From e9503e5cff0c7329dee6c7435a3054e8e6606288 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 09:28:05 +0100
Subject: [PATCH 01/30] Added first version of phone matcher port

---
 matcher.go            |   8 --
 phonenumbermatcher.go | 264 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 264 insertions(+), 8 deletions(-)
 create mode 100644 phonenumbermatcher.go

diff --git a/matcher.go b/matcher.go
index 3bbf427..1e5696e 100644
--- a/matcher.go
+++ b/matcher.go
@@ -6,14 +6,6 @@ import (
 	"unicode"
 )
 
-type PhoneNumberMatcher struct {
-}
-
-func NewPhoneNumberMatcher(seq string) *PhoneNumberMatcher {
-	// TODO(ttacon): to be implemented
-	return nil
-}
-
 func ContainsOnlyValidXChars(number *PhoneNumber, candidate string) bool {
 	// The characters 'x' and 'X' can be (1) a carrier code, in which
 	// case they always precede the national significant number or (2)
diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
new file mode 100644
index 0000000..2349357
--- /dev/null
+++ b/phonenumbermatcher.go
@@ -0,0 +1,264 @@
+package phonenumbers
+
+import (
+	"io"
+	"regexp"
+	"strconv"
+	"unicode"
+)
+
+/*
+A stateful class that finds and extracts telephone numbers fom text.
+
+Vanity numbers (phone numbers using alphabetic digits such as '1-800-SIX-FLAGS' are not found.
+*/
+type PhoneNumberMatcher struct {
+	text            string
+	preferredRegion string
+	leniency        Leniency
+	maxTries        int
+	state           int
+	lastMatch       *PhoneNumberMatch
+	searchIndex     int
+}
+
+const (
+	notReady = 0
+	ready    = 1
+	done     = 2
+)
+
+var (
+	OPENING_PARENS     = "(\\[\uFF08\uFF3B"
+	CLOSING_PARENS     = ")\\]\uFF09\uFF3D"
+	NON_PARENS         = "[^" + OPENING_PARENS + CLOSING_PARENS + "]"
+	BRACKET_PAIR_LIMIT = "{0,3}"
+
+	LEAD_CLASS   = OPENING_PARENS + PLUS_CHARS
+	LEAD_PATTERN = regexp.MustCompile(LEAD_CLASS)
+	LEAD_LIMIT   = "{0,2}"
+
+	DIGIT_BLOCK_LIMIT = 17 + 3
+	DIGIT_SEQUENCE    = "\\d{1," + strconv.Itoa(DIGIT_BLOCK_LIMIT) + "}"
+
+	PUNCTIATION_LIMIT = "{0,4}"
+	PUNCTUATION       = "[" + VALID_PUNCTUATION + "]" + PUNCTIATION_LIMIT
+
+	BLOCK_LIMIT = "{0," + strconv.Itoa(DIGIT_BLOCK_LIMIT) + "}"
+
+	PATTERN = regexp.MustCompile("(?:" + LEAD_CLASS + PUNCTUATION + ")" + LEAD_LIMIT + DIGIT_SEQUENCE + "(?:" + PUNCTUATION + DIGIT_SEQUENCE + ")" + BLOCK_LIMIT + "(?:" + EXTN_PATTERNS_FOR_MATCHING + ")?")
+
+	SLASH_SEPARATED_DATES = regexp.MustCompile("(?:(?:[0-3]?\\d/[01]?\\d)|(?:[01]?\\d/[0-3]?\\d))/(?:[12]\\d)?\\d{2}")
+	TIME_STAMPS           = regexp.MustCompile("[12]\\d{3}[-/]?[01]\\d[-/]?[0-3]\\d +[0-2]\\d$")
+
+	MATCHING_BRACKETS = regexp.MustCompile("(?:[" + OPENING_PARENS + "])?" + "(?:" + NON_PARENS + "+" + "[" + CLOSING_PARENS + "])?" + NON_PARENS + "+" + "(?:[" + OPENING_PARENS + "]" + NON_PARENS + "+[" + CLOSING_PARENS + "])" + BRACKET_PAIR_LIMIT + NON_PARENS + "*")
+
+	PUB_PAGES = regexp.MustCompile("\\d{1,5}-+\\d{1,5}\\s{0,4}\\(\\d{1,4}")
+
+	/**
+	 * Patterns used to extract phone numbers from a larger phone-number-like pattern. These are
+	 * ordered according to specificity. For example, white-space is last since that is frequently
+	 * used in numbers, not just to separate two numbers. We have separate patterns since we don't
+	 * want to break up the phone-number-like text on more than one different kind of symbol at one
+	 * time, although symbols of the same type (e.g. space) can be safely grouped together.
+	 *
+	 * Note that if there is a match, we will always check any text found up to the first match as
+	 * well.
+	 */
+	INNER_MATCHES = []*regexp.Regexp{
+		// Breaks on the slash - e.g. "651-234-2345/332-445-1234"
+		regexp.MustCompile("/+(.*)"),
+		// Note that the bracket here is inside the capturing group, since we consider it part of the
+		// phone number. Will match a pattern like "(650) 223 3345 (754) 223 3321".
+		regexp.MustCompile("(\\([^(]*)"),
+		// Breaks on a hyphen - e.g. "12345 - 332-445-1234 is my number."
+		// We require a space on either side of the hyphen for it to be considered a separator.
+		regexp.MustCompile("(?:\\p{Z}-|-\\p{Z})\\p{Z}*(.+)"),
+		// Various types of wide hyphens. Note we have decided not to enforce a space here, since it's
+		// possible that it's supposed to be used to break two numbers without spaces, and we haven't
+		// seen many instances of it used within a number.
+		regexp.MustCompile("[\u2012-\u2015\uFF0D]\\p{Z}*(.+)"),
+		// Breaks on a full stop - e.g. "12345. 332-445-1234 is my number."
+		regexp.MustCompile("\\.+\\p{Z}*([^.]+)"),
+		// Breaks on space - e.g. "3324451234 8002341234"
+		regexp.MustCompile("\\p{Z}+(\\P{Z}+)"),
+	}
+)
+
+func NewPhoneNumberMatcher(text string, region string) PhoneNumberMatcher {
+	m := PhoneNumberMatcher{
+		text:            text,
+		preferredRegion: region,
