of 0.75.
354   private RegexCache regexCache = new RegexCache(100);
355 
356   public static final String REGION_CODE_FOR_NON_GEO_ENTITY = "001";
357 
358   /**
359    * INTERNATIONAL and NATIONAL formats are consistent with the definition in ITU-T Recommendation
360    * E123. For example, the number of the Google Switzerland office will be written as
361    * "+41 44 668 1800" in INTERNATIONAL format, and as "044 668 1800" in NATIONAL format.
362    * E164 format is as per INTERNATIONAL format but with no formatting applied, e.g.
363    * "+41446681800". RFC3966 is as per INTERNATIONAL format, but with all spaces and other
364    * separating symbols replaced with a hyphen, and with any phone number extension appended with
365    * ";ext=". It also will have a prefix of "tel:" added, e.g. "tel:+41-44-668-1800".
366    *
367    * Note: If you are considering storing the number in a neutral format, you are highly advised to
368    * use the PhoneNumber class.
369    */
370   public enum PhoneNumberFormat {
371     E164,
372     INTERNATIONAL,
373     NATIONAL,
374     RFC3966
375   }
376 
377   /**
378    * Type of phone numbers.
379    */
380   public enum PhoneNumberType {
381     FIXED_LINE,
382     MOBILE,
383     // In some regions (e.g. the USA), it is impossible to distinguish between fixed-line and
384     // mobile numbers by looking at the phone number itself.
385     FIXED_LINE_OR_MOBILE,
386     // Freephone lines
387     TOLL_FREE,
388     PREMIUM_RATE,
389     // The cost of this call is shared between the caller and the recipient, and is hence typically
390     // less than PREMIUM_RATE calls. See // http://en.wikipedia.org/wiki/Shared_Cost_Service for
391     // more information.
392     SHARED_COST,
393     // Voice over IP numbers. This includes TSoIP (Telephony Service over IP).
394     VOIP,
395     // A personal number is associated with a particular person, and may be routed to either a
396     // MOBILE or FIXED_LINE number. Some more information can be found here:
397     // http://en.wikipedia.org/wiki/Personal_Numbers
398     PERSONAL_NUMBER,
399     PAGER,
400     // Used for "Universal Access Numbers" or "Company Numbers". They may be further routed to
401     // specific offices, but allow one number to be used for a company.
402     UAN,
403     // Used for "Voice Mail Access Numbers".
404     VOICEMAIL,
405     // A phone number is of type UNKNOWN when it does not fit any of the known patterns for a
406     // specific region.
407     UNKNOWN
408   }
409 
410   /**
411    * Types of phone number matches. See detailed description beside the isNumberMatch() method.
412    */
413   public enum MatchType {
414     NOT_A_NUMBER,
415     NO_MATCH,
416     SHORT_NSN_MATCH,
417     NSN_MATCH,
418     EXACT_MATCH,
419   }
420 
421   /**
422    * Possible outcomes when testing if a PhoneNumber is possible.
423    */
424   public enum ValidationResult {
425     IS_POSSIBLE,
426     INVALID_COUNTRY_CODE,
427     TOO_SHORT,
428     TOO_LONG,
429   }
430 
431   /**
432    * Leniency when {@linkplain PhoneNumberUtil#findNumbers finding} potential phone numbers in text
433    * segments. The levels here are ordered in increasing strictness.
434    */
435   public enum Leniency {
436     /**
437      * Phone numbers accepted are {@linkplain PhoneNumberUtil#isPossibleNumber(PhoneNumber)
438      * possible}, but not necessarily {@linkplain PhoneNumberUtil#isValidNumber(PhoneNumber) valid}.
439      */
440     POSSIBLE {
441       @Override
verify(PhoneNumber number, String candidate, PhoneNumberUtil util)442       boolean verify(PhoneNumber number, String candidate, PhoneNumberUtil util) {
443         return util.isPossibleNumber(number);
444       }
445     },
446     /**
447      * Phone numbers accepted are {@linkplain PhoneNumberUtil#isPossibleNumber(PhoneNumber)
448      * possible} and {@linkplain PhoneNumberUtil#isValidNumber(PhoneNumber) valid}. Numbers written
449      * in national format must have their national-prefix present if it is usually written for a
450      * number of this type.
451      */
452     VALID {
453       @Override
verify(PhoneNumber number, String candidate, PhoneNumberUtil util)454       boolean verify(PhoneNumber number, String candidate, PhoneNumberUtil util) {
455         if (!util.isValidNumber(number) ||
456             !containsOnlyValidXChars(number, candidate, util)) {
457           return false;
458         }
459         return isNationalPrefixPresentIfRequired(number, util);
460       }
461     },
462     /**
463      * Phone numbers accepted are {@linkplain PhoneNumberUtil#isValidNumber(PhoneNumber) valid} and
464      * are grouped in a possible way for this locale. For example, a US number written as
465      * "65 02 53 00 00" and "650253 0000" are not accepted at this leniency level, whereas
466      * "650 253 0000", "650 2530000" or "6502530000" are.
467      * Numbers with more than one '/' symbol are also dropped at this level.
468      * <p>
469      * Warning: This level might result in lower coverage especially for regions outside of country
470      * code "+1". If you are not sure about which level to use, email the discussion group
471      * libphonenumber-discuss@googlegroups.com.
472      */
473     STRICT_GROUPING {
474       @Override
verify(PhoneNumber number, String candidate, PhoneNumberUtil util)475       boolean verify(PhoneNumber number, String candidate, PhoneNumberUtil util) {
476         if (!util.isValidNumber(number) ||
477             !containsOnlyValidXChars(number, candidate, util) ||
478             containsMoreThanOneSlash(candidate) ||
479             !isNationalPrefixPresentIfRequired(number, util)) {
480           return false;
481         }
482         // TODO: Evaluate how this works for other locales (testing has been
483         // limited to NANPA regions) and optimise if necessary.
484         String[] formattedNumberGroups = getNationalNumberGroups(util, number);
485         StringBuilder normalizedCandidate = normalizeDigits(candidate,
486                                                             true /* keep strip non-digits */);
487         int fromIndex = 0;
488         // Check each group of consecutive digits are not broken into separate groups in the
489         // {@code candidate} string.
490         for (int i = 0; i < formattedNumberGroups.length; i++) {
491           // Fails if the substring of {@code candidate} starting from {@code fromIndex} doesn't
492           // contain the consecutive digits in formattedNumberGroups[i].
493           fromIndex = normalizedCandidate.indexOf(formattedNumberGroups[i], fromIndex);
494           if (fromIndex < 0) {
495             return false;
496           }
497           // Moves {@code fromIndex} forward.
498           fromIndex += formattedNumberGroups[i].length();
499           if (i == 0 && fromIndex < normalizedCandidate.length()) {
500             // We are at the position right after the NDC.
501             if (Character.isDigit(normalizedCandidate.charAt(fromIndex))) {
502               // This means there is no formatting symbol after the NDC. In this case, we only
503               // accept the number if there is no formatting symbol at all in the number, except
504               // for extensions.
505               String nationalSignificantNumber = util.getNationalSignificantNumber(number);
506               return normalizedCandidate.substring(fromIndex - formattedNumberGroups[i].length())
507                   .startsWith(nationalSignificantNumber);
508             }
509           }
510         }
511         // The check here makes sure that we haven't mistakenly already used the extension to
512         // match the last group of the subscriber number. Note the extension cannot have
513         // formatting in-between digits.
514         return normalizedCandidate.substring(fromIndex).contains(number.getExtension());
515       }
516     },
517     /**
518      * Phone numbers accepted are {@linkplain PhoneNumberUtil#isValidNumber(PhoneNumber) valid} and
519      * are grouped in the same way that we would have formatted it, or as a single block. For
520      * example, a US number written as "650 2530000" is not accepted at this leniency level, whereas
521      * "650 253 0000" or "6502530000" are.
522      * Numbers with more than one '/' symbol are also dropped at this level.
523      * <p>
524      * Warning: This level might result in lower coverage especially for regions outside of country
525      * code "+1". If you are not sure about which level to use, email the discussion group
526      * libphonenumber-discuss@googlegroups.com.
527      */
528     EXACT_GROUPING {
529       @Override
verify(PhoneNumber number, String candidate, PhoneNumberUtil util)530       boolean verify(PhoneNumber number, String candidate, PhoneNumberUtil util) {
531         if (!util.isValidNumber(number) ||
532             !containsOnlyValidXChars(number, candidate, util) ||
533             containsMoreThanOneSlash(candidate) ||
534             !isNationalPrefixPresentIfRequired(number, util)) {
535           return false;
536         }
537         // TODO: Evaluate how this works for other locales (testing has been
538         // limited to NANPA regions) and optimise if necessary.
539         StringBuilder normalizedCandidate = normalizeDigits(candidate,
540                                                             true /* keep strip non-digits */);
541         String[] candidateGroups =
542             NON_DIGITS_PATTERN.split(normalizedCandidate.toString());
543         // Set this to the last group, skipping it if the number has an extension.
544         int candidateNumberGroupIndex =
545             number.hasExtension() ? candidateGroups.length - 2 : candidateGroups.length - 1;
546         // First we check if the national significant number is formatted as a block.
547         // We use contains and not equals, since the national significant number may be present with
548         // a prefix such as a national number prefix, or the country code itself.
549         if (candidateGroups.length == 1 ||
550             candidateGroups[candidateNumberGroupIndex].contains(
551                 util.getNationalSignificantNumber(number))) {
552           return true;
553         }
554         String[] formattedNumberGroups = getNationalNumberGroups(util, number);
555         // Starting from the end, go through in reverse, excluding the first group, and check the
556         // candidate and number groups are the same.
557         for (int formattedNumberGroupIndex = (formattedNumberGroups.length - 1);
558              formattedNumberGroupIndex > 0 && candidateNumberGroupIndex >= 0;
559              formattedNumberGroupIndex--, candidateNumberGroupIndex--) {
560           if (!candidateGroups[candidateNumberGroupIndex].equals(
561               formattedNumberGroups[formattedNumberGroupIndex])) {
562             return false;
563           }
564         }
565         // Now check the first group. There may be a national prefix at the start, so we only check
566         // that the candidate group ends with the formatted number group.
567         return (candidateNumberGroupIndex >= 0 &&
568                 candidateGroups[candidateNumberGroupIndex].endsWith(formattedNumberGroups[0]));
569       }
570     };
571 
572     /**
573      * Helper method to get the national-number part of a number, formatted without any national
574      * prefix, and return it as a set of digit blocks that would be formatted together.
575      */
getNationalNumberGroups(PhoneNumberUtil util, PhoneNumber number)576     private static String[] getNationalNumberGroups(PhoneNumberUtil util, PhoneNumber number) {
577       // This will be in the format +CC-DG;ext=EXT where DG represents groups of digits.
578       String rfc3966Format = util.format(number, PhoneNumberFormat.RFC3966);
579       // We remove the extension part from the formatted string before splitting it into different
580       // groups.
581       int endIndex = rfc3966Format.indexOf(';');
582       if (endIndex < 0) {
583         endIndex = rfc3966Format.length();
584       }
585       // The country-code will have a '-' following it.
586       int startIndex = rfc3966Format.indexOf('-') + 1;
587       return rfc3966Format.substring(startIndex, endIndex).split("-");
588     }
589 
containsMoreThanOneSlash(String candidate)590     private static boolean containsMoreThanOneSlash(String candidate) {
591       int firstSlashIndex = candidate.indexOf('/');
592       return (firstSlashIndex > 0 && candidate.substring(firstSlashIndex + 1).contains("/"));
593     }
594 
containsOnlyValidXChars( PhoneNumber number, String candidate, PhoneNumberUtil util)595     private static boolean containsOnlyValidXChars(
596         PhoneNumber number, String candidate, PhoneNumberUtil util) {
597       // The characters 'x' and 'X' can be (1) a carrier code, in which case they always precede the
598       // national significant number or (2) an extension sign, in which case they always precede the
599       // extension number. We assume a carrier code is more than 1 digit, so the first case has to
600       // have more than 1 consecutive 'x' or 'X', whereas the second case can only have exactly 1
601       // 'x' or 'X'. We ignore the character if it appears as the last character of the string.
602       for (int index = 0; index < candidate.length() - 1; index++) {
603         char charAtIndex = candidate.charAt(index);
604         if (charAtIndex == 'x' || charAtIndex == 'X') {
605           char charAtNextIndex = candidate.charAt(index + 1);
606           if (charAtNextIndex == 'x' || charAtNextIndex == 'X') {
607             // This is the carrier code case, in which the 'X's always precede the national
608             // significant number.
609             index++;
610             if (util.isNumberMatch(number, candidate.substring(index)) != MatchType.NSN_MATCH) {
611               return false;
612             }
613           // This is the extension sign case, in which the 'x' or 'X' should always precede the
614           // extension number.
615           } else if (!PhoneNumberUtil.normalizeDigitsOnly(candidate.substring(index)).equals(
616               number.getExtension())) {
617               return false;
618           }
619         }
620       }
621       return true;
622     }
623 
isNationalPrefixPresentIfRequired( PhoneNumber number, PhoneNumberUtil util)624     private static boolean isNationalPrefixPresentIfRequired(
625         PhoneNumber number, PhoneNumberUtil util) {
626       // First, check how we deduced the country code. If it was written in international format,
627       // then the national prefix is not required.
628       if (number.getCountryCodeSource() != CountryCodeSource.FROM_DEFAULT_COUNTRY) {
629         return true;
630       }
631       String phoneNumberRegion =
632           util.getRegionCodeForCountryCode(number.getCountryCode());
633       PhoneMetadata metadata = util.getMetadataForRegion(phoneNumberRegion);
634       if (metadata == null) {
635         return true;
636       }
637       // Check if a national prefix should be present when formatting this number.
638       String nationalNumber = util.getNationalSignificantNumber(number);
639       NumberFormat formatRule =
640           util.chooseFormattingPatternForNumber(metadata.numberFormats(), nationalNumber);
641       // To do this, we check that a national prefix formatting rule was present and that it wasn't
642       // just the first-group symbol ($1) with punctuation.
643       if ((formatRule != null) && formatRule.getNationalPrefixFormattingRule().length() > 0) {
644         if (formatRule.isNationalPrefixOptionalWhenFormatting()) {
645           // The national-prefix is optional in these cases, so we don't need to check if it was
646           // present.
647           return true;
648         }
649         // Remove the first-group symbol.
650         String candidateNationalPrefixRule = formatRule.getNationalPrefixFormattingRule();
651         // We assume that the first-group symbol will never be _before_ the national prefix.
652         candidateNationalPrefixRule =
653             candidateNationalPrefixRule.substring(0, candidateNationalPrefixRule.indexOf("$1"));
654         candidateNationalPrefixRule = normalizeDigitsOnly(candidateNationalPrefixRule);
655         if (candidateNationalPrefixRule.length() == 0) {
656           // National Prefix not needed for this number.
657           return true;
658         }
659         // Normalize the remainder.
660         String rawInputCopy = normalizeDigitsOnly(number.getRawInput());
661         StringBuilder rawInput = new StringBuilder(rawInputCopy);
662         // Check if we found a national prefix and/or carrier code at the start of the raw input,
663         // and return the result.
664         return util.maybeStripNationalPrefixAndCarrierCode(rawInput, metadata, null);
665       }
666       return true;
667     }
668 
669     /** Returns true if {@code number} is a verified number according to this leniency. */
verify(PhoneNumber number, String candidate, PhoneNumberUtil util)670     abstract boolean verify(PhoneNumber number, String candidate, PhoneNumberUtil util);
671   }
672 
673   /**
674    * This class implements a singleton, so the only constructor is private.
675    */
PhoneNumberUtil()676   private PhoneNumberUtil() {
677   }
678 
init(String filePrefix)679   private void init(String filePrefix) {
680     currentFilePrefix = filePrefix;
681     for (List<String> regionCodes : countryCallingCodeToRegionCodeMap.values()) {
682       supportedRegions.addAll(regionCodes);
683     }
684     supportedRegions.remove(REGION_CODE_FOR_NON_GEO_ENTITY);
685     nanpaRegions.addAll(countryCallingCodeToRegionCodeMap.get(NANPA_COUNTRY_CODE));
686   }
687 
loadMetadataFromFile(String filePrefix, String regionCode, int countryCallingCode)688   private void loadMetadataFromFile(String filePrefix, String regionCode, int countryCallingCode) {
689     boolean isNonGeoRegion = REGION_CODE_FOR_NON_GEO_ENTITY.equals(regionCode);
690     InputStream source = isNonGeoRegion
691         ? PhoneNumberUtil.class.getResourceAsStream(filePrefix + "_" + countryCallingCode)
692         : PhoneNumberUtil.class.getResourceAsStream(filePrefix + "_" + regionCode);
693     ObjectInputStream in = null;
694     try {
695       in = new ObjectInputStream(source);
696       PhoneMetadataCollection metadataCollection = new PhoneMetadataCollection();
697       metadataCollection.readExternal(in);
698       for (PhoneMetadata metadata : metadataCollection.getMetadataList()) {
699         if (isNonGeoRegion) {
700           countryCodeToNonGeographicalMetadataMap.put(countryCallingCode, metadata);
701         } else {
702           regionToMetadataMap.put(regionCode, metadata);
703         }
704       }
705     } catch (IOException e) {
706       LOGGER.log(Level.WARNING, e.toString());
707     } finally {
708       close(in);
709     }
710   }
711 
close(InputStream in)712   private static void close(InputStream in) {
713     if (in != null) {
714       try {
715         in.close();
716       } catch (IOException e) {
717         LOGGER.log(Level.WARNING, e.toString());
718       }
719     }
720   }
721 
722   /**
723    * Attempts to extract a possible number from the string passed in. This currently strips all
724    * leading characters that cannot be used to start a phone number. Characters that can be used to
725    * start a phone number are defined in the VALID_START_CHAR_PATTERN. If none of these characters
726    * are found in the number passed in, an empty string is returned. This function also attempts to
727    * strip off any alternative extensions or endings if two or more are present, such as in the case
728    * of: (530) 583-6985 x302/x2303. The second extension here makes this actually two phone numbers,
729    * (530) 583-6985 x302 and (530) 583-6985 x2303. We remove the second extension so that the first
730    * number is parsed correctly.