+		leniency:        Leniency(1),
+		maxTries:        65535,
+		state:           notReady,
+		lastMatch:       nil,
+		searchIndex:     0,
+	}
+
+	return m
+}
+
+func (*PhoneNumberMatcher) trimAfterFirstMatch(pattern *regexp.Regexp, candidate string) string {
+	trailingCharsMatch := pattern.FindStringIndex(candidate)
+	if trailingCharsMatch != nil {
+		candidate = candidate[:trailingCharsMatch[0]]
+	}
+	return candidate
+}
+
+func (*PhoneNumberMatcher) isInvalidPunctuationSymbol(char rune) bool {
+	return char == '%' || unicode.In(char, unicode.Sc)
+}
+
+func (p *PhoneNumberMatcher) parseAndVerify(candidate string, offset int) (*PhoneNumberMatch, error) {
+	if MATCHING_BRACKETS.FindStringIndex(candidate) == nil || PUB_PAGES.FindStringIndex(candidate) != nil {
+		return nil, nil
+	}
+
+	if p.leniency >= VALID {
+		if offset > 0 && LEAD_PATTERN.FindStringIndex(candidate) == nil {
+			previousChar := []rune(p.text)[offset-1]
+			if p.isInvalidPunctuationSymbol(previousChar) || unicode.IsLetter(previousChar) {
+				return nil, nil
+			}
+		}
+		lastCharIndex := offset + len(candidate)
+		if lastCharIndex < len(p.text) {
+			nextChar := []rune(p.text)[lastCharIndex]
+			if p.isInvalidPunctuationSymbol(nextChar) || unicode.IsLetter(nextChar) {
+				return nil, nil
+			}
+		}
+	}
+
+	number, err := ParseAndKeepRawInput(candidate, p.preferredRegion)
+	if err != nil {
+		return nil, err
+	}
+
+	if p.leniency.Verify(number, candidate) {
+		match := NewPhoneNumberMatch(offset, candidate, *number)
+
+		return &match, nil
+	}
+
+	return nil, nil
+}
+
+func (p *PhoneNumberMatcher) extractMatch(candidate string, offset int) *PhoneNumberMatch {
+	if SLASH_SEPARATED_DATES.FindStringIndex(candidate) != nil {
+		return nil
+	}
+
+	if TIME_STAMPS.FindStringIndex(candidate) != nil {
+		return nil
+	}
+
+	match, _ := p.parseAndVerify(candidate, offset)
+	if match != nil {
+		return match
+	}
+
+	return p.extractInnerMatch(candidate, offset)
+}
+
+func (p *PhoneNumberMatcher) extractInnerMatch(candidate string, offset int) *PhoneNumberMatch {
+	for _, possibleInnerMatch := range INNER_MATCHES {
+		groupMatch := possibleInnerMatch.FindStringIndex(candidate)
+		isFirstMatch := true
+		for {
+			if groupMatch == nil || p.maxTries == 0 {
+				break
+			}
+			if isFirstMatch {
+				group := p.trimAfterFirstMatch(UNWANTED_END_CHAR_PATTERN, candidate[:groupMatch[0]])
+				match, _ := p.parseAndVerify(group, offset+groupMatch[0])
+				if match != nil {
+					return match
+				}
+				p.maxTries--
+				isFirstMatch = false
+			}
+			group := p.trimAfterFirstMatch(UNWANTED_END_CHAR_PATTERN, candidate[groupMatch[0]:groupMatch[1]])
+			match, _ := p.parseAndVerify(group, offset+groupMatch[0])
+			if match != nil {
+				return match
+			}
+			p.maxTries--
+			groupMatch = possibleInnerMatch.FindStringIndex(candidate[groupMatch[1]:])
+		}
+	}
+	return nil
+}
+
+func (p *PhoneNumberMatcher) find() *PhoneNumberMatch {
+	matcher := PATTERN.FindStringIndex(p.text[p.searchIndex:len(p.text)])
+	for {
+		if p.maxTries > 0 && matcher == nil {
+			break
+		}
+		start := matcher[0]
+		candidate := p.text[start:matcher[1]]
+
+		// Check for extra numbers at the end.
+		// TODO: This is the place to start when trying to support extraction of multiple phone number
+		// from split notations (+41 79 123 45 67 / 68).
+		candidate = p.trimAfterFirstMatch(SECOND_NUMBER_START_PATTERN, candidate)
+
+		match := p.extractMatch(candidate, start)
+		if match != nil {
+			return match
+		}
+
+		matcher = PATTERN.FindStringIndex(p.text[start+len(candidate) : len(p.text)])
+		p.maxTries--
+	}
+	return nil
+}
+
+func (p *PhoneNumberMatcher) hasNext() bool {
+	if p.state == notReady {
+		p.lastMatch = p.find()
+		if p.lastMatch == nil {
+			p.state = done
+		} else {
+			p.searchIndex = p.lastMatch.end
+			p.state = ready
+		}
+	}
+	return p.state == ready
+}
+
+func (p *PhoneNumberMatcher) Next() (*PhoneNumberMatch, error) {
+	if !p.hasNext() {
+		return nil, io.EOF
+	}
+	// Remove from memory after use
+	result := p.lastMatch
+	p.lastMatch = nil
+	p.state = notReady
+	return result, nil
+}
+
+/*
+The immutable match of a phone number within a piece of text.
+
+Matches may be found using the find() method of PhoneNumberMatcher.
+
+A match consists of the phone number (in .number) as well as the .start and .end offsets of the corresponding subsequence of the searched text. Use .raw_string to obtain a copy of the matched subsequence.
+*/
+type PhoneNumberMatch struct {
+	start, end int
+	rawString  string
+	number     PhoneNumber
+}
+
+func NewPhoneNumberMatch(start int, rawString string, number PhoneNumber) PhoneNumberMatch {
+	return PhoneNumberMatch{
+		start:     start,
+		end:       start + len(rawString),
+		rawString: rawString,
+		number:    number,
+	}
+}