731    *
732    * @param number  the string that might contain a phone number
733    * @return        the number, stripped of any non-phone-number prefix (such as "Tel:") or an empty
734    *                string if no character used to start phone numbers (such as + or any digit) is
735    *                found in the number
736    */
extractPossibleNumber(String number)737   static String extractPossibleNumber(String number) {
738     Matcher m = VALID_START_CHAR_PATTERN.matcher(number);
739     if (m.find()) {
740       number = number.substring(m.start());
741       // Remove trailing non-alpha non-numerical characters.
742       Matcher trailingCharsMatcher = UNWANTED_END_CHAR_PATTERN.matcher(number);
743       if (trailingCharsMatcher.find()) {
744         number = number.substring(0, trailingCharsMatcher.start());
745         LOGGER.log(Level.FINER, "Stripped trailing characters: " + number);
746       }
747       // Check for extra numbers at the end.
748       Matcher secondNumber = SECOND_NUMBER_START_PATTERN.matcher(number);
749       if (secondNumber.find()) {
750         number = number.substring(0, secondNumber.start());
751       }
752       return number;
753     } else {
754       return "";
755     }
756   }
757 
758   /**
759    * Checks to see if the string of characters could possibly be a phone number at all. At the
760    * moment, checks to see that the string begins with at least 3 digits, ignoring any punctuation
761    * commonly found in phone numbers.
762    * This method does not require the number to be normalized in advance - but does assume that
763    * leading non-number symbols have been removed, such as by the method extractPossibleNumber.
764    *
765    * @param number  string to be checked for viability as a phone number
766    * @return        true if the number could be a phone number of some sort, otherwise false
767    */
768   // @VisibleForTesting
isViablePhoneNumber(String number)769   static boolean isViablePhoneNumber(String number) {
770     if (number.length() < MIN_LENGTH_FOR_NSN) {
771       return false;
772     }
773     Matcher m = VALID_PHONE_NUMBER_PATTERN.matcher(number);
774     return m.matches();
775   }
776 
777   /**
778    * Normalizes a string of characters representing a phone number. This performs the following
779    * conversions:
780    *   Punctuation is stripped.
781    *   For ALPHA/VANITY numbers:
782    *   Letters are converted to their numeric representation on a telephone keypad. The keypad
783    *       used here is the one defined in ITU Recommendation E.161. This is only done if there are
784    *       3 or more letters in the number, to lessen the risk that such letters are typos.
785    *   For other numbers:
786    *   Wide-ascii digits are converted to normal ASCII (European) digits.
787    *   Arabic-Indic numerals are converted to European numerals.
788    *   Spurious alpha characters are stripped.
789    *
790    * @param number  a string of characters representing a phone number
791    * @return        the normalized string version of the phone number
792    */
normalize(String number)793   static String normalize(String number) {
794     Matcher m = VALID_ALPHA_PHONE_PATTERN.matcher(number);
795     if (m.matches()) {
796       return normalizeHelper(number, ALPHA_PHONE_MAPPINGS, true);
797     } else {
798       return normalizeDigitsOnly(number);
799     }
800   }
801 
802   /**
803    * Normalizes a string of characters representing a phone number. This is a wrapper for
804    * normalize(String number) but does in-place normalization of the StringBuilder provided.
805    *
806    * @param number  a StringBuilder of characters representing a phone number that will be
807    *     normalized in place
808    */
normalize(StringBuilder number)809   static void normalize(StringBuilder number) {
810     String normalizedNumber = normalize(number.toString());
811     number.replace(0, number.length(), normalizedNumber);
812   }
813 
814   /**
815    * Normalizes a string of characters representing a phone number. This converts wide-ascii and
816    * arabic-indic numerals to European numerals, and strips punctuation and alpha characters.
817    *
818    * @param number  a string of characters representing a phone number
819    * @return        the normalized string version of the phone number
820    */
normalizeDigitsOnly(String number)821   public static String normalizeDigitsOnly(String number) {
822     return normalizeDigits(number, false /* strip non-digits */).toString();
823   }
824 
normalizeDigits(String number, boolean keepNonDigits)825   private static StringBuilder normalizeDigits(String number, boolean keepNonDigits) {
826     StringBuilder normalizedDigits = new StringBuilder(number.length());
827     for (char c : number.toCharArray()) {
828       int digit = Character.digit(c, 10);
829       if (digit != -1) {
830         normalizedDigits.append(digit);
831       } else if (keepNonDigits) {
832         normalizedDigits.append(c);
833       }
834     }
835     return normalizedDigits;
836   }
837 
838   /**
839    * Converts all alpha characters in a number to their respective digits on a keypad, but retains
840    * existing formatting.
841    */
convertAlphaCharactersInNumber(String number)842   public static String convertAlphaCharactersInNumber(String number) {
843     return normalizeHelper(number, ALPHA_PHONE_MAPPINGS, false);
844   }
845 
846   /**
847    * Gets the length of the geographical area code in the {@code nationalNumber_} field of the
848    * PhoneNumber object passed in, so that clients could use it to split a national significant
849    * number into geographical area code and subscriber number. It works in such a way that the
850    * resultant subscriber number should be diallable, at least on some devices. An example of how
851    * this could be used:
852    *
853    * <pre>
854    * PhoneNumberUtil phoneUtil = PhoneNumberUtil.getInstance();
855    * PhoneNumber number = phoneUtil.parse("16502530000", "US");
856    * String nationalSignificantNumber = phoneUtil.getNationalSignificantNumber(number);
857    * String areaCode;
858    * String subscriberNumber;
859    *
860    * int areaCodeLength = phoneUtil.getLengthOfGeographicalAreaCode(number);
861    * if (areaCodeLength > 0) {
862    *   areaCode = nationalSignificantNumber.substring(0, areaCodeLength);
863    *   subscriberNumber = nationalSignificantNumber.substring(areaCodeLength);
864    * } else {
865    *   areaCode = "";
866    *   subscriberNumber = nationalSignificantNumber;
867    * }
868    * </pre>
869    *
870    * N.B.: area code is a very ambiguous concept, so the I18N team generally recommends against
871    * using it for most purposes, but recommends using the more general {@code national_number}
872    * instead. Read the following carefully before deciding to use this method:
873    * <ul>
874    *  <li> geographical area codes change over time, and this method honors those changes;
875    *    therefore, it doesn't guarantee the stability of the result it produces.
876    *  <li> subscriber numbers may not be diallable from all devices (notably mobile devices, which
877    *    typically requires the full national_number to be dialled in most regions).
878    *  <li> most non-geographical numbers have no area codes, including numbers from non-geographical
879    *    entities
880    *  <li> some geographical numbers have no area codes.
881    * </ul>
882    * @param number  the PhoneNumber object for which clients want to know the length of the area
883    *     code.
884    * @return  the length of area code of the PhoneNumber object passed in.
885    */
getLengthOfGeographicalAreaCode(PhoneNumber number)886   public int getLengthOfGeographicalAreaCode(PhoneNumber number) {
887     String regionCode = getRegionCodeForNumber(number);
888     if (!isValidRegionCode(regionCode)) {
889       return 0;
890     }
891     PhoneMetadata metadata = getMetadataForRegion(regionCode);
892     if (!metadata.hasNationalPrefix()) {
893       return 0;
894     }
895 
896     PhoneNumberType type = getNumberTypeHelper(getNationalSignificantNumber(number),
897                                                metadata);
898     // Most numbers other than the two types below have to be dialled in full.
899     if (type != PhoneNumberType.FIXED_LINE && type != PhoneNumberType.FIXED_LINE_OR_MOBILE) {
900       return 0;
901     }
902 
903     return getLengthOfNationalDestinationCode(number);
904   }
905 
906   /**
907    * Gets the length of the national destination code (NDC) from the PhoneNumber object passed in,
908    * so that clients could use it to split a national significant number into NDC and subscriber
909    * number. The NDC of a phone number is normally the first group of digit(s) right after the
910    * country calling code when the number is formatted in the international format, if there is a
911    * subscriber number part that follows. An example of how this could be used:
912    *
913    * <pre>
914    * PhoneNumberUtil phoneUtil = PhoneNumberUtil.getInstance();
915    * PhoneNumber number = phoneUtil.parse("18002530000", "US");
916    * String nationalSignificantNumber = phoneUtil.getNationalSignificantNumber(number);
917    * String nationalDestinationCode;
918    * String subscriberNumber;
919    *
920    * int nationalDestinationCodeLength = phoneUtil.getLengthOfNationalDestinationCode(number);
921    * if (nationalDestinationCodeLength > 0) {
922    *   nationalDestinationCode = nationalSignificantNumber.substring(0,
923    *       nationalDestinationCodeLength);
924    *   subscriberNumber = nationalSignificantNumber.substring(nationalDestinationCodeLength);
925    * } else {
926    *   nationalDestinationCode = "";
927    *   subscriberNumber = nationalSignificantNumber;
928    * }
929    * </pre>
930    *
931    * Refer to the unittests to see the difference between this function and
932    * {@link #getLengthOfGeographicalAreaCode}.
933    *
934    * @param number  the PhoneNumber object for which clients want to know the length of the NDC.
935    * @return  the length of NDC of the PhoneNumber object passed in.
936    */
getLengthOfNationalDestinationCode(PhoneNumber number)937   public int getLengthOfNationalDestinationCode(PhoneNumber number) {
938     PhoneNumber copiedProto;
939     if (number.hasExtension()) {
940       // We don't want to alter the proto given to us, but we don't want to include the extension
941       // when we format it, so we copy it and clear the extension here.
942       copiedProto = new PhoneNumber();
943       copiedProto.mergeFrom(number);
944       copiedProto.clearExtension();
945     } else {
946       copiedProto = number;
947     }
948 
949     String nationalSignificantNumber = format(copiedProto,
950                                               PhoneNumberUtil.PhoneNumberFormat.INTERNATIONAL);
951     String[] numberGroups = NON_DIGITS_PATTERN.split(nationalSignificantNumber);
952     // The pattern will start with "+COUNTRY_CODE " so the first group will always be the empty
953     // string (before the + symbol) and the second group will be the country calling code. The third
954     // group will be area code if it is not the last group.
955     if (numberGroups.length <= 3) {
956       return 0;
957     }
958 
959     if (getRegionCodeForCountryCode(number.getCountryCode()).equals("AR") &&
960         getNumberType(number) == PhoneNumberType.MOBILE) {
961       // Argentinian mobile numbers, when formatted in the international format, are in the form of
962       // +54 9 NDC XXXX.... As a result, we take the length of the third group (NDC) and add 1 for
963       // the digit 9, which also forms part of the national significant number.
964       //
965       // TODO: Investigate the possibility of better modeling the metadata to make it
966       // easier to obtain the NDC.
967       return numberGroups[3].length() + 1;
968     }
969     return numberGroups[2].length();
970   }
971 
972   /**
973    * Normalizes a string of characters representing a phone number by replacing all characters found
974    * in the accompanying map with the values therein, and stripping all other characters if
975    * removeNonMatches is true.
976    *
977    * @param number                     a string of characters representing a phone number
978    * @param normalizationReplacements  a mapping of characters to what they should be replaced by in
979    *                                   the normalized version of the phone number
980    * @param removeNonMatches           indicates whether characters that are not able to be replaced
981    *                                   should be stripped from the number. If this is false, they
982    *                                   will be left unchanged in the number.
983    * @return  the normalized string version of the phone number
984    */
normalizeHelper(String number, Map<Character, Character> normalizationReplacements, boolean removeNonMatches)985   private static String normalizeHelper(String number,
986                                         Map<Character, Character> normalizationReplacements,
987                                         boolean removeNonMatches) {
988     StringBuilder normalizedNumber = new StringBuilder(number.length());
989     char[] numberAsCharArray = number.toCharArray();
990     for (char character : numberAsCharArray) {
991       Character newDigit = normalizationReplacements.get(Character.toUpperCase(character));
992       if (newDigit != null) {
993         normalizedNumber.append(newDigit);
994       } else if (!removeNonMatches) {
995         normalizedNumber.append(character);
996       }
997       // If neither of the above are true, we remove this character.
998     }
999     return normalizedNumber.toString();
1000   }
1001 
1002   // @VisibleForTesting
getInstance( String baseFileLocation, Map<Integer, List<String>> countryCallingCodeToRegionCodeMap)1003   static synchronized PhoneNumberUtil getInstance(
1004       String baseFileLocation,
1005       Map<Integer, List<String>> countryCallingCodeToRegionCodeMap) {
1006     if (instance == null) {
1007       instance = new PhoneNumberUtil();
1008       instance.countryCallingCodeToRegionCodeMap = countryCallingCodeToRegionCodeMap;
1009       instance.init(baseFileLocation);
1010     }
1011     return instance;
1012   }
1013 
1014   /**
1015    * Used for testing purposes only to reset the PhoneNumberUtil singleton to null.
1016    */
1017   // @VisibleForTesting
resetInstance()1018   static synchronized void resetInstance() {
1019     instance = null;
1020   }
1021 
1022   /**
1023    * Convenience method to get a list of what regions the library has metadata for.
1024    */
getSupportedRegions()1025   public Set<String> getSupportedRegions() {
1026     return supportedRegions;
1027   }
1028 
1029   /**
1030    * Convenience method to get a list of what global network calling codes the library has metadata
1031    * for.
1032    */
getSupportedGlobalNetworkCallingCodes()1033   public Set<Integer> getSupportedGlobalNetworkCallingCodes() {
1034     return countryCodeToNonGeographicalMetadataMap.keySet();
1035   }
1036 
1037   /**
1038    * Gets a {@link PhoneNumberUtil} instance to carry out international phone number formatting,
1039    * parsing, or validation. The instance is loaded with phone number metadata for a number of most
1040    * commonly used regions.
1041    *
1042    * <p>The {@link PhoneNumberUtil} is implemented as a singleton. Therefore, calling getInstance
1043    * multiple times will only result in one instance being created.
1044    *
1045    * @return a PhoneNumberUtil instance
1046    */
getInstance()1047   public static synchronized PhoneNumberUtil getInstance() {
1048     if (instance == null) {
1049       return getInstance(META_DATA_FILE_PREFIX,
1050           CountryCodeToRegionCodeMap.getCountryCodeToRegionCodeMap());
1051     }
1052     return instance;
1053   }
1054 
1055   /**
1056    * Helper function to check region code is not unknown or null.
1057    */
isValidRegionCode(String regionCode)1058   private boolean isValidRegionCode(String regionCode) {
1059     return regionCode != null && supportedRegions.contains(regionCode);
1060   }
1061 
1062   /**
1063    * Helper function to check the country calling code is valid.
1064    */
hasValidCountryCallingCode(int countryCallingCode)1065   private boolean hasValidCountryCallingCode(int countryCallingCode) {
1066     return countryCallingCodeToRegionCodeMap.containsKey(countryCallingCode);
1067   }
1068 
1069   /**
1070    * Formats a phone number in the specified format using default rules. Note that this does not
1071    * promise to produce a phone number that the user can dial from where they are - although we do
1072    * format in either 'national' or 'international' format depending on what the client asks for, we
1073    * do not currently support a more abbreviated format, such as for users in the same "area" who
1074    * could potentially dial the number without area code. Note that if the phone number has a
1075    * country calling code of 0 or an otherwise invalid country calling code, we cannot work out
1076    * which formatting rules to apply so we return the national significant number with no formatting
1077    * applied.
1078    *
1079    * @param number         the phone number to be formatted
1080    * @param numberFormat   the format the phone number should be formatted into
1081    * @return  the formatted phone number
1082    */
format(PhoneNumber number, PhoneNumberFormat numberFormat)1083   public String format(PhoneNumber number, PhoneNumberFormat numberFormat) {
1084     if (number.getNationalNumber() == 0 && number.hasRawInput()) {
1085       String rawInput = number.getRawInput();
1086       if (rawInput.length() > 0) {
1087         return rawInput;
1088       }
1089     }
1090     StringBuilder formattedNumber = new StringBuilder(20);
1091     format(number, numberFormat, formattedNumber);
1092     return formattedNumber.toString();
1093   }
1094 
1095   /**
1096    * Same as {@link #format(PhoneNumber, PhoneNumberFormat)}, but accepts a mutable StringBuilder as
1097    * a parameter to decrease object creation when invoked many times.