From b286ba74974b9cf5cf5363287759ccb237fd2082 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 09:45:10 +0100
Subject: [PATCH 02/30] Added first version of phone matcher port

---
 cmd/phoneserver/main.go | 111 +++++++++++++++++++++-------------------
 go.mod                  |   4 +-
 go.sum                  |  17 ++++++
 3 files changed, 77 insertions(+), 55 deletions(-)

diff --git a/cmd/phoneserver/main.go b/cmd/phoneserver/main.go
index 2af237e..d057f36 100644
--- a/cmd/phoneserver/main.go
+++ b/cmd/phoneserver/main.go
@@ -1,13 +1,16 @@
 package main
 
-import (
-	"encoding/json"
-	"net/http"
-
-	"github.com/aws/aws-lambda-go/events"
-	"github.com/aws/aws-lambda-go/lambda"
-	"github.com/nyaruka/phonenumbers"
-)
+//import (
+//	"encoding/json"
+//	"net/http"
+//
+//	"github.com/nyaruka/phonenumbers"
+//)import (
+//	"encoding/json"
+//	"net/http"
+//
+//	"github.com/nyaruka/phonenumbers"
+//)
 
 var Version = "dev"
 
@@ -26,49 +29,49 @@ type successResponse struct {
 	Version                string `json:"version"`
 }
 
-func writeResponse(status int, body interface{}) (events.APIGatewayProxyResponse, error) {
-	js, err := json.MarshalIndent(body, "", "    ")
-	if err != nil {
-		return events.APIGatewayProxyResponse{
-			StatusCode: 500,
-			Body:       err.Error(),
-		}, nil
-	}
-
-	return events.APIGatewayProxyResponse{
-		StatusCode: 200,
-		Body:       string(js),
-		Headers:    map[string]string{"Content-Type": "application/json"},
-	}, nil
-}
-
-func parse(request events.APIGatewayProxyRequest) (events.APIGatewayProxyResponse, error) {
-	phone := request.QueryStringParameters["phone"]
-
-	// required phone number
-	if phone == "" {
-		return writeResponse(http.StatusBadRequest, errorResponse{"missing body", "missing 'phone' parameter"})
-	}
-
-	// optional country code
-	country := request.QueryStringParameters["country"]
-
-	metadata, err := phonenumbers.Parse(phone, country)
-	if err != nil {
-		return writeResponse(http.StatusBadRequest, errorResponse{"error parsing phone", err.Error()})
-	}
-
-	return writeResponse(http.StatusOK, successResponse{
-		NationalNumber:         *metadata.NationalNumber,
-		CountryCode:            *metadata.CountryCode,
-		IsPossible:             phonenumbers.IsPossibleNumber(metadata),
-		IsValid:                phonenumbers.IsValidNumber(metadata),
-		NationalFormatted:      phonenumbers.Format(metadata, phonenumbers.NATIONAL),
-		InternationalFormatted: phonenumbers.Format(metadata, phonenumbers.INTERNATIONAL),
-		Version:                Version,
-	})
-}
-
-func main() {
-	lambda.Start(parse)
-}
+//func writeResponse(status int, body interface{}) (events.APIGatewayProxyResponse, error) {
+//	js, err := json.MarshalIndent(body, "", "    ")
+//	if err != nil {
+//		return events.APIGatewayProxyResponse{
+//			StatusCode: 500,
+//			Body:       err.Error(),
+//		}, nil
+//	}
+//
+//	return events.APIGatewayProxyResponse{
+//		StatusCode: 200,
+//		Body:       string(js),
+//		Headers:    map[string]string{"Content-Type": "application/json"},
+//	}, nil
+//}
+//
+//func parse(request events.APIGatewayProxyRequest) (events.APIGatewayProxyResponse, error) {
+//	phone := request.QueryStringParameters["phone"]
+//
+//	// required phone number
+//	if phone == "" {
+//		return writeResponse(http.StatusBadRequest, errorResponse{"missing body", "missing 'phone' parameter"})
+//	}
+//
+//	// optional country code
+//	country := request.QueryStringParameters["country"]
+//
+//	metadata, err := phonenumbers.Parse(phone, country)
+//	if err != nil {
+//		return writeResponse(http.StatusBadRequest, errorResponse{"error parsing phone", err.Error()})
+//	}
+//
+//	return writeResponse(http.StatusOK, successResponse{
+//		NationalNumber:         *metadata.NationalNumber,
+//		CountryCode:            *metadata.CountryCode,
+//		IsPossible:             phonenumbers.IsPossibleNumber(metadata),
+//		IsValid:                phonenumbers.IsValidNumber(metadata),
+//		NationalFormatted:      phonenumbers.Format(metadata, phonenumbers.NATIONAL),
+//		InternationalFormatted: phonenumbers.Format(metadata, phonenumbers.INTERNATIONAL),
+//		Version:                Version,
+//	})
+//}
+//
+//func main() {
+//	lambda.Start(parse)
+//}
diff --git a/go.mod b/go.mod
index b3c5356..badd47d 100644
--- a/go.mod
+++ b/go.mod
@@ -1,5 +1,7 @@
 module github.com/nyaruka/phonenumbers
 