1098    */
format(PhoneNumber number, PhoneNumberFormat numberFormat, StringBuilder formattedNumber)1099   public void format(PhoneNumber number, PhoneNumberFormat numberFormat,
1100                      StringBuilder formattedNumber) {
1101     // Clear the StringBuilder first.
1102     formattedNumber.setLength(0);
1103     int countryCallingCode = number.getCountryCode();
1104     String nationalSignificantNumber = getNationalSignificantNumber(number);
1105     if (numberFormat == PhoneNumberFormat.E164) {
1106       // Early exit for E164 case since no formatting of the national number needs to be applied.
1107       // Extensions are not formatted.
1108       formattedNumber.append(nationalSignificantNumber);
1109       prefixNumberWithCountryCallingCode(countryCallingCode, PhoneNumberFormat.E164,
1110                                          formattedNumber);
1111       return;
1112     }
1113     // Note getRegionCodeForCountryCode() is used because formatting information for regions which
1114     // share a country calling code is contained by only one region for performance reasons. For
1115     // example, for NANPA regions it will be contained in the metadata for US.
1116     String regionCode = getRegionCodeForCountryCode(countryCallingCode);
1117     if (!hasValidCountryCallingCode(countryCallingCode)) {
1118       formattedNumber.append(nationalSignificantNumber);
1119       return;
1120     }
1121 
1122     PhoneMetadata metadata =
1123         getMetadataForRegionOrCallingCode(countryCallingCode, regionCode);
1124     formattedNumber.append(formatNsn(nationalSignificantNumber, metadata, numberFormat));
1125     maybeAppendFormattedExtension(number, metadata, numberFormat, formattedNumber);
1126     prefixNumberWithCountryCallingCode(countryCallingCode, numberFormat, formattedNumber);
1127   }
1128 
1129   /**
1130    * Formats a phone number in the specified format using client-defined formatting rules. Note that
1131    * if the phone number has a country calling code of zero or an otherwise invalid country calling
1132    * code, we cannot work out things like whether there should be a national prefix applied, or how
1133    * to format extensions, so we return the national significant number with no formatting applied.
1134    *
1135    * @param number                        the phone number to be formatted
1136    * @param numberFormat                  the format the phone number should be formatted into
1137    * @param userDefinedFormats            formatting rules specified by clients
1138    * @return  the formatted phone number
1139    */
formatByPattern(PhoneNumber number, PhoneNumberFormat numberFormat, List<NumberFormat> userDefinedFormats)1140   public String formatByPattern(PhoneNumber number,
1141                                 PhoneNumberFormat numberFormat,
1142                                 List<NumberFormat> userDefinedFormats) {
1143     int countryCallingCode = number.getCountryCode();
1144     String nationalSignificantNumber = getNationalSignificantNumber(number);
1145     // Note getRegionCodeForCountryCode() is used because formatting information for regions which
1146     // share a country calling code is contained by only one region for performance reasons. For
1147     // example, for NANPA regions it will be contained in the metadata for US.
1148     String regionCode = getRegionCodeForCountryCode(countryCallingCode);
1149     if (!hasValidCountryCallingCode(countryCallingCode)) {
1150       return nationalSignificantNumber;
1151     }
1152     PhoneMetadata metadata =
1153         getMetadataForRegionOrCallingCode(countryCallingCode, regionCode);
1154 
1155     StringBuilder formattedNumber = new StringBuilder(20);
1156 
1157     NumberFormat formattingPattern =
1158         chooseFormattingPatternForNumber(userDefinedFormats, nationalSignificantNumber);
1159     if (formattingPattern == null) {
1160       // If no pattern above is matched, we format the number as a whole.
1161       formattedNumber.append(nationalSignificantNumber);
1162     } else {
1163       NumberFormat numFormatCopy = new NumberFormat();
1164       // Before we do a replacement of the national prefix pattern $NP with the national prefix, we
1165       // need to copy the rule so that subsequent replacements for different numbers have the
1166       // appropriate national prefix.
1167       numFormatCopy.mergeFrom(formattingPattern);
1168       String nationalPrefixFormattingRule = formattingPattern.getNationalPrefixFormattingRule();
1169       if (nationalPrefixFormattingRule.length() > 0) {
1170         String nationalPrefix = metadata.getNationalPrefix();
1171         if (nationalPrefix.length() > 0) {
1172           // Replace $NP with national prefix and $FG with the first group ($1).
1173           nationalPrefixFormattingRule =
1174               NP_PATTERN.matcher(nationalPrefixFormattingRule).replaceFirst(nationalPrefix);
1175           nationalPrefixFormattingRule =
1176               FG_PATTERN.matcher(nationalPrefixFormattingRule).replaceFirst("\\$1");
1177           numFormatCopy.setNationalPrefixFormattingRule(nationalPrefixFormattingRule);
1178         } else {
1179           // We don't want to have a rule for how to format the national prefix if there isn't one.
1180           numFormatCopy.clearNationalPrefixFormattingRule();
1181         }
1182       }
1183       formattedNumber.append(
1184           formatNsnUsingPattern(nationalSignificantNumber, numFormatCopy, numberFormat));
1185     }
1186     maybeAppendFormattedExtension(number, metadata, numberFormat, formattedNumber);
1187     prefixNumberWithCountryCallingCode(countryCallingCode, numberFormat, formattedNumber);
1188     return formattedNumber.toString();
1189   }
1190 
1191   /**
1192    * Formats a phone number in national format for dialing using the carrier as specified in the
1193    * {@code carrierCode}. The {@code carrierCode} will always be used regardless of whether the
1194    * phone number already has a preferred domestic carrier code stored. If {@code carrierCode}
1195    * contains an empty string, returns the number in national format without any carrier code.
1196    *
1197    * @param number  the phone number to be formatted
1198    * @param carrierCode  the carrier selection code to be used
1199    * @return  the formatted phone number in national format for dialing using the carrier as
1200    *          specified in the {@code carrierCode}
1201    */
formatNationalNumberWithCarrierCode(PhoneNumber number, String carrierCode)1202   public String formatNationalNumberWithCarrierCode(PhoneNumber number, String carrierCode) {
1203     int countryCallingCode = number.getCountryCode();
1204     String nationalSignificantNumber = getNationalSignificantNumber(number);
1205     // Note getRegionCodeForCountryCode() is used because formatting information for regions which
1206     // share a country calling code is contained by only one region for performance reasons. For
1207     // example, for NANPA regions it will be contained in the metadata for US.
1208     String regionCode = getRegionCodeForCountryCode(countryCallingCode);
1209     if (!hasValidCountryCallingCode(countryCallingCode)) {
1210       return nationalSignificantNumber;
1211     }
1212 
1213     StringBuilder formattedNumber = new StringBuilder(20);
1214     PhoneMetadata metadata = getMetadataForRegionOrCallingCode(countryCallingCode, regionCode);
1215     formattedNumber.append(formatNsn(nationalSignificantNumber, metadata,
1216                                      PhoneNumberFormat.NATIONAL, carrierCode));
1217     maybeAppendFormattedExtension(number, metadata, PhoneNumberFormat.NATIONAL, formattedNumber);
1218     prefixNumberWithCountryCallingCode(countryCallingCode, PhoneNumberFormat.NATIONAL,
1219                                        formattedNumber);
1220     return formattedNumber.toString();
1221   }
1222 
getMetadataForRegionOrCallingCode( int countryCallingCode, String regionCode)1223   private PhoneMetadata getMetadataForRegionOrCallingCode(
1224       int countryCallingCode, String regionCode) {
1225     return REGION_CODE_FOR_NON_GEO_ENTITY.equals(regionCode)
1226         ? getMetadataForNonGeographicalRegion(countryCallingCode)
1227         : getMetadataForRegion(regionCode);
1228   }
1229 
1230   /**
1231    * Formats a phone number in national format for dialing using the carrier as specified in the
1232    * preferredDomesticCarrierCode field of the PhoneNumber object passed in. If that is missing,
1233    * use the {@code fallbackCarrierCode} passed in instead. If there is no
1234    * {@code preferredDomesticCarrierCode}, and the {@code fallbackCarrierCode} contains an empty
1235    * string, return the number in national format without any carrier code.
1236    *
1237    * <p>Use {@link #formatNationalNumberWithCarrierCode} instead if the carrier code passed in
1238    * should take precedence over the number's {@code preferredDomesticCarrierCode} when formatting.
1239    *
1240    * @param number  the phone number to be formatted
1241    * @param fallbackCarrierCode  the carrier selection code to be used, if none is found in the
1242    *     phone number itself
1243    * @return  the formatted phone number in national format for dialing using the number's
1244    *     {@code preferredDomesticCarrierCode}, or the {@code fallbackCarrierCode} passed in if
1245    *     none is found
1246    */
formatNationalNumberWithPreferredCarrierCode(PhoneNumber number, String fallbackCarrierCode)1247   public String formatNationalNumberWithPreferredCarrierCode(PhoneNumber number,
1248                                                              String fallbackCarrierCode) {
1249     return formatNationalNumberWithCarrierCode(number, number.hasPreferredDomesticCarrierCode()
1250                                                        ? number.getPreferredDomesticCarrierCode()
1251                                                        : fallbackCarrierCode);
1252   }
1253 
1254   /**
1255    * Returns a number formatted in such a way that it can be dialed from a mobile phone in a
1256    * specific region. If the number cannot be reached from the region (e.g. some countries block
1257    * toll-free numbers from being called outside of the country), the method returns an empty
1258    * string.
1259    *
1260    * @param number  the phone number to be formatted
1261    * @param regionCallingFrom  the region where the call is being placed
1262    * @param withFormatting  whether the number should be returned with formatting symbols, such as
1263    *     spaces and dashes.
1264    * @return  the formatted phone number
1265    */
formatNumberForMobileDialing(PhoneNumber number, String regionCallingFrom, boolean withFormatting)1266   public String formatNumberForMobileDialing(PhoneNumber number, String regionCallingFrom,
1267                                              boolean withFormatting) {
1268     int countryCallingCode = number.getCountryCode();
1269     if (!hasValidCountryCallingCode(countryCallingCode)) {
1270       return number.hasRawInput() ? number.getRawInput() : "";
1271     }
1272 
1273     String formattedNumber;
1274     // Clear the extension, as that part cannot normally be dialed together with the main number.
1275     PhoneNumber numberNoExt = new PhoneNumber().mergeFrom(number).clearExtension();
1276     PhoneNumberType numberType = getNumberType(numberNoExt);
1277     String regionCode = getRegionCodeForCountryCode(countryCallingCode);
1278     if (regionCode.equals("CO") && regionCallingFrom.equals("CO")) {
1279       if (numberType == PhoneNumberType.FIXED_LINE) {
1280         formattedNumber =
1281             formatNationalNumberWithCarrierCode(numberNoExt, COLOMBIA_MOBILE_TO_FIXED_LINE_PREFIX);
1282       } else {
1283         // E164 doesn't work at all when dialing within Colombia.
1284         formattedNumber = format(numberNoExt, PhoneNumberFormat.NATIONAL);
1285       }
1286     } else if (regionCode.equals("PE") && regionCallingFrom.equals("PE")) {
1287       // In Peru, numbers cannot be dialled using E164 format from a mobile phone for Movistar.
1288       // Instead they must be dialled in national format.
1289       formattedNumber = format(numberNoExt, PhoneNumberFormat.NATIONAL);
1290     } else if (regionCode.equals("BR") && regionCallingFrom.equals("BR") &&
1291         ((numberType == PhoneNumberType.FIXED_LINE) || (numberType == PhoneNumberType.MOBILE) ||
1292          (numberType == PhoneNumberType.FIXED_LINE_OR_MOBILE))) {
1293       formattedNumber = numberNoExt.hasPreferredDomesticCarrierCode()
1294           ? formatNationalNumberWithPreferredCarrierCode(numberNoExt, "")
1295           // Brazilian fixed line and mobile numbers need to be dialed with a carrier code when
1296           // called within Brazil. Without that, most of the carriers won't connect the call.
1297           // Because of that, we return an empty string here.
1298           : "";
1299     } else if (canBeInternationallyDialled(numberNoExt)) {
1300       return withFormatting ? format(numberNoExt, PhoneNumberFormat.INTERNATIONAL)
1301                             : format(numberNoExt, PhoneNumberFormat.E164);
1302     } else {
1303       formattedNumber = (regionCallingFrom.equals(regionCode))
1304           ? format(numberNoExt, PhoneNumberFormat.NATIONAL) : "";
1305     }
1306     return withFormatting ? formattedNumber
1307                           : normalizeHelper(formattedNumber, DIALLABLE_CHAR_MAPPINGS,
1308                                             true /* remove non matches */);
1309   }
1310 
1311   /**
1312    * Formats a phone number for out-of-country dialing purposes. If no regionCallingFrom is
1313    * supplied, we format the number in its INTERNATIONAL format. If the country calling code is the
1314    * same as that of the region where the number is from, then NATIONAL formatting will be applied.
1315    *
1316    * <p>If the number itself has a country calling code of zero or an otherwise invalid country
1317    * calling code, then we return the number with no formatting applied.
1318    *
1319    * <p>Note this function takes care of the case for calling inside of NANPA and between Russia and
1320    * Kazakhstan (who share the same country calling code). In those cases, no international prefix
1321    * is used. For regions which have multiple international prefixes, the number in its
1322    * INTERNATIONAL format will be returned instead.
1323    *
1324    * @param number               the phone number to be formatted
1325    * @param regionCallingFrom    the region where the call is being placed
1326    * @return  the formatted phone number
1327    */
formatOutOfCountryCallingNumber(PhoneNumber number, String regionCallingFrom)1328   public String formatOutOfCountryCallingNumber(PhoneNumber number,
1329                                                 String regionCallingFrom) {
1330     if (!isValidRegionCode(regionCallingFrom)) {
1331       LOGGER.log(Level.WARNING,
1332                  "Trying to format number from invalid region "
1333                  + regionCallingFrom
1334                  + ". International formatting applied.");
1335       return format(number, PhoneNumberFormat.INTERNATIONAL);
1336     }
1337     int countryCallingCode = number.getCountryCode();
1338     String nationalSignificantNumber = getNationalSignificantNumber(number);
1339     if (!hasValidCountryCallingCode(countryCallingCode)) {
1340       return nationalSignificantNumber;
1341     }
1342     if (countryCallingCode == NANPA_COUNTRY_CODE) {
1343       if (isNANPACountry(regionCallingFrom)) {
1344         // For NANPA regions, return the national format for these regions but prefix it with the
1345         // country calling code.
1346         return countryCallingCode + " " + format(number, PhoneNumberFormat.NATIONAL);
1347       }
1348     } else if (countryCallingCode == getCountryCodeForValidRegion(regionCallingFrom)) {
1349     // For regions that share a country calling code, the country calling code need not be dialled.
1350     // This also applies when dialling within a region, so this if clause covers both these cases.
1351     // Technically this is the case for dialling from La Reunion to other overseas departments of
1352     // France (French Guiana, Martinique, Guadeloupe), but not vice versa - so we don't cover this
1353     // edge case for now and for those cases return the version including country calling code.
1354     // Details here: http://www.petitfute.com/voyage/225-info-pratiques-reunion
1355       return format(number, PhoneNumberFormat.NATIONAL);
1356     }
1357     PhoneMetadata metadataForRegionCallingFrom = getMetadataForRegion(regionCallingFrom);
1358     String internationalPrefix = metadataForRegionCallingFrom.getInternationalPrefix();
1359 
1360     // For regions that have multiple international prefixes, the international format of the
1361     // number is returned, unless there is a preferred international prefix.
1362     String internationalPrefixForFormatting = "";
1363     if (UNIQUE_INTERNATIONAL_PREFIX.matcher(internationalPrefix).matches()) {
1364       internationalPrefixForFormatting = internationalPrefix;
1365     } else if (metadataForRegionCallingFrom.hasPreferredInternationalPrefix()) {
1366       internationalPrefixForFormatting =
1367           metadataForRegionCallingFrom.getPreferredInternationalPrefix();
1368     }
1369 
1370     String regionCode = getRegionCodeForCountryCode(countryCallingCode);
1371     PhoneMetadata metadataForRegion =
1372         getMetadataForRegionOrCallingCode(countryCallingCode, regionCode);
1373     String formattedNationalNumber =
1374         formatNsn(nationalSignificantNumber, metadataForRegion, PhoneNumberFormat.INTERNATIONAL);
1375     StringBuilder formattedNumber = new StringBuilder(formattedNationalNumber);
1376     maybeAppendFormattedExtension(number, metadataForRegion, PhoneNumberFormat.INTERNATIONAL,
1377                                   formattedNumber);
1378     if (internationalPrefixForFormatting.length() > 0) {
1379       formattedNumber.insert(0, " ").insert(0, countryCallingCode).insert(0, " ")
1380           .insert(0, internationalPrefixForFormatting);
1381     } else {
1382       prefixNumberWithCountryCallingCode(countryCallingCode,
1383                                          PhoneNumberFormat.INTERNATIONAL,
1384                                          formattedNumber);
1385     }
1386     return formattedNumber.toString();
1387   }
1388 
1389   /**
1390    * Formats a phone number using the original phone number format that the number is parsed from.