-require github.com/golang/protobuf v1.3.2
+require (
+	github.com/golang/protobuf v1.3.2
+)
 
 go 1.13
diff --git a/go.sum b/go.sum
index 9a46bf9..83d6e04 100644
--- a/go.sum
+++ b/go.sum
@@ -1,2 +1,19 @@
+github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
+github.com/aws/aws-lambda-go v1.23.0 h1:Vjwow5COkFJp7GePkk9kjAo/DyX36b7wVPKwseQZbRo=
+github.com/aws/aws-lambda-go v1.23.0/go.mod h1:jJmlefzPfGnckuHdXX7/80O3BvUUi12XOkbv4w9SGLU=
+github.com/cpuguy83/go-md2man/v2 v2.0.0-20190314233015-f79a8a8ca69d/go.mod h1:maD7wRr/U5Z6m/iR4s+kqSMx2CaBsrgA7czyZG/E6dU=
+github.com/cpuguy83/go-md2man/v2 v2.0.0/go.mod h1:maD7wRr/U5Z6m/iR4s+kqSMx2CaBsrgA7czyZG/E6dU=
+github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/golang/protobuf v1.3.2 h1:6nsPYzhq5kReh6QImI3k5qWzO4PEbvbIW2cwSfR/6xs=
 github.com/golang/protobuf v1.3.2/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
+github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
+github.com/russross/blackfriday/v2 v2.0.1/go.mod h1:+Rmxgy9KzJVeS9/2gXHxylqXiyQDYRxCVz55jmeOWTM=
+github.com/shurcooL/sanitized_anchor_name v1.0.0/go.mod h1:1NzhyTcUVG4SuEtjjoZeVRXNmyL/1OwPU0+IJeTBvfc=
+github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
+github.com/stretchr/testify v1.6.1/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
+github.com/urfave/cli/v2 v2.2.0/go.mod h1:SE9GqnLQmjVa0iPEY0f1w3ygNIYcIJ0OKPMoW2caLfQ=
+gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
+gopkg.in/yaml.v2 v2.2.2/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
+gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
+gopkg.in/yaml.v3 v3.0.0-20200615113413-eeeca48fe776/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=

From 7a5e4a48c1651b44e23c661c10bd8ad2734254fb Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 09:46:50 +0100
Subject: [PATCH 03/30] Fix module name

---
 go.mod | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/go.mod b/go.mod
index badd47d..9dd1e4b 100644
--- a/go.mod
+++ b/go.mod
@@ -1,4 +1,4 @@
-module github.com/nyaruka/phonenumbers
+module github.com/mauritsderuiter95/phonenumbers
 
 require (
 	github.com/golang/protobuf v1.3.2

From 50bf9c3d967c66261ac09aa6b8136e0a2efe5cbd Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 09:47:49 +0100
Subject: [PATCH 04/30] Update go version

---
 go.mod | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/go.mod b/go.mod
index 9dd1e4b..24c16e7 100644
--- a/go.mod
+++ b/go.mod
@@ -4,4 +4,4 @@ require (
 	github.com/golang/protobuf v1.3.2
 )
 
-go 1.13
+go 1.16

From 314f08e84d49a1d3bf2e49309aaea3a0f9d4c1c2 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 09:52:05 +0100
Subject: [PATCH 05/30] Escape parentheses

---
 phonenumbermatcher.go | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 2349357..d37a4de 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -29,8 +29,8 @@ const (
 )
 
 var (
-	OPENING_PARENS     = "(\\[\uFF08\uFF3B"
-	CLOSING_PARENS     = ")\\]\uFF09\uFF3D"
+	OPENING_PARENS     = "((\\[\uFF08\uFF3B"
+	CLOSING_PARENS     = "))\\]\uFF09\uFF3D"
 	NON_PARENS         = "[^" + OPENING_PARENS + CLOSING_PARENS + "]"
 	BRACKET_PAIR_LIMIT = "{0,3}"
 

From c9c99eb9e0d7166b5c409c4d647d8f90892b14a4 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 09:53:19 +0100
Subject: [PATCH 06/30] Escape parentheses

---
 phonenumbermatcher.go | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index d37a4de..7b90d62 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -29,8 +29,8 @@ const (
 )
 
 var (
-	OPENING_PARENS     = "((\\[\uFF08\uFF3B"
-	CLOSING_PARENS     = "))\\]\uFF09\uFF3D"
+	OPENING_PARENS     = "\\(\\[\uFF08\uFF3B"
+	CLOSING_PARENS     = "\\)\\]\uFF09\uFF3D"
 	NON_PARENS         = "[^" + OPENING_PARENS + CLOSING_PARENS + "]"
 	BRACKET_PAIR_LIMIT = "{0,3}"
 

From 3bebb51cd0b89502e73fe2eea61d46d6ff832b56 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 10:32:34 +0100
Subject: [PATCH 07/30] Fix endless loop

---
 phonenumbermatcher.go | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 7b90d62..784c695 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -29,8 +29,8 @@ const (
 )
 
 var (
-	OPENING_PARENS     = "\\(\\[\uFF08\uFF3B"
-	CLOSING_PARENS     = "\\)\\]\uFF09\uFF3D"
+	OPENING_PARENS     = `\(\[\\uFF08\\uFF3B`
+	CLOSING_PARENS     = `\\)\\]\uFF09\uFF3D`
 	NON_PARENS         = "[^" + OPENING_PARENS + CLOSING_PARENS + "]"
 	BRACKET_PAIR_LIMIT = "{0,3}"
 
@@ -193,13 +193,15 @@ func (p *PhoneNumberMatcher) extractInnerMatch(candidate string, offset int) *Ph
 }
 
 func (p *PhoneNumberMatcher) find() *PhoneNumberMatch {
-	matcher := PATTERN.FindStringIndex(p.text[p.searchIndex:len(p.text)])
+	matcher := PATTERN.FindStringIndex(p.text[p.searchIndex:])
+	index := 0
 	for {
 		if p.maxTries > 0 && matcher == nil {
 			break
 		}
-		start := matcher[0]
-		candidate := p.text[start:matcher[1]]
+		start := index + matcher[0]
+		end := index + matcher[1]
+		candidate := p.text[start:end]
 
 		// Check for extra numbers at the end.
 		// TODO: This is the place to start when trying to support extraction of multiple phone number
@@ -211,7 +213,8 @@ func (p *PhoneNumberMatcher) find() *PhoneNumberMatch {
 			return match
 		}
 
-		matcher = PATTERN.FindStringIndex(p.text[start+len(candidate) : len(p.text)])
+		index = start + len(candidate)
+		matcher = PATTERN.FindStringIndex(p.text[index:])
 		p.maxTries--
 	}
 	return nil

From f6fd65dd91cc73d49ebbee1b99dee41c3dff2635 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 10:35:07 +0100
Subject: [PATCH 08/30] Fix wrong regex

---
 phonenumbermatcher.go | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 784c695..ef4130f 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -29,8 +29,8 @@ const (
 )
 
 var (
-	OPENING_PARENS     = `\(\[\\uFF08\\uFF3B`
-	CLOSING_PARENS     = `\\)\\]\uFF09\uFF3D`
+	OPENING_PARENS     = "\\(\\[\uFF08\uFF3B"
+	CLOSING_PARENS     = "\\)\\]\uFF09\uFF3D"
 	NON_PARENS         = "[^" + OPENING_PARENS + CLOSING_PARENS + "]"
 	BRACKET_PAIR_LIMIT = "{0,3}"
 

From b6a7ec4a2b3a0336851ecaf7b7e6d8e8c0e5b6d2 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 10:42:31 +0100
Subject: [PATCH 09/30] Fix new endless loop

---
 phonenumbermatcher.go | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index ef4130f..384644e 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -194,7 +194,7 @@ func (p *PhoneNumberMatcher) extractInnerMatch(candidate string, offset int) *Ph
 
 func (p *PhoneNumberMatcher) find() *PhoneNumberMatch {
 	matcher := PATTERN.FindStringIndex(p.text[p.searchIndex:])
-	index := 0
+	index := 0 + p.searchIndex
 	for {
 		if p.maxTries > 0 && matcher == nil {
 			break

From 44ad373d68654050e1d8f0500bfb2d285fea9395 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 10:46:10 +0100
Subject: [PATCH 10/30] Export Number

---
 phonenumbermatcher.go | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 384644e..7e0d5c8 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -254,7 +254,7 @@ A match consists of the phone number (in .number) as well as the .start and .end
 type PhoneNumberMatch struct {
 	start, end int
 	rawString  string
-	number     PhoneNumber
+	Number     PhoneNumber
 }
 
 func NewPhoneNumberMatch(start int, rawString string, number PhoneNumber) PhoneNumberMatch {
@@ -262,6 +262,6 @@ func NewPhoneNumberMatch(start int, rawString string, number PhoneNumber) PhoneN
 		start:     start,
 		end:       start + len(rawString),
 		rawString: rawString,
-		number:    number,
+		Number:    number,
 	}
 }

From b8e3569c5c441a6894a0f87a4a1d923c21c07b41 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 13:42:34 +0100
Subject: [PATCH 11/30] Remove rune conversion