1391    * The original format is embedded in the country_code_source field of the PhoneNumber object
1392    * passed in. If such information is missing, the number will be formatted into the NATIONAL
1393    * format by default. When the number contains a leading zero and this is unexpected for this
1394    * country, or we don't have a formatting pattern for the number, the method returns the raw input
1395    * when it is available.
1396    *
1397    * Note this method guarantees no digit will be inserted, removed or modified as a result of
1398    * formatting.
1399    *
1400    * @param number  the phone number that needs to be formatted in its original number format
1401    * @param regionCallingFrom  the region whose IDD needs to be prefixed if the original number
1402    *     has one
1403    * @return  the formatted phone number in its original number format
1404    */
formatInOriginalFormat(PhoneNumber number, String regionCallingFrom)1405   public String formatInOriginalFormat(PhoneNumber number, String regionCallingFrom) {
1406     if (number.hasRawInput() &&
1407         (hasUnexpectedItalianLeadingZero(number) || !hasFormattingPatternForNumber(number))) {
1408       // We check if we have the formatting pattern because without that, we might format the number
1409       // as a group without national prefix.
1410       return number.getRawInput();
1411     }
1412     if (!number.hasCountryCodeSource()) {
1413       return format(number, PhoneNumberFormat.NATIONAL);
1414     }
1415     String formattedNumber;
1416     switch (number.getCountryCodeSource()) {
1417       case FROM_NUMBER_WITH_PLUS_SIGN:
1418         formattedNumber = format(number, PhoneNumberFormat.INTERNATIONAL);
1419         break;
1420       case FROM_NUMBER_WITH_IDD:
1421         formattedNumber = formatOutOfCountryCallingNumber(number, regionCallingFrom);
1422         break;
1423       case FROM_NUMBER_WITHOUT_PLUS_SIGN:
1424         formattedNumber = format(number, PhoneNumberFormat.INTERNATIONAL).substring(1);
1425         break;
1426       case FROM_DEFAULT_COUNTRY:
1427         // Fall-through to default case.
1428       default:
1429         String regionCode = getRegionCodeForCountryCode(number.getCountryCode());
1430         // We strip non-digits from the NDD here, and from the raw input later, so that we can
1431         // compare them easily.
1432         String nationalPrefix = getNddPrefixForRegion(regionCode, true /* strip non-digits */);
1433         String nationalFormat = format(number, PhoneNumberFormat.NATIONAL);
1434         if (nationalPrefix == null || nationalPrefix.length() == 0) {
1435           // If the region doesn't have a national prefix at all, we can safely return the national
1436           // format without worrying about a national prefix being added.
1437           formattedNumber = nationalFormat;
1438           break;
1439         }
1440         // Otherwise, we check if the original number was entered with a national prefix.
1441         if (rawInputContainsNationalPrefix(
1442             number.getRawInput(), nationalPrefix, regionCode)) {
1443           // If so, we can safely return the national format.
1444           formattedNumber = nationalFormat;
1445           break;
1446         }
1447         PhoneMetadata metadata = getMetadataForRegion(regionCode);
1448         String nationalNumber = getNationalSignificantNumber(number);
1449         NumberFormat formatRule =
1450             chooseFormattingPatternForNumber(metadata.numberFormats(), nationalNumber);
1451         // When the format we apply to this number doesn't contain national prefix, we can just
1452         // return the national format.
1453         // TODO: Refactor the code below with the code in isNationalPrefixPresentIfRequired.
1454         String candidateNationalPrefixRule = formatRule.getNationalPrefixFormattingRule();
1455         // We assume that the first-group symbol will never be _before_ the national prefix.
1456         int indexOfFirstGroup = candidateNationalPrefixRule.indexOf("$1");
1457         if (indexOfFirstGroup <= 0) {
1458           formattedNumber = nationalFormat;
1459           break;
1460         }
1461         candidateNationalPrefixRule =
1462             candidateNationalPrefixRule.substring(0, indexOfFirstGroup);
1463         candidateNationalPrefixRule = normalizeDigitsOnly(candidateNationalPrefixRule);
1464         if (candidateNationalPrefixRule.length() == 0) {
1465           // National prefix not used when formatting this number.
1466           formattedNumber = nationalFormat;
1467           break;
1468         }
1469         // Otherwise, we need to remove the national prefix from our output.
1470         NumberFormat numFormatCopy = new NumberFormat();
1471         numFormatCopy.mergeFrom(formatRule);
1472         numFormatCopy.clearNationalPrefixFormattingRule();
1473         List<NumberFormat> numberFormats = new ArrayList<NumberFormat>(1);
1474         numberFormats.add(numFormatCopy);
1475         formattedNumber = formatByPattern(number, PhoneNumberFormat.NATIONAL, numberFormats);
1476         break;
1477     }
1478     String rawInput = number.getRawInput();
1479     // If no digit is inserted/removed/modified as a result of our formatting, we return the
1480     // formatted phone number; otherwise we return the raw input the user entered.
1481     return (formattedNumber != null &&
1482             normalizeDigitsOnly(formattedNumber).equals(normalizeDigitsOnly(rawInput)))
1483         ? formattedNumber
1484         : rawInput;
1485   }
1486 
1487   // Check if rawInput, which is assumed to be in the national format, has a national prefix. The
1488   // national prefix is assumed to be in digits-only form.
rawInputContainsNationalPrefix(String rawInput, String nationalPrefix, String regionCode)1489   private boolean rawInputContainsNationalPrefix(String rawInput, String nationalPrefix,
1490       String regionCode) {
1491     String normalizedNationalNumber = normalizeDigitsOnly(rawInput);
1492     if (normalizedNationalNumber.startsWith(nationalPrefix)) {
1493       try {
1494         // Some Japanese numbers (e.g. 00777123) might be mistaken to contain the national prefix
1495         // when written without it (e.g. 0777123) if we just do prefix matching. To tackle that, we
1496         // check the validity of the number if the assumed national prefix is removed (777123 won't
1497         // be valid in Japan).
1498         return isValidNumber(
1499             parse(normalizedNationalNumber.substring(nationalPrefix.length()), regionCode));
1500       } catch (NumberParseException e) {
1501         return false;
1502       }
1503     }
1504     return false;
1505   }
1506 
1507   /**
1508    * Returns true if a number is from a region whose national significant number couldn't contain a
1509    * leading zero, but has the italian_leading_zero field set to true.
1510    */
hasUnexpectedItalianLeadingZero(PhoneNumber number)1511   private boolean hasUnexpectedItalianLeadingZero(PhoneNumber number) {
1512     return number.isItalianLeadingZero() && !isLeadingZeroPossible(number.getCountryCode());
1513   }
1514 
hasFormattingPatternForNumber(PhoneNumber number)1515   private boolean hasFormattingPatternForNumber(PhoneNumber number) {
1516     int countryCallingCode = number.getCountryCode();
1517     String phoneNumberRegion = getRegionCodeForCountryCode(countryCallingCode);
1518     PhoneMetadata metadata =
1519         getMetadataForRegionOrCallingCode(countryCallingCode, phoneNumberRegion);
1520     if (metadata == null) {
1521       return false;
1522     }
1523     String nationalNumber = getNationalSignificantNumber(number);
1524     NumberFormat formatRule =
1525         chooseFormattingPatternForNumber(metadata.numberFormats(), nationalNumber);
1526     return formatRule != null;
1527   }
1528 
1529   /**
1530    * Formats a phone number for out-of-country dialing purposes.
1531    *
1532    * Note that in this version, if the number was entered originally using alpha characters and
1533    * this version of the number is stored in raw_input, this representation of the number will be
1534    * used rather than the digit representation. Grouping information, as specified by characters
1535    * such as "-" and " ", will be retained.
1536    *
1537    * <p><b>Caveats:</b></p>
1538    * <ul>
1539    *  <li> This will not produce good results if the country calling code is both present in the raw
1540    *       input _and_ is the start of the national number. This is not a problem in the regions
1541    *       which typically use alpha numbers.
1542    *  <li> This will also not produce good results if the raw input has any grouping information
1543    *       within the first three digits of the national number, and if the function needs to strip
1544    *       preceding digits/words in the raw input before these digits. Normally people group the
1545    *       first three digits together so this is not a huge problem - and will be fixed if it
1546    *       proves to be so.
1547    * </ul>
1548    *
1549    * @param number  the phone number that needs to be formatted
1550    * @param regionCallingFrom  the region where the call is being placed
1551    * @return  the formatted phone number
1552    */
formatOutOfCountryKeepingAlphaChars(PhoneNumber number, String regionCallingFrom)1553   public String formatOutOfCountryKeepingAlphaChars(PhoneNumber number,
1554                                                     String regionCallingFrom) {
1555     String rawInput = number.getRawInput();
1556     // If there is no raw input, then we can't keep alpha characters because there aren't any.
1557     // In this case, we return formatOutOfCountryCallingNumber.
1558     if (rawInput.length() == 0) {
1559       return formatOutOfCountryCallingNumber(number, regionCallingFrom);
1560     }
1561     int countryCode = number.getCountryCode();
1562     if (!hasValidCountryCallingCode(countryCode)) {
1563       return rawInput;
1564     }
1565     // Strip any prefix such as country calling code, IDD, that was present. We do this by comparing
1566     // the number in raw_input with the parsed number.
1567     // To do this, first we normalize punctuation. We retain number grouping symbols such as " "
1568     // only.
1569     rawInput = normalizeHelper(rawInput, ALL_PLUS_NUMBER_GROUPING_SYMBOLS, true);
1570     // Now we trim everything before the first three digits in the parsed number. We choose three
1571     // because all valid alpha numbers have 3 digits at the start - if it does not, then we don't
1572     // trim anything at all. Similarly, if the national number was less than three digits, we don't
1573     // trim anything at all.
1574     String nationalNumber = getNationalSignificantNumber(number);
1575     if (nationalNumber.length() > 3) {
1576       int firstNationalNumberDigit = rawInput.indexOf(nationalNumber.substring(0, 3));
1577       if (firstNationalNumberDigit != -1) {
1578         rawInput = rawInput.substring(firstNationalNumberDigit);
1579       }
1580     }
1581     PhoneMetadata metadataForRegionCallingFrom = getMetadataForRegion(regionCallingFrom);
1582     if (countryCode == NANPA_COUNTRY_CODE) {
1583       if (isNANPACountry(regionCallingFrom)) {
1584         return countryCode + " " + rawInput;
1585       }
1586     } else if (isValidRegionCode(regionCallingFrom) &&
1587                countryCode == getCountryCodeForValidRegion(regionCallingFrom)) {
1588       NumberFormat formattingPattern =
1589           chooseFormattingPatternForNumber(metadataForRegionCallingFrom.numberFormats(),
1590                                            nationalNumber);
1591       if (formattingPattern == null) {
1592         // If no pattern above is matched, we format the original input.
1593         return rawInput;
1594       }
1595       NumberFormat newFormat = new NumberFormat();
1596       newFormat.mergeFrom(formattingPattern);
1597       // The first group is the first group of digits that the user wrote together.
1598       newFormat.setPattern("(\\d+)(.*)");
1599       // Here we just concatenate them back together after the national prefix has been fixed.
1600       newFormat.setFormat("$1$2");
1601       // Now we format using this pattern instead of the default pattern, but with the national
1602       // prefix prefixed if necessary.
1603       // This will not work in the cases where the pattern (and not the leading digits) decide
1604       // whether a national prefix needs to be used, since we have overridden the pattern to match
1605       // anything, but that is not the case in the metadata to date.
1606       return formatNsnUsingPattern(rawInput, newFormat, PhoneNumberFormat.NATIONAL);
1607     }
1608     String internationalPrefixForFormatting = "";
1609     // If an unsupported region-calling-from is entered, or a country with multiple international
1610     // prefixes, the international format of the number is returned, unless there is a preferred
1611     // international prefix.
1612     if (metadataForRegionCallingFrom != null) {
1613       String internationalPrefix = metadataForRegionCallingFrom.getInternationalPrefix();
1614       internationalPrefixForFormatting =
1615           UNIQUE_INTERNATIONAL_PREFIX.matcher(internationalPrefix).matches()
1616           ? internationalPrefix
1617           : metadataForRegionCallingFrom.getPreferredInternationalPrefix();
1618     }
1619     StringBuilder formattedNumber = new StringBuilder(rawInput);
1620     String regionCode = getRegionCodeForCountryCode(countryCode);
1621     PhoneMetadata metadataForRegion = getMetadataForRegionOrCallingCode(countryCode, regionCode);
1622     maybeAppendFormattedExtension(number, metadataForRegion,
1623                                   PhoneNumberFormat.INTERNATIONAL, formattedNumber);
1624     if (internationalPrefixForFormatting.length() > 0) {
1625       formattedNumber.insert(0, " ").insert(0, countryCode).insert(0, " ")
1626           .insert(0, internationalPrefixForFormatting);
1627     } else {
1628       // Invalid region entered as country-calling-from (so no metadata was found for it) or the
1629       // region chosen has multiple international dialling prefixes.
1630       LOGGER.log(Level.WARNING,
1631                  "Trying to format number from invalid region "
1632                  + regionCallingFrom
1633                  + ". International formatting applied.");
1634       prefixNumberWithCountryCallingCode(countryCode,
1635                                          PhoneNumberFormat.INTERNATIONAL,
1636                                          formattedNumber);
1637     }
1638     return formattedNumber.toString();
1639   }
1640 
1641   /**
1642    * Gets the national significant number of the a phone number. Note a national significant number
1643    * doesn't contain a national prefix or any formatting.
1644    *
1645    * @param number  the phone number for which the national significant number is needed
1646    * @return  the national significant number of the PhoneNumber object passed in
1647    */
getNationalSignificantNumber(PhoneNumber number)1648   public String getNationalSignificantNumber(PhoneNumber number) {
1649     // If a leading zero has been set, we prefix this now. Note this is not a national prefix.
1650     StringBuilder nationalNumber = new StringBuilder(number.isItalianLeadingZero() ? "0" : "");
1651     nationalNumber.append(number.getNationalNumber());
1652     return nationalNumber.toString();
1653   }
1654 
1655   /**
1656    * A helper function that is used by format and formatByPattern.
1657    */
prefixNumberWithCountryCallingCode(int countryCallingCode, PhoneNumberFormat numberFormat, StringBuilder formattedNumber)1658   private void prefixNumberWithCountryCallingCode(int countryCallingCode,
1659                                                   PhoneNumberFormat numberFormat,
1660                                                   StringBuilder formattedNumber) {
1661     switch (numberFormat) {
1662       case E164:
1663         formattedNumber.insert(0, countryCallingCode).insert(0, PLUS_SIGN);
1664         return;
1665       case INTERNATIONAL:
1666         formattedNumber.insert(0, " ").insert(0, countryCallingCode).insert(0, PLUS_SIGN);
1667         return;
1668       case RFC3966:
1669         formattedNumber.insert(0, "-").insert(0, countryCallingCode).insert(0, PLUS_SIGN)
1670             .insert(0, "tel:");
1671         return;
1672       case NATIONAL:
1673       default:
1674         return;
1675     }
1676   }
1677 
1678   // Simple wrapper of formatNsn for the common case of no carrier code.
formatNsn(String number, PhoneMetadata metadata, PhoneNumberFormat numberFormat)1679   private String formatNsn(String number, PhoneMetadata metadata, PhoneNumberFormat numberFormat) {
1680     return formatNsn(number, metadata, numberFormat, null);
1681   }
1682 
1683   // Note in some regions, the national number can be written in two completely different ways
1684   // depending on whether it forms part of the NATIONAL format or INTERNATIONAL format. The
1685   // numberFormat parameter here is used to specify which format to use for those cases. If a
1686   // carrierCode is specified, this will be inserted into the formatted string to replace $CC.
formatNsn(String number, PhoneMetadata metadata, PhoneNumberFormat numberFormat, String carrierCode)1687   private String formatNsn(String number,
1688                            PhoneMetadata metadata,
1689                            PhoneNumberFormat numberFormat,
1690                            String carrierCode) {
1691     List<NumberFormat> intlNumberFormats = metadata.intlNumberFormats();
1692     // When the intlNumberFormats exists, we use that to format national number for the
1693     // INTERNATIONAL format instead of using the numberDesc.numberFormats.
1694     List<NumberFormat> availableFormats =
1695         (intlNumberFormats.size() == 0 || numberFormat == PhoneNumberFormat.NATIONAL)
1696         ? metadata.numberFormats()
1697         : metadata.intlNumberFormats();
1698     NumberFormat formattingPattern = chooseFormattingPatternForNumber(availableFormats, number);
1699     return (formattingPattern == null)
1700         ? number
1701         : formatNsnUsingPattern(number, formattingPattern, numberFormat, carrierCode);
1702   }
1703 
chooseFormattingPatternForNumber(List<NumberFormat> availableFormats, String nationalNumber)1704   private NumberFormat chooseFormattingPatternForNumber(List<NumberFormat> availableFormats,
1705                                                         String nationalNumber) {
1706     for (NumberFormat numFormat : availableFormats) {
1707       int size = numFormat.leadingDigitsPatternSize();
1708       if (size == 0 || regexCache.getPatternForRegex(
1709               // We always use the last leading_digits_pattern, as it is the most detailed.