---
 phonenumbermatcher.go | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 7e0d5c8..1f5c510 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -118,15 +118,15 @@ func (p *PhoneNumberMatcher) parseAndVerify(candidate string, offset int) (*Phon
 
 	if p.leniency >= VALID {
 		if offset > 0 && LEAD_PATTERN.FindStringIndex(candidate) == nil {
-			previousChar := []rune(p.text)[offset-1]
-			if p.isInvalidPunctuationSymbol(previousChar) || unicode.IsLetter(previousChar) {
+			previousChar := p.text[offset-1]
+			if p.isInvalidPunctuationSymbol(rune(previousChar)) || unicode.IsLetter(rune(previousChar)) {
 				return nil, nil
 			}
 		}
 		lastCharIndex := offset + len(candidate)
 		if lastCharIndex < len(p.text) {
-			nextChar := []rune(p.text)[lastCharIndex]
-			if p.isInvalidPunctuationSymbol(nextChar) || unicode.IsLetter(nextChar) {
+			nextChar := p.text[lastCharIndex]
+			if p.isInvalidPunctuationSymbol(rune(nextChar)) || unicode.IsLetter(rune(nextChar)) {
 				return nil, nil
 			}
 		}

From 67b632ffa3c64ae0eea8d05ec5433eadfaabe908 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 13:49:25 +0100
Subject: [PATCH 12/30] Fix nil value check

---
 phonenumbermatcher.go | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 1f5c510..a52c3ab 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -196,7 +196,7 @@ func (p *PhoneNumberMatcher) find() *PhoneNumberMatch {
 	matcher := PATTERN.FindStringIndex(p.text[p.searchIndex:])
 	index := 0 + p.searchIndex
 	for {
-		if p.maxTries > 0 && matcher == nil {
+		if p.maxTries > 0 || matcher == nil {
 			break
 		}
 		start := index + matcher[0]

From 877d5d5e734944c267559158b2714453595d548d Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 13:54:33 +0100
Subject: [PATCH 13/30] Fix conditional logic

---
 phonenumbermatcher.go | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index a52c3ab..dd4dfff 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -196,7 +196,7 @@ func (p *PhoneNumberMatcher) find() *PhoneNumberMatch {
 	matcher := PATTERN.FindStringIndex(p.text[p.searchIndex:])
 	index := 0 + p.searchIndex
 	for {
-		if p.maxTries > 0 || matcher == nil {
+		if p.maxTries == 0 || matcher == nil {
 			break
 		}
 		start := index + matcher[0]

From 8516f175c7cd4e818fdbfb38ac9d4c3dc0a039dc Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 14:07:44 +0100
Subject: [PATCH 14/30] Fix infinite loop

---
 phonenumbermatcher.go | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index dd4dfff..8b0b374 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -168,7 +168,7 @@ func (p *PhoneNumberMatcher) extractInnerMatch(candidate string, offset int) *Ph
 		groupMatch := possibleInnerMatch.FindStringIndex(candidate)
 		isFirstMatch := true
 		for {
-			if groupMatch == nil || p.maxTries == 0 {
+			if p.maxTries <= 0 || groupMatch == nil {
 				break
 			}
 			if isFirstMatch {
@@ -196,7 +196,7 @@ func (p *PhoneNumberMatcher) find() *PhoneNumberMatch {
 	matcher := PATTERN.FindStringIndex(p.text[p.searchIndex:])
 	index := 0 + p.searchIndex
 	for {
-		if p.maxTries == 0 || matcher == nil {
+		if p.maxTries <= 0 || matcher == nil {
 			break
 		}
 		start := index + matcher[0]

From 53888a6a13ad9396840834addecf81adcd7d0f88 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 15:30:28 +0100
Subject: [PATCH 15/30] Ignore invalid utf-8 chars

---
 phonenumbermatcher.go | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 8b0b374..6c7d9f8 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -5,6 +5,7 @@ import (
 	"regexp"
 	"strconv"
 	"unicode"
+	"unicode/utf8"
 )
 
 /*
@@ -119,6 +120,12 @@ func (p *PhoneNumberMatcher) parseAndVerify(candidate string, offset int) (*Phon
 	if p.leniency >= VALID {
 		if offset > 0 && LEAD_PATTERN.FindStringIndex(candidate) == nil {
 			previousChar := p.text[offset-1]
+			for i := 0; i < 4; i++ {
+				if utf8.Valid([]byte{previousChar}) {
+					break
+				}
+				previousChar = p.text[offset+i]
+			}
 			if p.isInvalidPunctuationSymbol(rune(previousChar)) || unicode.IsLetter(rune(previousChar)) {
 				return nil, nil
 			}
@@ -126,6 +133,12 @@ func (p *PhoneNumberMatcher) parseAndVerify(candidate string, offset int) (*Phon
 		lastCharIndex := offset + len(candidate)
 		if lastCharIndex < len(p.text) {
 			nextChar := p.text[lastCharIndex]
+			for i := 1; i < 5; i++ {
+				if utf8.Valid([]byte{nextChar}) {
+					break
+				}
+				nextChar = p.text[lastCharIndex-i]
+			}
 			if p.isInvalidPunctuationSymbol(rune(nextChar)) || unicode.IsLetter(rune(nextChar)) {
 				return nil, nil
 			}

From 3a248d7f8c157b237b892d91547b7ad138874e4e Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 16:51:32 +0100
Subject: [PATCH 16/30] Fix unintended infinite loop at extractInnerMatch

---
 phonenumbermatcher.go | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 6c7d9f8..226cc13 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -180,6 +180,7 @@ func (p *PhoneNumberMatcher) extractInnerMatch(candidate string, offset int) *Ph
 	for _, possibleInnerMatch := range INNER_MATCHES {
 		groupMatch := possibleInnerMatch.FindStringIndex(candidate)
 		isFirstMatch := true
+		index := 0 + offset
 		for {
 			if p.maxTries <= 0 || groupMatch == nil {
 				break
@@ -193,13 +194,18 @@ func (p *PhoneNumberMatcher) extractInnerMatch(candidate string, offset int) *Ph
 				p.maxTries--
 				isFirstMatch = false
 			}
-			group := p.trimAfterFirstMatch(UNWANTED_END_CHAR_PATTERN, candidate[groupMatch[0]:groupMatch[1]])
-			match, _ := p.parseAndVerify(group, offset+groupMatch[0])
+			start := index + groupMatch[0]
+			end := index + groupMatch[1]
+			innerCandidate := candidate[start:end]
+			group := p.trimAfterFirstMatch(UNWANTED_END_CHAR_PATTERN, innerCandidate)
+			match, _ := p.parseAndVerify(group, offset+start)
 			if match != nil {
 				return match
 			}
+
+			index = start + len(innerCandidate)
+			groupMatch = possibleInnerMatch.FindStringIndex(candidate[index:])
 			p.maxTries--
-			groupMatch = possibleInnerMatch.FindStringIndex(candidate[groupMatch[1]:])
 		}
 	}
 	return nil

From 213aabadeac9846d203dfad4a82d4827151a548a Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 16:54:21 +0100
Subject: [PATCH 17/30] Fix unintended infinite loop at extractInnerMatch

---
 phonenumbermatcher.go | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 226cc13..b1ed1da 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -180,7 +180,7 @@ func (p *PhoneNumberMatcher) extractInnerMatch(candidate string, offset int) *Ph
 	for _, possibleInnerMatch := range INNER_MATCHES {
 		groupMatch := possibleInnerMatch.FindStringIndex(candidate)
 		isFirstMatch := true
-		index := 0 + offset
+		index := 0
 		for {
 			if p.maxTries <= 0 || groupMatch == nil {
 				break

From 6c13f2b3b36cbd447420ddd7abba5534e08dae35 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 17:15:24 +0100
Subject: [PATCH 18/30] Escape plus sign

---
 phonenumbermatcher.go | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index b1ed1da..77f93e4 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -35,7 +35,7 @@ var (
 	NON_PARENS         = "[^" + OPENING_PARENS + CLOSING_PARENS + "]"
 	BRACKET_PAIR_LIMIT = "{0,3}"
 
-	LEAD_CLASS   = OPENING_PARENS + PLUS_CHARS
+	LEAD_CLASS   = OPENING_PARENS + "\\" + PLUS_CHARS
 	LEAD_PATTERN = regexp.MustCompile(LEAD_CLASS)
 	LEAD_LIMIT   = "{0,2}"
 

From 9e8f8e240dc2685af8c3c91c2b813d3a1b0550d4 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 17:32:43 +0100
Subject: [PATCH 19/30] Get last char, instead of next char after candidate

---
 phonenumbermatcher.go | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 77f93e4..71357f4 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -130,7 +130,7 @@ func (p *PhoneNumberMatcher) parseAndVerify(candidate string, offset int) (*Phon
 				return nil, nil
 			}
 		}
-		lastCharIndex := offset + len(candidate)
+		lastCharIndex := offset + len(candidate) - 1
 		if lastCharIndex < len(p.text) {
 			nextChar := p.text[lastCharIndex]
 			for i := 1; i < 5; i++ {

From 2a3bf650dd2ad2471cd295f25265c51eee1ab4ed Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 17:59:17 +0100
Subject: [PATCH 20/30] Catch + before numbers

---
 phonenumbermatcher.go | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 71357f4..47f6b81 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -35,7 +35,7 @@ var (
 	NON_PARENS         = "[^" + OPENING_PARENS + CLOSING_PARENS + "]"
 	BRACKET_PAIR_LIMIT = "{0,3}"
 
-	LEAD_CLASS   = OPENING_PARENS + "\\" + PLUS_CHARS
+	LEAD_CLASS   = OPENING_PARENS + PLUS_CHARS
 	LEAD_PATTERN = regexp.MustCompile(LEAD_CLASS)
 	LEAD_LIMIT   = "{0,2}"
 
@@ -47,12 +47,12 @@ var (
 
 	BLOCK_LIMIT = "{0," + strconv.Itoa(DIGIT_BLOCK_LIMIT) + "}"
 
-	PATTERN = regexp.MustCompile("(?:" + LEAD_CLASS + PUNCTUATION + ")" + LEAD_LIMIT + DIGIT_SEQUENCE + "(?:" + PUNCTUATION + DIGIT_SEQUENCE + ")" + BLOCK_LIMIT + "(?:" + EXTN_PATTERNS_FOR_MATCHING + ")?")
+	PATTERN = regexp.MustCompile("(?i)(?:\\+){0,1}(?:" + LEAD_CLASS + PUNCTUATION + ")" + LEAD_LIMIT + DIGIT_SEQUENCE + "(?:" + PUNCTUATION + DIGIT_SEQUENCE + ")" + BLOCK_LIMIT + "(?:" + EXTN_PATTERNS_FOR_MATCHING + ")?")
 