1710               numFormat.getLeadingDigitsPattern(size - 1)).matcher(nationalNumber).lookingAt()) {
1711         Matcher m = regexCache.getPatternForRegex(numFormat.getPattern()).matcher(nationalNumber);
1712         if (m.matches()) {
1713           return numFormat;
1714         }
1715       }
1716     }
1717     return null;
1718   }
1719 
1720   // Simple wrapper of formatNsnUsingPattern for the common case of no carrier code.
formatNsnUsingPattern(String nationalNumber, NumberFormat formattingPattern, PhoneNumberFormat numberFormat)1721   private String formatNsnUsingPattern(String nationalNumber,
1722                                        NumberFormat formattingPattern,
1723                                        PhoneNumberFormat numberFormat) {
1724     return formatNsnUsingPattern(nationalNumber, formattingPattern, numberFormat, null);
1725   }
1726 
1727   // Note that carrierCode is optional - if NULL or an empty string, no carrier code replacement
1728   // will take place.
formatNsnUsingPattern(String nationalNumber, NumberFormat formattingPattern, PhoneNumberFormat numberFormat, String carrierCode)1729   private String formatNsnUsingPattern(String nationalNumber,
1730                                        NumberFormat formattingPattern,
1731                                        PhoneNumberFormat numberFormat,
1732                                        String carrierCode) {
1733     String numberFormatRule = formattingPattern.getFormat();
1734     Matcher m =
1735         regexCache.getPatternForRegex(formattingPattern.getPattern()).matcher(nationalNumber);
1736     String formattedNationalNumber = "";
1737     if (numberFormat == PhoneNumberFormat.NATIONAL &&
1738         carrierCode != null && carrierCode.length() > 0 &&
1739         formattingPattern.getDomesticCarrierCodeFormattingRule().length() > 0) {
1740       // Replace the $CC in the formatting rule with the desired carrier code.
1741       String carrierCodeFormattingRule = formattingPattern.getDomesticCarrierCodeFormattingRule();
1742       carrierCodeFormattingRule =
1743           CC_PATTERN.matcher(carrierCodeFormattingRule).replaceFirst(carrierCode);
1744       // Now replace the $FG in the formatting rule with the first group and the carrier code
1745       // combined in the appropriate way.
1746       numberFormatRule = FIRST_GROUP_PATTERN.matcher(numberFormatRule)
1747           .replaceFirst(carrierCodeFormattingRule);
1748       formattedNationalNumber = m.replaceAll(numberFormatRule);
1749     } else {
1750       // Use the national prefix formatting rule instead.
1751       String nationalPrefixFormattingRule = formattingPattern.getNationalPrefixFormattingRule();
1752       if (numberFormat == PhoneNumberFormat.NATIONAL &&
1753           nationalPrefixFormattingRule != null &&
1754           nationalPrefixFormattingRule.length() > 0) {
1755         Matcher firstGroupMatcher = FIRST_GROUP_PATTERN.matcher(numberFormatRule);
1756         formattedNationalNumber =
1757             m.replaceAll(firstGroupMatcher.replaceFirst(nationalPrefixFormattingRule));
1758       } else {
1759         formattedNationalNumber = m.replaceAll(numberFormatRule);
1760       }
1761     }
1762     if (numberFormat == PhoneNumberFormat.RFC3966) {
1763       // Strip any leading punctuation.
1764       Matcher matcher = SEPARATOR_PATTERN.matcher(formattedNationalNumber);
1765       if (matcher.lookingAt()) {
1766         formattedNationalNumber = matcher.replaceFirst("");
1767       }
1768       // Replace the rest with a dash between each number group.
1769       formattedNationalNumber = matcher.reset(formattedNationalNumber).replaceAll("-");
1770     }
1771     return formattedNationalNumber;
1772   }
1773 
1774   /**
1775    * Gets a valid number for the specified region.
1776    *
1777    * @param regionCode  the region for which an example number is needed
1778    * @return  a valid fixed-line number for the specified region. Returns null when the metadata
1779    *    does not contain such information, or the region 001 is passed in. For 001 (representing
1780    *    non-geographical numbers), call {@link #getExampleNumberForNonGeoEntity} instead.
1781    */
getExampleNumber(String regionCode)1782   public PhoneNumber getExampleNumber(String regionCode) {
1783     return getExampleNumberForType(regionCode, PhoneNumberType.FIXED_LINE);
1784   }
1785 
1786   /**
1787    * Gets a valid number for the specified region and number type.
1788    *
1789    * @param regionCode  the region for which an example number is needed
1790    * @param type  the type of number that is needed
1791    * @return  a valid number for the specified region and type. Returns null when the metadata
1792    *     does not contain such information or if an invalid region or region 001 was entered.
1793    *     For 001 (representing non-geographical numbers), call
1794    *     {@link #getExampleNumberForNonGeoEntity} instead.
1795    */
getExampleNumberForType(String regionCode, PhoneNumberType type)1796   public PhoneNumber getExampleNumberForType(String regionCode, PhoneNumberType type) {
1797     // Check the region code is valid.
1798     if (!isValidRegionCode(regionCode)) {
1799       LOGGER.log(Level.WARNING, "Invalid or unknown region code provided: " + regionCode);
1800       return null;
1801     }
1802     PhoneNumberDesc desc = getNumberDescByType(getMetadataForRegion(regionCode), type);
1803     try {
1804       if (desc.hasExampleNumber()) {
1805         return parse(desc.getExampleNumber(), regionCode);
1806       }
1807     } catch (NumberParseException e) {
1808       LOGGER.log(Level.SEVERE, e.toString());
1809     }
1810     return null;
1811   }
1812 
1813   /**
1814    * Gets a valid number for the specified country calling code for a non-geographical entity.
1815    *
1816    * @param countryCallingCode  the country calling code for a non-geographical entity
1817    * @return  a valid number for the non-geographical entity. Returns null when the metadata
1818    *    does not contain such information, or the country calling code passed in does not belong
1819    *    to a non-geographical entity.
1820    */
getExampleNumberForNonGeoEntity(int countryCallingCode)1821   public PhoneNumber getExampleNumberForNonGeoEntity(int countryCallingCode) {
1822     PhoneMetadata metadata = getMetadataForNonGeographicalRegion(countryCallingCode);
1823     if (metadata != null) {
1824       PhoneNumberDesc desc = metadata.getGeneralDesc();
1825       try {
1826         if (desc.hasExampleNumber()) {
1827           return parse("+" + countryCallingCode + desc.getExampleNumber(), "ZZ");
1828         }
1829       } catch (NumberParseException e) {
1830         LOGGER.log(Level.SEVERE, e.toString());
1831       }
1832     } else {
1833       LOGGER.log(Level.WARNING,
1834                  "Invalid or unknown country calling code provided: " + countryCallingCode);
1835     }
1836     return null;
1837   }
1838 
1839   /**
1840    * Appends the formatted extension of a phone number to formattedNumber, if the phone number had
1841    * an extension specified.
1842    */
maybeAppendFormattedExtension(PhoneNumber number, PhoneMetadata metadata, PhoneNumberFormat numberFormat, StringBuilder formattedNumber)1843   private void maybeAppendFormattedExtension(PhoneNumber number, PhoneMetadata metadata,
1844                                              PhoneNumberFormat numberFormat,
1845                                              StringBuilder formattedNumber) {
1846     if (number.hasExtension() && number.getExtension().length() > 0) {
1847       if (numberFormat == PhoneNumberFormat.RFC3966) {
1848         formattedNumber.append(RFC3966_EXTN_PREFIX).append(number.getExtension());
1849       } else {
1850         if (metadata.hasPreferredExtnPrefix()) {
1851           formattedNumber.append(metadata.getPreferredExtnPrefix()).append(number.getExtension());
1852         } else {
1853           formattedNumber.append(DEFAULT_EXTN_PREFIX).append(number.getExtension());
1854         }
1855       }
1856     }
1857   }
1858 
getNumberDescByType(PhoneMetadata metadata, PhoneNumberType type)1859   PhoneNumberDesc getNumberDescByType(PhoneMetadata metadata, PhoneNumberType type) {
1860     switch (type) {
1861       case PREMIUM_RATE:
1862         return metadata.getPremiumRate();
1863       case TOLL_FREE:
1864         return metadata.getTollFree();
1865       case MOBILE:
1866         return metadata.getMobile();
1867       case FIXED_LINE:
1868       case FIXED_LINE_OR_MOBILE:
1869         return metadata.getFixedLine();
1870       case SHARED_COST:
1871         return metadata.getSharedCost();
1872       case VOIP:
1873         return metadata.getVoip();
1874       case PERSONAL_NUMBER:
1875         return metadata.getPersonalNumber();
1876       case PAGER:
1877         return metadata.getPager();
1878       case UAN:
1879         return metadata.getUan();
1880       case VOICEMAIL:
1881         return metadata.getVoicemail();
1882       default:
1883         return metadata.getGeneralDesc();
1884     }
1885   }
1886 
1887   /**
1888    * Gets the type of a phone number.
1889    *
1890    * @param number  the phone number that we want to know the type
1891    * @return  the type of the phone number
1892    */
getNumberType(PhoneNumber number)1893   public PhoneNumberType getNumberType(PhoneNumber number) {
1894     String regionCode = getRegionCodeForNumber(number);
1895     if (!isValidRegionCode(regionCode) && !REGION_CODE_FOR_NON_GEO_ENTITY.equals(regionCode)) {
1896       return PhoneNumberType.UNKNOWN;
1897     }
1898     String nationalSignificantNumber = getNationalSignificantNumber(number);
1899     PhoneMetadata metadata = getMetadataForRegionOrCallingCode(number.getCountryCode(), regionCode);
1900     return getNumberTypeHelper(nationalSignificantNumber, metadata);
1901   }
1902 
getNumberTypeHelper(String nationalNumber, PhoneMetadata metadata)1903   private PhoneNumberType getNumberTypeHelper(String nationalNumber, PhoneMetadata metadata) {
1904     PhoneNumberDesc generalNumberDesc = metadata.getGeneralDesc();
1905     if (!generalNumberDesc.hasNationalNumberPattern() ||
1906         !isNumberMatchingDesc(nationalNumber, generalNumberDesc)) {
1907       return PhoneNumberType.UNKNOWN;
1908     }
1909 
1910     if (isNumberMatchingDesc(nationalNumber, metadata.getPremiumRate())) {
1911       return PhoneNumberType.PREMIUM_RATE;
1912     }
1913     if (isNumberMatchingDesc(nationalNumber, metadata.getTollFree())) {
1914       return PhoneNumberType.TOLL_FREE;
1915     }
1916     if (isNumberMatchingDesc(nationalNumber, metadata.getSharedCost())) {
1917       return PhoneNumberType.SHARED_COST;
1918     }
1919     if (isNumberMatchingDesc(nationalNumber, metadata.getVoip())) {
1920       return PhoneNumberType.VOIP;
1921     }
1922     if (isNumberMatchingDesc(nationalNumber, metadata.getPersonalNumber())) {
1923       return PhoneNumberType.PERSONAL_NUMBER;
1924     }
1925     if (isNumberMatchingDesc(nationalNumber, metadata.getPager())) {
1926       return PhoneNumberType.PAGER;
1927     }
1928     if (isNumberMatchingDesc(nationalNumber, metadata.getUan())) {
1929       return PhoneNumberType.UAN;
1930     }
1931     if (isNumberMatchingDesc(nationalNumber, metadata.getVoicemail())) {
1932       return PhoneNumberType.VOICEMAIL;
1933     }
1934 
1935     boolean isFixedLine = isNumberMatchingDesc(nationalNumber, metadata.getFixedLine());
1936     if (isFixedLine) {
1937       if (metadata.isSameMobileAndFixedLinePattern()) {
1938         return PhoneNumberType.FIXED_LINE_OR_MOBILE;
1939       } else if (isNumberMatchingDesc(nationalNumber, metadata.getMobile())) {
1940         return PhoneNumberType.FIXED_LINE_OR_MOBILE;
1941       }
1942       return PhoneNumberType.FIXED_LINE;
1943     }
1944     // Otherwise, test to see if the number is mobile. Only do this if certain that the patterns for
1945     // mobile and fixed line aren't the same.
1946     if (!metadata.isSameMobileAndFixedLinePattern() &&
1947         isNumberMatchingDesc(nationalNumber, metadata.getMobile())) {
1948       return PhoneNumberType.MOBILE;
1949     }
1950     return PhoneNumberType.UNKNOWN;
1951   }
1952 
getMetadataForRegion(String regionCode)1953   PhoneMetadata getMetadataForRegion(String regionCode) {
1954     if (!isValidRegionCode(regionCode)) {
1955       return null;
1956     }
1957     synchronized (regionToMetadataMap) {
1958       if (!regionToMetadataMap.containsKey(regionCode)) {
1959         // The regionCode here will be valid and won't be '001', so we don't need to worry about
1960         // what to pass in for the country calling code.
1961         loadMetadataFromFile(currentFilePrefix, regionCode, 0);
1962       }
1963     }
1964     return regionToMetadataMap.get(regionCode);
1965   }
1966 
getMetadataForNonGeographicalRegion(int countryCallingCode)1967   PhoneMetadata getMetadataForNonGeographicalRegion(int countryCallingCode) {
1968     synchronized (countryCodeToNonGeographicalMetadataMap) {
1969       if (!countryCallingCodeToRegionCodeMap.containsKey(countryCallingCode)) {
1970         return null;
1971       }
1972       if (!countryCodeToNonGeographicalMetadataMap.containsKey(countryCallingCode)) {
1973         loadMetadataFromFile(currentFilePrefix, REGION_CODE_FOR_NON_GEO_ENTITY, countryCallingCode);
1974       }
1975     }
1976     return countryCodeToNonGeographicalMetadataMap.get(countryCallingCode);
1977   }
1978 
isNumberMatchingDesc(String nationalNumber, PhoneNumberDesc numberDesc)1979   private boolean isNumberMatchingDesc(String nationalNumber, PhoneNumberDesc numberDesc) {
1980     Matcher possibleNumberPatternMatcher =
1981         regexCache.getPatternForRegex(numberDesc.getPossibleNumberPattern())
1982             .matcher(nationalNumber);
1983     Matcher nationalNumberPatternMatcher =
1984         regexCache.getPatternForRegex(numberDesc.getNationalNumberPattern())
1985             .matcher(nationalNumber);
1986     return possibleNumberPatternMatcher.matches() && nationalNumberPatternMatcher.matches();
1987   }
1988 
1989   /**
1990    * Tests whether a phone number matches a valid pattern. Note this doesn't verify the number
1991    * is actually in use, which is impossible to tell by just looking at a number itself.
1992    *
1993    * @param number       the phone number that we want to validate
1994    * @return  a boolean that indicates whether the number is of a valid pattern
1995    */
isValidNumber(PhoneNumber number)1996   public boolean isValidNumber(PhoneNumber number) {
1997     String regionCode = getRegionCodeForNumber(number);
1998     return isValidNumberForRegion(number, regionCode);
1999   }
2000 
2001   /**
2002    * Tests whether a phone number is valid for a certain region. Note this doesn't verify the number
2003    * is actually in use, which is impossible to tell by just looking at a number itself. If the
2004    * country calling code is not the same as the country calling code for the region, this
2005    * immediately exits with false. After this, the specific number pattern rules for the region are
2006    * examined. This is useful for determining for example whether a particular number is valid for
2007    * Canada, rather than just a valid NANPA number.
2008    *
2009    * @param number       the phone number that we want to validate
2010    * @param regionCode   the region that we want to validate the phone number for
2011    * @return  a boolean that indicates whether the number is of a valid pattern
2012    */
isValidNumberForRegion(PhoneNumber number, String regionCode)2013   public boolean isValidNumberForRegion(PhoneNumber number, String regionCode) {
2014     int countryCode = number.getCountryCode();
2015     PhoneMetadata metadata = getMetadataForRegionOrCallingCode(countryCode, regionCode);
2016     if ((metadata == null) ||
2017         (!REGION_CODE_FOR_NON_GEO_ENTITY.equals(regionCode) &&
2018          countryCode != getCountryCodeForValidRegion(regionCode))) {
2019       // Either the region code was invalid, or the country calling code for this number does not
2020       // match that of the region code.
2021       return false;
2022     }
2023     PhoneNumberDesc generalNumDesc = metadata.getGeneralDesc();
2024     String nationalSignificantNumber = getNationalSignificantNumber(number);
2025 
2026     // For regions where we don't have metadata for PhoneNumberDesc, we treat any number passed in
2027     // as a valid number if its national significant number is between the minimum and maximum
2028     // lengths defined by ITU for a national significant number.
2029     if (!generalNumDesc.hasNationalNumberPattern()) {
2030       int numberLength = nationalSignificantNumber.length();
2031       return numberLength > MIN_LENGTH_FOR_NSN && numberLength <= MAX_LENGTH_FOR_NSN;
2032     }
2033     return getNumberTypeHelper(nationalSignificantNumber, metadata) != PhoneNumberType.UNKNOWN;
2034   }
2035 
2036   /**
2037    * Returns the region where a phone number is from. This could be used for geocoding at the region
2038    * level.
2039    *
2040    * @param number  the phone number whose origin we want to know
2041    * @return  the region where the phone number is from, or null if no region matches this calling
2042    *     code
2043    */
getRegionCodeForNumber(PhoneNumber number)2044   public String getRegionCodeForNumber(PhoneNumber number) {
2045     int countryCode = number.getCountryCode();
2046     List<String> regions = countryCallingCodeToRegionCodeMap.get(countryCode);
2047     if (regions == null) {
2048       String numberString = getNationalSignificantNumber(number);
2049       LOGGER.log(Level.WARNING,
2050                  "Missing/invalid country_code (" + countryCode + ") for number " + numberString);
2051       return null;
2052     }
2053     if (regions.size() == 1) {
2054       return regions.get(0);
2055     } else {
2056       return getRegionCodeForNumberFromRegionList(number, regions);
2057     }
2058   }
2059 
getRegionCodeForNumberFromRegionList(PhoneNumber number, List<String> regionCodes)2060   private String getRegionCodeForNumberFromRegionList(PhoneNumber number,
2061                                                       List<String> regionCodes) {
2062     String nationalNumber = getNationalSignificantNumber(number);
2063     for (String regionCode : regionCodes) {
2064       // If leadingDigits is present, use this. Otherwise, do full validation.
2065       PhoneMetadata metadata = getMetadataForRegion(regionCode);
2066       if (metadata.hasLeadingDigits()) {
2067         if (regexCache.getPatternForRegex(metadata.getLeadingDigits())
2068                 .matcher(nationalNumber).lookingAt()) {
2069           return regionCode;
2070         }
2071       } else if (getNumberTypeHelper(nationalNumber, metadata) != PhoneNumberType.UNKNOWN) {
2072         return regionCode;
2073       }
2074     }
2075     return null;
2076   }
2077 
2078   /**
2079    * Returns the region code that matches the specific country calling code. In the case of no
2080    * region code being found, ZZ will be returned. In the case of multiple regions, the one
2081    * designated in the metadata as the "main" region for this calling code will be returned.
2082    */
getRegionCodeForCountryCode(int countryCallingCode)2083   public String getRegionCodeForCountryCode(int countryCallingCode) {
2084     List<String> regionCodes = countryCallingCodeToRegionCodeMap.get(countryCallingCode);
2085     return regionCodes == null ? UNKNOWN_REGION : regionCodes.get(0);
2086   }
2087 
2088   /**
2089    * Returns the country calling code for a specific region. For example, this would be 1 for the
2090    * United States, and 64 for New Zealand.
2091    *
2092    * @param regionCode  the region that we want to get the country calling code for
2093    * @return  the country calling code for the region denoted by regionCode
2094    */
getCountryCodeForRegion(String regionCode)2095   public int getCountryCodeForRegion(String regionCode) {
2096     if (!isValidRegionCode(regionCode)) {
2097       LOGGER.log(Level.WARNING,
2098                  "Invalid or missing region code ("
2099                   + ((regionCode == null) ? "null" : regionCode)
2100                   + ") provided.");
2101       return 0;
2102     }
2103     return getCountryCodeForValidRegion(regionCode);
2104   }
2105 
2106   /**
2107    * Returns the country calling code for a specific region. For example, this would be 1 for the
2108    * United States, and 64 for New Zealand. Assumes the region is already valid.
2109    *
2110    * @param regionCode  the region that we want to get the country calling code for
2111    * @return  the country calling code for the region denoted by regionCode
2112    */
getCountryCodeForValidRegion(String regionCode)2113   private int getCountryCodeForValidRegion(String regionCode) {
2114     PhoneMetadata metadata = getMetadataForRegion(regionCode);
2115     return metadata.getCountryCode();
2116   }
2117 
2118   /**
2119    * Returns the national dialling prefix for a specific region. For example, this would be 1 for
2120    * the United States, and 0 for New Zealand. Set stripNonDigits to true to strip symbols like "~"
2121    * (which indicates a wait for a dialling tone) from the prefix returned. If no national prefix is
2122    * present, we return null.
2123    *
2124    * <p>Warning: Do not use this method for do-your-own formatting - for some regions, the
2125    * national dialling prefix is used only for certain types of numbers. Use the library's
2126    * formatting functions to prefix the national prefix when required.
2127    *
2128    * @param regionCode  the region that we want to get the dialling prefix for
2129    * @param stripNonDigits  true to strip non-digits from the national dialling prefix
2130    * @return  the dialling prefix for the region denoted by regionCode
2131    */
getNddPrefixForRegion(String regionCode, boolean stripNonDigits)2132   public String getNddPrefixForRegion(String regionCode, boolean stripNonDigits) {
2133     if (!isValidRegionCode(regionCode)) {
2134       LOGGER.log(Level.WARNING,
2135                  "Invalid or missing region code ("
2136                   + ((regionCode == null) ? "null" : regionCode)
2137                   + ") provided.");
2138       return null;
2139     }
2140     PhoneMetadata metadata = getMetadataForRegion(regionCode);
2141     String nationalPrefix = metadata.getNationalPrefix();
2142     // If no national prefix was found, we return null.
2143     if (nationalPrefix.length() == 0) {
2144       return null;
2145     }
2146     if (stripNonDigits) {
2147       // Note: if any other non-numeric symbols are ever used in national prefixes, these would have
2148       // to be removed here as well.
2149       nationalPrefix = nationalPrefix.replace("~", "");
2150     }
2151     return nationalPrefix;
2152   }
2153 
2154   /**
2155    * Checks if this is a region under the North American Numbering Plan Administration (NANPA).
2156    *
2157    * @return  true if regionCode is one of the regions under NANPA
2158    */
isNANPACountry(String regionCode)2159   public boolean isNANPACountry(String regionCode) {
2160     return nanpaRegions.contains(regionCode);
2161   }
2162 
2163   /**
2164    * Checks whether the country calling code is from a region whose national significant number
2165    * could contain a leading zero. An example of such a region is Italy. Returns false if no
2166    * metadata for the country is found.
2167    */
isLeadingZeroPossible(int countryCallingCode)2168   boolean isLeadingZeroPossible(int countryCallingCode) {
2169     PhoneMetadata mainMetadataForCallingCode = getMetadataForRegion(
2170         getRegionCodeForCountryCode(countryCallingCode));
2171     if (mainMetadataForCallingCode == null) {
2172       return false;
2173     }
2174     return mainMetadataForCallingCode.isLeadingZeroPossible();
2175   }
2176 
2177   /**
2178    * Checks if the number is a valid vanity (alpha) number such as 800 MICROSOFT. A valid vanity
2179    * number will start with at least 3 digits and will have three or more alpha characters. This
2180    * does not do region-specific checks - to work out if this number is actually valid for a region,
2181    * it should be parsed and methods such as {@link #isPossibleNumberWithReason} and
2182    * {@link #isValidNumber} should be used.
2183    *
2184    * @param number  the number that needs to be checked
2185    * @return  true if the number is a valid vanity number
2186    */
isAlphaNumber(String number)2187   public boolean isAlphaNumber(String number) {
2188     if (!isViablePhoneNumber(number)) {
2189       // Number is too short, or doesn't match the basic phone number pattern.
2190       return false;
2191     }
2192     StringBuilder strippedNumber = new StringBuilder(number);
2193     maybeStripExtension(strippedNumber);
2194     return VALID_ALPHA_PHONE_PATTERN.matcher(strippedNumber).matches();
2195   }
2196 
2197   /**
2198    * Convenience wrapper around {@link #isPossibleNumberWithReason}. Instead of returning the reason
2199    * for failure, this method returns a boolean value.
2200    * @param number  the number that needs to be checked
2201    * @return  true if the number is possible
2202    */
isPossibleNumber(PhoneNumber number)2203   public boolean isPossibleNumber(PhoneNumber number) {
2204     return isPossibleNumberWithReason(number) == ValidationResult.IS_POSSIBLE;
2205   }
2206 
2207   /**
2208    * Helper method to check a number against a particular pattern and determine whether it matches,
2209    * or is too short or too long. Currently, if a number pattern suggests that numbers of length 7
2210    * and 10 are possible, and a number in between these possible lengths is entered, such as of
2211    * length 8, this will return TOO_LONG.
2212    */
testNumberLengthAgainstPattern(Pattern numberPattern, String number)2213   private ValidationResult testNumberLengthAgainstPattern(Pattern numberPattern, String number) {
2214     Matcher numberMatcher = numberPattern.matcher(number);
2215     if (numberMatcher.matches()) {
2216       return ValidationResult.IS_POSSIBLE;
2217     }
2218     if (numberMatcher.lookingAt()) {
2219       return ValidationResult.TOO_LONG;
2220     } else {
2221       return ValidationResult.TOO_SHORT;
2222     }
2223   }
2224 
2225   /**
2226    * Check whether a phone number is a possible number. It provides a more lenient check than
2227    * {@link #isValidNumber} in the following sense:
2228    *<ol>
2229    * <li> It only checks the length of phone numbers. In particular, it doesn't check starting
2230    *      digits of the number.
2231    * <li> It doesn't attempt to figure out the type of the number, but uses general rules which
2232    *      applies to all types of phone numbers in a region. Therefore, it is much faster than
2233    *      isValidNumber.
2234    * <li> For fixed line numbers, many regions have the concept of area code, which together with
2235    *      subscriber number constitute the national significant number. It is sometimes okay to dial
2236    *      the subscriber number only when dialing in the same area. This function will return
2237    *      true if the subscriber-number-only version is passed in. On the other hand, because
2238    *      isValidNumber validates using information on both starting digits (for fixed line
2239    *      numbers, that would most likely be area codes) and length (obviously includes the
2240    *      length of area codes for fixed line numbers), it will return false for the
2241    *      subscriber-number-only version.
2242    * </ol
2243    * @param number  the number that needs to be checked
2244    * @return  a ValidationResult object which indicates whether the number is possible
2245    */
isPossibleNumberWithReason(PhoneNumber number)2246   public ValidationResult isPossibleNumberWithReason(PhoneNumber number) {
2247     String nationalNumber = getNationalSignificantNumber(number);
2248     int countryCode = number.getCountryCode();
2249     // Note: For Russian Fed and NANPA numbers, we just use the rules from the default region (US or
2250     // Russia) since the getRegionCodeForNumber will not work if the number is possible but not
2251     // valid. This would need to be revisited if the possible number pattern ever differed between
2252     // various regions within those plans.
2253     if (!hasValidCountryCallingCode(countryCode)) {
2254       return ValidationResult.INVALID_COUNTRY_CODE;
2255     }
2256     String regionCode = getRegionCodeForCountryCode(countryCode);
2257     PhoneMetadata metadata = getMetadataForRegionOrCallingCode(countryCode, regionCode);
2258     PhoneNumberDesc generalNumDesc = metadata.getGeneralDesc();
2259     // Handling case of numbers with no metadata.
2260     if (!generalNumDesc.hasNationalNumberPattern()) {
2261       LOGGER.log(Level.FINER, "Checking if number is possible with incomplete metadata.");
2262       int numberLength = nationalNumber.length();
2263       if (numberLength < MIN_LENGTH_FOR_NSN) {
2264         return ValidationResult.TOO_SHORT;
2265       } else if (numberLength > MAX_LENGTH_FOR_NSN) {
2266         return ValidationResult.TOO_LONG;
2267       } else {
2268         return ValidationResult.IS_POSSIBLE;
2269       }
2270     }
2271     Pattern possibleNumberPattern =
2272         regexCache.getPatternForRegex(generalNumDesc.getPossibleNumberPattern());
2273     return testNumberLengthAgainstPattern(possibleNumberPattern, nationalNumber);
2274   }
2275 
2276   /**
2277    * Check whether a phone number is a possible number given a number in the form of a string, and
2278    * the region where the number could be dialed from. It provides a more lenient check than
2279    * {@link #isValidNumber}. See {@link #isPossibleNumber(PhoneNumber)} for details.
2280    *
2281    * <p>This method first parses the number, then invokes {@link #isPossibleNumber(PhoneNumber)}
2282    * with the resultant PhoneNumber object.
2283    *
2284    * @param number  the number that needs to be checked, in the form of a string
2285    * @param regionDialingFrom  the region that we are expecting the number to be dialed from.
2286    *     Note this is different from the region where the number belongs.  For example, the number
2287    *     +1 650 253 0000 is a number that belongs to US. When written in this form, it can be
2288    *     dialed from any region. When it is written as 00 1 650 253 0000, it can be dialed from any
2289    *     region which uses an international dialling prefix of 00. When it is written as
2290    *     650 253 0000, it can only be dialed from within the US, and when written as 253 0000, it
2291    *     can only be dialed from within a smaller area in the US (Mountain View, CA, to be more
2292    *     specific).
2293    * @return  true if the number is possible
2294    */
isPossibleNumber(String number, String regionDialingFrom)2295   public boolean isPossibleNumber(String number, String regionDialingFrom) {
2296     try {
2297       return isPossibleNumber(parse(number, regionDialingFrom));
2298     } catch (NumberParseException e) {
2299       return false;
2300     }
2301   }
2302 
2303   /**
2304    * Attempts to extract a valid number from a phone number that is too long to be valid, and resets
2305    * the PhoneNumber object passed in to that valid version. If no valid number could be extracted,
2306    * the PhoneNumber object passed in will not be modified.
2307    * @param number a PhoneNumber object which contains a number that is too long to be valid.
2308    * @return  true if a valid phone number can be successfully extracted.
2309    */
truncateTooLongNumber(PhoneNumber number)2310   public boolean truncateTooLongNumber(PhoneNumber number) {
2311     if (isValidNumber(number)) {
2312       return true;
2313     }
2314     PhoneNumber numberCopy = new PhoneNumber();
2315     numberCopy.mergeFrom(number);
2316     long nationalNumber = number.getNationalNumber();
2317     do {
2318       nationalNumber /= 10;
2319       numberCopy.setNationalNumber(nationalNumber);
2320       if (isPossibleNumberWithReason(numberCopy) == ValidationResult.TOO_SHORT ||
2321           nationalNumber == 0) {
2322         return false;
2323       }
2324     } while (!isValidNumber(numberCopy));
2325     number.setNationalNumber(nationalNumber);
2326     return true;
2327   }
2328 
2329   /**
2330    * Gets an {@link com.android.i18n.phonenumbers.AsYouTypeFormatter} for the specific region.
2331    *
2332    * @param regionCode  the region where the phone number is being entered
2333    * @return  an {@link com.android.i18n.phonenumbers.AsYouTypeFormatter} object, which can be used
2334    *     to format phone numbers in the specific region "as you type"
2335    */
getAsYouTypeFormatter(String regionCode)2336   public AsYouTypeFormatter getAsYouTypeFormatter(String regionCode) {
2337     return new AsYouTypeFormatter(regionCode);
2338   }
2339 
2340   // Extracts country calling code from fullNumber, returns it and places the remaining number in
2341   // nationalNumber. It assumes that the leading plus sign or IDD has already been removed. Returns
2342   // 0 if fullNumber doesn't start with a valid country calling code, and leaves nationalNumber
2343   // unmodified.
extractCountryCode(StringBuilder fullNumber, StringBuilder nationalNumber)2344   int extractCountryCode(StringBuilder fullNumber, StringBuilder nationalNumber) {
2345     if ((fullNumber.length() == 0) || (fullNumber.charAt(0) == '0')) {
2346       // Country codes do not begin with a '0'.
2347       return 0;
2348     }
2349     int potentialCountryCode;
2350     int numberLength = fullNumber.length();
2351     for (int i = 1; i <= MAX_LENGTH_COUNTRY_CODE && i <= numberLength; i++) {
2352       potentialCountryCode = Integer.parseInt(fullNumber.substring(0, i));
2353       if (countryCallingCodeToRegionCodeMap.containsKey(potentialCountryCode)) {
2354         nationalNumber.append(fullNumber.substring(i));
2355         return potentialCountryCode;
2356       }
2357     }
2358     return 0;
2359   }
2360 
2361   /**
2362    * Tries to extract a country calling code from a number. This method will return zero if no
2363    * country calling code is considered to be present. Country calling codes are extracted in the
2364    * following ways:
2365    * <ul>
2366    *  <li> by stripping the international dialing prefix of the region the person is dialing from,
2367    *       if this is present in the number, and looking at the next digits
2368    *  <li> by stripping the '+' sign if present and then looking at the next digits
2369    *  <li> by comparing the start of the number and the country calling code of the default region.
2370    *       If the number is not considered possible for the numbering plan of the default region
2371    *       initially, but starts with the country calling code of this region, validation will be
2372    *       reattempted after stripping this country calling code. If this number is considered a
2373    *       possible number, then the first digits will be considered the country calling code and
2374    *       removed as such.
2375    * </ul>
2376    * It will throw a NumberParseException if the number starts with a '+' but the country calling
2377    * code supplied after this does not match that of any known region.
2378    *
2379    * @param number  non-normalized telephone number that we wish to extract a country calling
2380    *     code from - may begin with '+'
2381    * @param defaultRegionMetadata  metadata about the region this number may be from
2382    * @param nationalNumber  a string buffer to store the national significant number in, in the case
2383    *     that a country calling code was extracted. The number is appended to any existing contents.
2384    *     If no country calling code was extracted, this will be left unchanged.
2385    * @param keepRawInput  true if the country_code_source and preferred_carrier_code fields of
2386    *     phoneNumber should be populated.
2387    * @param phoneNumber  the PhoneNumber object where the country_code and country_code_source need
2388    *     to be populated. Note the country_code is always populated, whereas country_code_source is
2389    *     only populated when keepCountryCodeSource is true.
2390    * @return  the country calling code extracted or 0 if none could be extracted
2391    */
2392   // @VisibleForTesting
maybeExtractCountryCode(String number, PhoneMetadata defaultRegionMetadata, StringBuilder nationalNumber, boolean keepRawInput, PhoneNumber phoneNumber)2393   int maybeExtractCountryCode(String number, PhoneMetadata defaultRegionMetadata,
2394                               StringBuilder nationalNumber, boolean keepRawInput,
2395                               PhoneNumber phoneNumber)
2396       throws NumberParseException {
2397     if (number.length() == 0) {
2398       return 0;
2399     }
2400     StringBuilder fullNumber = new StringBuilder(number);
2401     // Set the default prefix to be something that will never match.
2402     String possibleCountryIddPrefix = "NonMatch";
2403     if (defaultRegionMetadata != null) {
2404       possibleCountryIddPrefix = defaultRegionMetadata.getInternationalPrefix();
2405     }
2406 
2407     CountryCodeSource countryCodeSource =
2408         maybeStripInternationalPrefixAndNormalize(fullNumber, possibleCountryIddPrefix);
2409     if (keepRawInput) {
2410       phoneNumber.setCountryCodeSource(countryCodeSource);
2411     }
2412     if (countryCodeSource != CountryCodeSource.FROM_DEFAULT_COUNTRY) {
2413       if (fullNumber.length() < MIN_LENGTH_FOR_NSN) {
2414         throw new NumberParseException(NumberParseException.ErrorType.TOO_SHORT_AFTER_IDD,
2415                                        "Phone number had an IDD, but after this was not "
2416                                        + "long enough to be a viable phone number.");
2417       }
2418       int potentialCountryCode = extractCountryCode(fullNumber, nationalNumber);
2419       if (potentialCountryCode != 0) {
2420         phoneNumber.setCountryCode(potentialCountryCode);
2421         return potentialCountryCode;
2422       }
2423 
2424       // If this fails, they must be using a strange country calling code that we don't recognize,
2425       // or that doesn't exist.
2426       throw new NumberParseException(NumberParseException.ErrorType.INVALID_COUNTRY_CODE,
2427                                      "Country calling code supplied was not recognised.");
2428     } else if (defaultRegionMetadata != null) {
2429       // Check to see if the number starts with the country calling code for the default region. If
2430       // so, we remove the country calling code, and do some checks on the validity of the number
2431       // before and after.
2432       int defaultCountryCode = defaultRegionMetadata.getCountryCode();
2433       String defaultCountryCodeString = String.valueOf(defaultCountryCode);
2434       String normalizedNumber = fullNumber.toString();
2435       if (normalizedNumber.startsWith(defaultCountryCodeString)) {
2436         StringBuilder potentialNationalNumber =
2437             new StringBuilder(normalizedNumber.substring(defaultCountryCodeString.length()));
2438         PhoneNumberDesc generalDesc = defaultRegionMetadata.getGeneralDesc();
2439         Pattern validNumberPattern =
2440             regexCache.getPatternForRegex(generalDesc.getNationalNumberPattern());
2441         maybeStripNationalPrefixAndCarrierCode(
2442             potentialNationalNumber, defaultRegionMetadata, null /* Don't need the carrier code */);
2443         Pattern possibleNumberPattern =
2444             regexCache.getPatternForRegex(generalDesc.getPossibleNumberPattern());
2445         // If the number was not valid before but is valid now, or if it was too long before, we
2446         // consider the number with the country calling code stripped to be a better result and
2447         // keep that instead.
2448         if ((!validNumberPattern.matcher(fullNumber).matches() &&
2449              validNumberPattern.matcher(potentialNationalNumber).matches()) ||
2450              testNumberLengthAgainstPattern(possibleNumberPattern, fullNumber.toString())
2451                   == ValidationResult.TOO_LONG) {
2452           nationalNumber.append(potentialNationalNumber);
2453           if (keepRawInput) {
2454             phoneNumber.setCountryCodeSource(CountryCodeSource.FROM_NUMBER_WITHOUT_PLUS_SIGN);
2455           }
2456           phoneNumber.setCountryCode(defaultCountryCode);
2457           return defaultCountryCode;
2458         }
2459       }
2460     }
2461     // No country calling code present.
2462     phoneNumber.setCountryCode(0);
2463     return 0;
2464   }
2465 
2466   /**
2467    * Strips the IDD from the start of the number if present. Helper function used by
2468    * maybeStripInternationalPrefixAndNormalize.
2469    */
parsePrefixAsIdd(Pattern iddPattern, StringBuilder number)2470   private boolean parsePrefixAsIdd(Pattern iddPattern, StringBuilder number) {
2471     Matcher m = iddPattern.matcher(number);
2472     if (m.lookingAt()) {
2473       int matchEnd = m.end();
2474       // Only strip this if the first digit after the match is not a 0, since country calling codes
2475       // cannot begin with 0.
2476       Matcher digitMatcher = CAPTURING_DIGIT_PATTERN.matcher(number.substring(matchEnd));
2477       if (digitMatcher.find()) {
2478         String normalizedGroup = normalizeDigitsOnly(digitMatcher.group(1));
2479         if (normalizedGroup.equals("0")) {
2480           return false;
2481         }
2482       }
2483       number.delete(0, matchEnd);
2484       return true;
2485     }
2486     return false;
2487   }
2488 
2489   /**
2490    * Strips any international prefix (such as +, 00, 011) present in the number provided, normalizes
2491    * the resulting number, and indicates if an international prefix was present.
2492    *
2493    * @param number  the non-normalized telephone number that we wish to strip any international
2494    *     dialing prefix from.
2495    * @param possibleIddPrefix  the international direct dialing prefix from the region we
2496    *     think this number may be dialed in
2497    * @return  the corresponding CountryCodeSource if an international dialing prefix could be
2498    *     removed from the number, otherwise CountryCodeSource.FROM_DEFAULT_COUNTRY if the number did
2499    *     not seem to be in international format.
2500    */
2501   // @VisibleForTesting
maybeStripInternationalPrefixAndNormalize( StringBuilder number, String possibleIddPrefix)2502   CountryCodeSource maybeStripInternationalPrefixAndNormalize(
2503       StringBuilder number,
2504       String possibleIddPrefix) {
2505     if (number.length() == 0) {
2506       return CountryCodeSource.FROM_DEFAULT_COUNTRY;
2507     }
2508     // Check to see if the number begins with one or more plus signs.
2509     Matcher m = PLUS_CHARS_PATTERN.matcher(number);
2510     if (m.lookingAt()) {
2511       number.delete(0, m.end());
2512       // Can now normalize the rest of the number since we've consumed the "+" sign at the start.
2513       normalize(number);
2514       return CountryCodeSource.FROM_NUMBER_WITH_PLUS_SIGN;
2515     }
2516     // Attempt to parse the first digits as an international prefix.
2517     Pattern iddPattern = regexCache.getPatternForRegex(possibleIddPrefix);
2518     normalize(number);
2519     return parsePrefixAsIdd(iddPattern, number)
2520            ? CountryCodeSource.FROM_NUMBER_WITH_IDD
2521            : CountryCodeSource.FROM_DEFAULT_COUNTRY;
2522   }
2523 
2524   /**
2525    * Strips any national prefix (such as 0, 1) present in the number provided.
2526    *
2527    * @param number  the normalized telephone number that we wish to strip any national
2528    *     dialing prefix from
2529    * @param metadata  the metadata for the region that we think this number is from
2530    * @param carrierCode  a place to insert the carrier code if one is extracted
2531    * @return true if a national prefix or carrier code (or both) could be extracted.
2532    */
2533   // @VisibleForTesting
maybeStripNationalPrefixAndCarrierCode( StringBuilder number, PhoneMetadata metadata, StringBuilder carrierCode)2534   boolean maybeStripNationalPrefixAndCarrierCode(
2535       StringBuilder number, PhoneMetadata metadata, StringBuilder carrierCode) {
2536     int numberLength = number.length();
2537     String possibleNationalPrefix = metadata.getNationalPrefixForParsing();
2538     if (numberLength == 0 || possibleNationalPrefix.length() == 0) {
2539       // Early return for numbers of zero length.
2540       return false;
2541     }
2542     // Attempt to parse the first digits as a national prefix.
2543     Matcher prefixMatcher = regexCache.getPatternForRegex(possibleNationalPrefix).matcher(number);
2544     if (prefixMatcher.lookingAt()) {
2545       Pattern nationalNumberRule =
2546           regexCache.getPatternForRegex(metadata.getGeneralDesc().getNationalNumberPattern());
2547       // Check if the original number is viable.
2548       boolean isViableOriginalNumber = nationalNumberRule.matcher(number).matches();
2549       // prefixMatcher.group(numOfGroups) == null implies nothing was captured by the capturing
2550       // groups in possibleNationalPrefix; therefore, no transformation is necessary, and we just
2551       // remove the national prefix.
2552       int numOfGroups = prefixMatcher.groupCount();
2553       String transformRule = metadata.getNationalPrefixTransformRule();
2554       if (transformRule == null || transformRule.length() == 0 ||
2555           prefixMatcher.group(numOfGroups) == null) {
2556         // If the original number was viable, and the resultant number is not, we return.
2557         if (isViableOriginalNumber &&
2558             !nationalNumberRule.matcher(number.substring(prefixMatcher.end())).matches()) {
2559           return false;
2560         }
2561         if (carrierCode != null && numOfGroups > 0 && prefixMatcher.group(numOfGroups) != null) {
2562           carrierCode.append(prefixMatcher.group(1));
2563         }
2564         number.delete(0, prefixMatcher.end());
2565         return true;
2566       } else {
2567         // Check that the resultant number is still viable. If not, return. Check this by copying
2568         // the string buffer and making the transformation on the copy first.
2569         StringBuilder transformedNumber = new StringBuilder(number);
2570         transformedNumber.replace(0, numberLength, prefixMatcher.replaceFirst(transformRule));
2571         if (isViableOriginalNumber &&
2572             !nationalNumberRule.matcher(transformedNumber.toString()).matches()) {
2573           return false;
2574         }
2575         if (carrierCode != null && numOfGroups > 1) {
2576           carrierCode.append(prefixMatcher.group(1));
2577         }
2578         number.replace(0, number.length(), transformedNumber.toString());
2579         return true;
2580       }
2581     }
2582     return false;
2583   }
2584 
2585   /**
2586    * Strips any extension (as in, the part of the number dialled after the call is connected,
2587    * usually indicated with extn, ext, x or similar) from the end of the number, and returns it.
2588    *
2589    * @param number  the non-normalized telephone number that we wish to strip the extension from
2590    * @return        the phone extension
2591    */
2592   // @VisibleForTesting
maybeStripExtension(StringBuilder number)2593   String maybeStripExtension(StringBuilder number) {
2594     Matcher m = EXTN_PATTERN.matcher(number);
2595     // If we find a potential extension, and the number preceding this is a viable number, we assume
2596     // it is an extension.
2597     if (m.find() && isViablePhoneNumber(number.substring(0, m.start()))) {
2598       // The numbers are captured into groups in the regular expression.
2599       for (int i = 1, length = m.groupCount(); i <= length; i++) {
2600         if (m.group(i) != null) {
2601           // We go through the capturing groups until we find one that captured some digits. If none
2602           // did, then we will return the empty string.
2603           String extension = m.group(i);
2604           number.delete(m.start(), number.length());
2605           return extension;
2606         }
2607       }
2608     }
2609     return "";
2610   }
2611 
2612   /**
2613    * Checks to see that the region code used is valid, or if it is not valid, that the number to
2614    * parse starts with a + symbol so that we can attempt to infer the region from the number.
2615    * Returns false if it cannot use the region provided and the region cannot be inferred.
2616    */
checkRegionForParsing(String numberToParse, String defaultRegion)2617   private boolean checkRegionForParsing(String numberToParse, String defaultRegion) {
2618     if (!isValidRegionCode(defaultRegion)) {
2619       // If the number is null or empty, we can't infer the region.
2620       if (numberToParse == null || numberToParse.length() == 0 ||
2621           !PLUS_CHARS_PATTERN.matcher(numberToParse).lookingAt()) {
2622         return false;
2623       }
2624     }
2625     return true;
2626   }
2627 
2628   /**
2629    * Parses a string and returns it in proto buffer format. This method will throw a
2630    * {@link com.android.i18n.phonenumbers.NumberParseException} if the number is not considered to be
2631    * a possible number. Note that validation of whether the number is actually a valid number for a
2632    * particular region is not performed. This can be done separately with {@link #isValidNumber}.
2633    *
2634    * @param numberToParse     number that we are attempting to parse. This can contain formatting
2635    *                          such as +, ( and -, as well as a phone number extension.
2636    * @param defaultRegion     region that we are expecting the number to be from. This is only used
2637    *                          if the number being parsed is not written in international format.
2638    *                          The country_code for the number in this case would be stored as that
2639    *                          of the default region supplied. If the number is guaranteed to
2640    *                          start with a '+' followed by the country calling code, then
2641    *                          "ZZ" or null can be supplied.
2642    * @return                  a phone number proto buffer filled with the parsed number
2643    * @throws NumberParseException  if the string is not considered to be a viable phone number or if
2644    *                               no default region was supplied and the number is not in
2645    *                               international format (does not start with +)
2646    */
parse(String numberToParse, String defaultRegion)2647   public PhoneNumber parse(String numberToParse, String defaultRegion)
2648       throws NumberParseException {
2649     PhoneNumber phoneNumber = new PhoneNumber();
2650     parse(numberToParse, defaultRegion, phoneNumber);
2651     return phoneNumber;
2652   }
2653 
2654   /**
2655    * Same as {@link #parse(String, String)}, but accepts mutable PhoneNumber as a parameter to
2656    * decrease object creation when invoked many times.
2657    */
parse(String numberToParse, String defaultRegion, PhoneNumber phoneNumber)2658   public void parse(String numberToParse, String defaultRegion, PhoneNumber phoneNumber)
2659       throws NumberParseException {
2660     parseHelper(numberToParse, defaultRegion, false, true, phoneNumber);
2661   }
2662 
2663   /**
2664    * Parses a string and returns it in proto buffer format. This method differs from {@link #parse}
2665    * in that it always populates the raw_input field of the protocol buffer with numberToParse as
2666    * well as the country_code_source field.
2667    *
2668    * @param numberToParse     number that we are attempting to parse. This can contain formatting
2669    *                          such as +, ( and -, as well as a phone number extension.
2670    * @param defaultRegion     region that we are expecting the number to be from. This is only used
2671    *                          if the number being parsed is not written in international format.
2672    *                          The country calling code for the number in this case would be stored
2673    *                          as that of the default region supplied.
2674    * @return                  a phone number proto buffer filled with the parsed number
2675    * @throws NumberParseException  if the string is not considered to be a viable phone number or if
2676    *                               no default region was supplied
2677    */
parseAndKeepRawInput(String numberToParse, String defaultRegion)2678   public PhoneNumber parseAndKeepRawInput(String numberToParse, String defaultRegion)
2679       throws NumberParseException {
2680     PhoneNumber phoneNumber = new PhoneNumber();
2681     parseAndKeepRawInput(numberToParse, defaultRegion, phoneNumber);
2682     return phoneNumber;
2683   }
2684 
2685   /**
2686    * Same as{@link #parseAndKeepRawInput(String, String)}, but accepts a mutable PhoneNumber as
2687    * a parameter to decrease object creation when invoked many times.
2688    */
parseAndKeepRawInput(String numberToParse, String defaultRegion, PhoneNumber phoneNumber)2689   public void parseAndKeepRawInput(String numberToParse, String defaultRegion,
2690                                    PhoneNumber phoneNumber)
2691       throws NumberParseException {
2692     parseHelper(numberToParse, defaultRegion, true, true, phoneNumber);
2693   }
2694 
2695   /**
2696    * Returns an iterable over all {@link PhoneNumberMatch PhoneNumberMatches} in {@code text}. This
2697    * is a shortcut for {@link #findNumbers(CharSequence, String, Leniency, long)
2698    * getMatcher(text, defaultRegion, Leniency.VALID, Long.MAX_VALUE)}.
2699    *
2700    * @param text              the text to search for phone numbers, null for no text
2701    * @param defaultRegion     region that we are expecting the number to be from. This is only used
2702    *                          if the number being parsed is not written in international format. The
2703    *                          country_code for the number in this case would be stored as that of
2704    *                          the default region supplied. May be null if only international
2705    *                          numbers are expected.
2706    */
findNumbers(CharSequence text, String defaultRegion)2707   public Iterable<PhoneNumberMatch> findNumbers(CharSequence text, String defaultRegion) {
2708     return findNumbers(text, defaultRegion, Leniency.VALID, Long.MAX_VALUE);
2709   }
2710 
2711   /**
2712    * Returns an iterable over all {@link PhoneNumberMatch PhoneNumberMatches} in {@code text}.
2713    *
2714    * @param text              the text to search for phone numbers, null for no text
2715    * @param defaultRegion     region that we are expecting the number to be from. This is only used
2716    *                          if the number being parsed is not written in international format. The
2717    *                          country_code for the number in this case would be stored as that of
2718    *                          the default region supplied. May be null if only international
2719    *                          numbers are expected.
2720    * @param leniency          the leniency to use when evaluating candidate phone numbers
2721    * @param maxTries          the maximum number of invalid numbers to try before giving up on the
2722    *                          text. This is to cover degenerate cases where the text has a lot of
2723    *                          false positives in it. Must be {@code >= 0}.
2724    */
findNumbers( final CharSequence text, final String defaultRegion, final Leniency leniency, final long maxTries)2725   public Iterable<PhoneNumberMatch> findNumbers(
2726       final CharSequence text, final String defaultRegion, final Leniency leniency,
2727       final long maxTries) {
2728 
2729     return new Iterable<PhoneNumberMatch>() {
2730       public Iterator<PhoneNumberMatch> iterator() {
2731         return new PhoneNumberMatcher(
2732             PhoneNumberUtil.this, text, defaultRegion, leniency, maxTries);
2733       }
2734     };
2735   }
2736 
2737   /**
2738    * Parses a string and fills up the phoneNumber. This method is the same as the public
2739    * parse() method, with the exception that it allows the default region to be null, for use by
2740    * isNumberMatch(). checkRegion should be set to false if it is permitted for the default region
2741    * to be null or unknown ("ZZ").
2742    */
2743   private void parseHelper(String numberToParse, String defaultRegion, boolean keepRawInput,
2744                            boolean checkRegion, PhoneNumber phoneNumber)
2745       throws NumberParseException {
2746     if (numberToParse == null) {
2747       throw new NumberParseException(NumberParseException.ErrorType.NOT_A_NUMBER,
2748                                      "The phone number supplied was null.");
2749     } else if (numberToParse.length() > MAX_INPUT_STRING_LENGTH) {
2750       throw new NumberParseException(NumberParseException.ErrorType.TOO_LONG,
2751                                      "The string supplied was too long to parse.");
2752     }
2753     // Extract a possible number from the string passed in (this strips leading characters that
2754     // could not be the start of a phone number.)
2755     String number = extractPossibleNumber(numberToParse);
2756     if (!isViablePhoneNumber(number)) {
2757       throw new NumberParseException(NumberParseException.ErrorType.NOT_A_NUMBER,
2758                                      "The string supplied did not seem to be a phone number.");
2759     }
2760 
2761     // Check the region supplied is valid, or that the extracted number starts with some sort of +
2762     // sign so the number's region can be determined.
2763     if (checkRegion && !checkRegionForParsing(number, defaultRegion)) {
2764       throw new NumberParseException(NumberParseException.ErrorType.INVALID_COUNTRY_CODE,
2765                                      "Missing or invalid default region.");
2766     }
2767 
2768     if (keepRawInput) {
2769       phoneNumber.setRawInput(numberToParse);
2770     }
2771     StringBuilder nationalNumber = new StringBuilder(number);
2772     // Attempt to parse extension first, since it doesn't require region-specific data and we want
2773     // to have the non-normalised number here.
2774     String extension = maybeStripExtension(nationalNumber);
2775     if (extension.length() > 0) {
2776       phoneNumber.setExtension(extension);
2777     }
2778 
2779     PhoneMetadata regionMetadata = getMetadataForRegion(defaultRegion);
2780     // Check to see if the number is given in international format so we know whether this number is
2781     // from the default region or not.
2782     StringBuilder normalizedNationalNumber = new StringBuilder();
2783     int countryCode = 0;
2784     try {
2785       // TODO: This method should really just take in the string buffer that has already
2786       // been created, and just remove the prefix, rather than taking in a string and then
2787       // outputting a string buffer.
2788       countryCode = maybeExtractCountryCode(nationalNumber.toString(), regionMetadata,
2789                                             normalizedNationalNumber, keepRawInput, phoneNumber);
2790     } catch (NumberParseException e) {
2791       Matcher matcher = PLUS_CHARS_PATTERN.matcher(nationalNumber.toString());
2792       if (e.getErrorType() == NumberParseException.ErrorType.INVALID_COUNTRY_CODE &&
2793           matcher.lookingAt()) {
2794         // Strip the plus-char, and try again.
2795         countryCode = maybeExtractCountryCode(nationalNumber.substring(matcher.end()),
2796                                               regionMetadata, normalizedNationalNumber,
2797                                               keepRawInput, phoneNumber);
2798         if (countryCode == 0) {
2799           throw new NumberParseException(NumberParseException.ErrorType.INVALID_COUNTRY_CODE,
2800                                          "Could not interpret numbers after plus-sign.");
2801         }
2802       } else {
2803         throw new NumberParseException(e.getErrorType(), e.getMessage());
2804       }
2805     }
2806     if (countryCode != 0) {
2807       String phoneNumberRegion = getRegionCodeForCountryCode(countryCode);
2808       if (!phoneNumberRegion.equals(defaultRegion)) {
2809         regionMetadata = getMetadataForRegionOrCallingCode(countryCode, phoneNumberRegion);
2810       }
2811     } else {
2812       // If no extracted country calling code, use the region supplied instead. The national number
2813       // is just the normalized version of the number we were given to parse.
2814       normalize(nationalNumber);
2815       normalizedNationalNumber.append(nationalNumber);
2816       if (defaultRegion != null) {
2817         countryCode = regionMetadata.getCountryCode();
2818         phoneNumber.setCountryCode(countryCode);
2819       } else if (keepRawInput) {
2820         phoneNumber.clearCountryCodeSource();
2821       }
2822     }
2823     if (normalizedNationalNumber.length() < MIN_LENGTH_FOR_NSN) {
2824       throw new NumberParseException(NumberParseException.ErrorType.TOO_SHORT_NSN,
2825                                      "The string supplied is too short to be a phone number.");
2826     }
2827     if (regionMetadata != null) {
2828       StringBuilder carrierCode = new StringBuilder();
2829       maybeStripNationalPrefixAndCarrierCode(normalizedNationalNumber, regionMetadata, carrierCode);
2830       if (keepRawInput) {
2831         phoneNumber.setPreferredDomesticCarrierCode(carrierCode.toString());
2832       }
2833     }
2834     int lengthOfNationalNumber = normalizedNationalNumber.length();
2835     if (lengthOfNationalNumber < MIN_LENGTH_FOR_NSN) {
2836       throw new NumberParseException(NumberParseException.ErrorType.TOO_SHORT_NSN,
2837                                      "The string supplied is too short to be a phone number.");
2838     }
2839     if (lengthOfNationalNumber > MAX_LENGTH_FOR_NSN) {
2840       throw new NumberParseException(NumberParseException.ErrorType.TOO_LONG,
2841                                      "The string supplied is too long to be a phone number.");
2842     }
2843     if (normalizedNationalNumber.charAt(0) == '0') {
2844       phoneNumber.setItalianLeadingZero(true);
2845     }
2846     phoneNumber.setNationalNumber(Long.parseLong(normalizedNationalNumber.toString()));
2847   }
2848 
2849   /**
2850    * Takes two phone numbers and compares them for equality.
2851    *
2852    * <p>Returns EXACT_MATCH if the country_code, NSN, presence of a leading zero for Italian numbers
2853    * and any extension present are the same.
2854    * Returns NSN_MATCH if either or both has no region specified, and the NSNs and extensions are
2855    * the same.
2856    * Returns SHORT_NSN_MATCH if either or both has no region specified, or the region specified is
2857    * the same, and one NSN could be a shorter version of the other number. This includes the case
2858    * where one has an extension specified, and the other does not.
2859    * Returns NO_MATCH otherwise.
2860    * For example, the numbers +1 345 657 1234 and 657 1234 are a SHORT_NSN_MATCH.
2861    * The numbers +1 345 657 1234 and 345 657 are a NO_MATCH.
2862    *
2863    * @param firstNumberIn  first number to compare
2864    * @param secondNumberIn  second number to compare
2865    *
2866    * @return  NO_MATCH, SHORT_NSN_MATCH, NSN_MATCH or EXACT_MATCH depending on the level of equality
2867    *     of the two numbers, described in the method definition.
2868    */
2869   public MatchType isNumberMatch(PhoneNumber firstNumberIn, PhoneNumber secondNumberIn) {
2870     // Make copies of the phone number so that the numbers passed in are not edited.
2871     PhoneNumber firstNumber = new PhoneNumber();
2872     firstNumber.mergeFrom(firstNumberIn);
2873     PhoneNumber secondNumber = new PhoneNumber();
2874     secondNumber.mergeFrom(secondNumberIn);
2875     // First clear raw_input, country_code_source and preferred_domestic_carrier_code fields and any
2876     // empty-string extensions so that we can use the proto-buffer equality method.
2877     firstNumber.clearRawInput();
2878     firstNumber.clearCountryCodeSource();
2879     firstNumber.clearPreferredDomesticCarrierCode();
2880     secondNumber.clearRawInput();
2881     secondNumber.clearCountryCodeSource();
2882     secondNumber.clearPreferredDomesticCarrierCode();
2883     if (firstNumber.hasExtension() &&
2884         firstNumber.getExtension().length() == 0) {
2885         firstNumber.clearExtension();
2886     }
2887     if (secondNumber.hasExtension() &&
2888         secondNumber.getExtension().length() == 0) {
2889         secondNumber.clearExtension();
2890     }
2891     // Early exit if both had extensions and these are different.
2892     if (firstNumber.hasExtension() && secondNumber.hasExtension() &&
2893         !firstNumber.getExtension().equals(secondNumber.getExtension())) {
2894       return MatchType.NO_MATCH;
2895     }
2896     int firstNumberCountryCode = firstNumber.getCountryCode();
2897     int secondNumberCountryCode = secondNumber.getCountryCode();
2898     // Both had country_code specified.
2899     if (firstNumberCountryCode != 0 && secondNumberCountryCode != 0) {
2900       if (firstNumber.exactlySameAs(secondNumber)) {
2901         return MatchType.EXACT_MATCH;
2902       } else if (firstNumberCountryCode == secondNumberCountryCode &&
2903                  isNationalNumberSuffixOfTheOther(firstNumber, secondNumber)) {
2904         // A SHORT_NSN_MATCH occurs if there is a difference because of the presence or absence of
2905         // an 'Italian leading zero', the presence or absence of an extension, or one NSN being a
2906         // shorter variant of the other.
2907         return MatchType.SHORT_NSN_MATCH;
2908       }
2909       // This is not a match.
2910       return MatchType.NO_MATCH;
2911     }
2912     // Checks cases where one or both country_code fields were not specified. To make equality
2913     // checks easier, we first set the country_code fields to be equal.
2914     firstNumber.setCountryCode(secondNumberCountryCode);
2915     // If all else was the same, then this is an NSN_MATCH.
2916     if (firstNumber.exactlySameAs(secondNumber)) {
2917       return MatchType.NSN_MATCH;
2918     }
2919     if (isNationalNumberSuffixOfTheOther(firstNumber, secondNumber)) {
2920       return MatchType.SHORT_NSN_MATCH;
2921     }
2922     return MatchType.NO_MATCH;
2923   }
2924 
2925   // Returns true when one national number is the suffix of the other or both are the same.
2926   private boolean isNationalNumberSuffixOfTheOther(PhoneNumber firstNumber,
2927                                                    PhoneNumber secondNumber) {
2928     String firstNumberNationalNumber = String.valueOf(firstNumber.getNationalNumber());
2929     String secondNumberNationalNumber = String.valueOf(secondNumber.getNationalNumber());
2930     // Note that endsWith returns true if the numbers are equal.
2931     return firstNumberNationalNumber.endsWith(secondNumberNationalNumber) ||
2932            secondNumberNationalNumber.endsWith(firstNumberNationalNumber);
2933   }
2934 
2935   /**
2936    * Takes two phone numbers as strings and compares them for equality. This is a convenience
2937    * wrapper for {@link #isNumberMatch(PhoneNumber, PhoneNumber)}. No default region is known.
2938    *
2939    * @param firstNumber  first number to compare. Can contain formatting, and can have country
2940    *     calling code specified with + at the start.
2941    * @param secondNumber  second number to compare. Can contain formatting, and can have country
2942    *     calling code specified with + at the start.
2943    * @return  NOT_A_NUMBER, NO_MATCH, SHORT_NSN_MATCH, NSN_MATCH, EXACT_MATCH. See
2944    *     {@link #isNumberMatch(PhoneNumber, PhoneNumber)} for more details.
2945    */
2946   public MatchType isNumberMatch(String firstNumber, String secondNumber) {
2947     try {
2948       PhoneNumber firstNumberAsProto = parse(firstNumber, UNKNOWN_REGION);
2949       return isNumberMatch(firstNumberAsProto, secondNumber);
2950     } catch (NumberParseException e) {
2951       if (e.getErrorType() == NumberParseException.ErrorType.INVALID_COUNTRY_CODE) {
2952         try {
2953           PhoneNumber secondNumberAsProto = parse(secondNumber, UNKNOWN_REGION);
2954           return isNumberMatch(secondNumberAsProto, firstNumber);
2955         } catch (NumberParseException e2) {
2956           if (e2.getErrorType() == NumberParseException.ErrorType.INVALID_COUNTRY_CODE) {
2957             try {
2958               PhoneNumber firstNumberProto = new PhoneNumber();
2959               PhoneNumber secondNumberProto = new PhoneNumber();
2960               parseHelper(firstNumber, null, false, false, firstNumberProto);
2961               parseHelper(secondNumber, null, false, false, secondNumberProto);
2962               return isNumberMatch(firstNumberProto, secondNumberProto);
2963             } catch (NumberParseException e3) {
2964               // Fall through and return MatchType.NOT_A_NUMBER.
2965             }
2966           }
2967         }
2968       }
2969     }
2970     // One or more of the phone numbers we are trying to match is not a viable phone number.
2971     return MatchType.NOT_A_NUMBER;
2972   }
2973 
2974   /**
2975    * Takes two phone numbers and compares them for equality. This is a convenience wrapper for
2976    * {@link #isNumberMatch(PhoneNumber, PhoneNumber)}. No default region is known.
2977    *
2978    * @param firstNumber  first number to compare in proto buffer format.
2979    * @param secondNumber  second number to compare. Can contain formatting, and can have country
2980    *     calling code specified with + at the start.
2981    * @return  NOT_A_NUMBER, NO_MATCH, SHORT_NSN_MATCH, NSN_MATCH, EXACT_MATCH. See
2982    *     {@link #isNumberMatch(PhoneNumber, PhoneNumber)} for more details.
2983    */
2984   public MatchType isNumberMatch(PhoneNumber firstNumber, String secondNumber) {
2985     // First see if the second number has an implicit country calling code, by attempting to parse
2986     // it.
2987     try {
2988       PhoneNumber secondNumberAsProto = parse(secondNumber, UNKNOWN_REGION);
2989       return isNumberMatch(firstNumber, secondNumberAsProto);
2990     } catch (NumberParseException e) {
2991       if (e.getErrorType() == NumberParseException.ErrorType.INVALID_COUNTRY_CODE) {
2992         // The second number has no country calling code. EXACT_MATCH is no longer possible.
2993         // We parse it as if the region was the same as that for the first number, and if
2994         // EXACT_MATCH is returned, we replace this with NSN_MATCH.
2995         String firstNumberRegion = getRegionCodeForCountryCode(firstNumber.getCountryCode());
2996         try {
2997           if (!firstNumberRegion.equals(UNKNOWN_REGION)) {
2998             PhoneNumber secondNumberWithFirstNumberRegion = parse(secondNumber, firstNumberRegion);
2999             MatchType match = isNumberMatch(firstNumber, secondNumberWithFirstNumberRegion);
3000             if (match == MatchType.EXACT_MATCH) {
3001               return MatchType.NSN_MATCH;
3002             }
3003             return match;
3004           } else {
3005             // If the first number didn't have a valid country calling code, then we parse the
3006             // second number without one as well.
3007             PhoneNumber secondNumberProto = new PhoneNumber();
3008             parseHelper(secondNumber, null, false, false, secondNumberProto);
3009             return isNumberMatch(firstNumber, secondNumberProto);
3010           }
3011         } catch (NumberParseException e2) {
3012           // Fall-through to return NOT_A_NUMBER.
3013         }
3014       }
3015     }
3016     // One or more of the phone numbers we are trying to match is not a viable phone number.
3017     return MatchType.NOT_A_NUMBER;
3018   }
3019 
3020   /**
3021    * Returns true if the number can be dialled from outside the region, or unknown. If the number
3022    * can only be dialled from within the region, returns false. Does not check the number is a valid
3023    * number.
3024    * TODO: Make this method public when we have enough metadata to make it worthwhile.
3025    *
3026    * @param number  the phone-number for which we want to know whether it is diallable from
3027    *     outside the region
3028    */
3029   // @VisibleForTesting
3030   boolean canBeInternationallyDialled(PhoneNumber number) {
3031     String regionCode = getRegionCodeForNumber(number);
3032     if (!isValidRegionCode(regionCode)) {
3033       // Note numbers belonging to non-geographical entities (e.g. +800 numbers) are always
3034       // internationally diallable, and will be caught here.
3035       return true;
3036     }
3037     PhoneMetadata metadata = getMetadataForRegion(regionCode);
3038     String nationalSignificantNumber = getNationalSignificantNumber(number);
3039     return !isNumberMatchingDesc(nationalSignificantNumber, metadata.getNoInternationalDialling());
3040   }
3041 }
3042