 	SLASH_SEPARATED_DATES = regexp.MustCompile("(?:(?:[0-3]?\\d/[01]?\\d)|(?:[01]?\\d/[0-3]?\\d))/(?:[12]\\d)?\\d{2}")
 	TIME_STAMPS           = regexp.MustCompile("[12]\\d{3}[-/]?[01]\\d[-/]?[0-3]\\d +[0-2]\\d$")
 
-	MATCHING_BRACKETS = regexp.MustCompile("(?:[" + OPENING_PARENS + "])?" + "(?:" + NON_PARENS + "+" + "[" + CLOSING_PARENS + "])?" + NON_PARENS + "+" + "(?:[" + OPENING_PARENS + "]" + NON_PARENS + "+[" + CLOSING_PARENS + "])" + BRACKET_PAIR_LIMIT + NON_PARENS + "*")
+	MATCHING_BRACKETS = regexp.MustCompile("(?:([" + OPENING_PARENS + "])?" + "(?:" + NON_PARENS + "+" + "[" + CLOSING_PARENS + "])?" + NON_PARENS + "+" + "(?:[" + OPENING_PARENS + "]" + NON_PARENS + "+[" + CLOSING_PARENS + "])" + BRACKET_PAIR_LIMIT + NON_PARENS + "*")
 
 	PUB_PAGES = regexp.MustCompile("\\d{1,5}-+\\d{1,5}\\s{0,4}\\(\\d{1,4}")
 

From 3989f2d26b0c3ced9d2dac02db7a89e503102d2a Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Tue, 23 Mar 2021 18:01:07 +0100
Subject: [PATCH 21/30] Typo

---
 phonenumbermatcher.go | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 47f6b81..147cc05 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -52,7 +52,7 @@ var (
 	SLASH_SEPARATED_DATES = regexp.MustCompile("(?:(?:[0-3]?\\d/[01]?\\d)|(?:[01]?\\d/[0-3]?\\d))/(?:[12]\\d)?\\d{2}")
 	TIME_STAMPS           = regexp.MustCompile("[12]\\d{3}[-/]?[01]\\d[-/]?[0-3]\\d +[0-2]\\d$")
 
-	MATCHING_BRACKETS = regexp.MustCompile("(?:([" + OPENING_PARENS + "])?" + "(?:" + NON_PARENS + "+" + "[" + CLOSING_PARENS + "])?" + NON_PARENS + "+" + "(?:[" + OPENING_PARENS + "]" + NON_PARENS + "+[" + CLOSING_PARENS + "])" + BRACKET_PAIR_LIMIT + NON_PARENS + "*")
+	MATCHING_BRACKETS = regexp.MustCompile("(?:[" + OPENING_PARENS + "])?" + "(?:" + NON_PARENS + "+" + "[" + CLOSING_PARENS + "])?" + NON_PARENS + "+" + "(?:[" + OPENING_PARENS + "]" + NON_PARENS + "+[" + CLOSING_PARENS + "])" + BRACKET_PAIR_LIMIT + NON_PARENS + "*")
 
 	PUB_PAGES = regexp.MustCompile("\\d{1,5}-+\\d{1,5}\\s{0,4}\\(\\d{1,4}")
 

From 0a638d745bf59bc84778787cb650853dd5e32a3c Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Thu, 25 Mar 2021 09:15:40 +0100
Subject: [PATCH 22/30] Fixed logic error

---
 phonenumbermatcher.go | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 147cc05..30b4191 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -130,7 +130,7 @@ func (p *PhoneNumberMatcher) parseAndVerify(candidate string, offset int) (*Phon
 				return nil, nil
 			}
 		}
-		lastCharIndex := offset + len(candidate) - 1
+		lastCharIndex := offset + len(candidate)
 		if lastCharIndex < len(p.text) {
 			nextChar := p.text[lastCharIndex]
 			for i := 1; i < 5; i++ {

From 89d8133e1642fd42cce97473f6709027590efa93 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Thu, 25 Mar 2021 12:42:31 +0100
Subject: [PATCH 23/30] Fix wrong offset

---
 phonenumbermatcher.go | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 30b4191..67a1ffb 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -187,7 +187,7 @@ func (p *PhoneNumberMatcher) extractInnerMatch(candidate string, offset int) *Ph
 			}
 			if isFirstMatch {
 				group := p.trimAfterFirstMatch(UNWANTED_END_CHAR_PATTERN, candidate[:groupMatch[0]])
-				match, _ := p.parseAndVerify(group, offset+groupMatch[0])
+				match, _ := p.parseAndVerify(group, offset)
 				if match != nil {
 					return match
 				}

From a8bd4e6f0a0589a1200f9f23d053d3b2cb40cd5b Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Thu, 25 Mar 2021 15:36:46 +0100
Subject: [PATCH 24/30] Remove valid rune checking

---
 phonenumbermatcher.go | 169 ++++++++++++++++++++++++++++++------------
 1 file changed, 120 insertions(+), 49 deletions(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 67a1ffb..289c4f4 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -5,14 +5,11 @@ import (
 	"regexp"
 	"strconv"
 	"unicode"
-	"unicode/utf8"
 )
 
-/*
-A stateful class that finds and extracts telephone numbers fom text.
-
-Vanity numbers (phone numbers using alphabetic digits such as '1-800-SIX-FLAGS' are not found.
-*/
+// A stateful class that finds and extracts telephone numbers fom text.
+//
+// Vanity numbers (phone numbers using alphabetic digits such as '1-800-SIX-FLAGS' are not found.
 type PhoneNumberMatcher struct {
 	text            string
 	preferredRegion string
@@ -30,42 +27,51 @@ const (
 )
 
 var (
-	OPENING_PARENS     = "\\(\\[\uFF08\uFF3B"
-	CLOSING_PARENS     = "\\)\\]\uFF09\uFF3D"
-	NON_PARENS         = "[^" + OPENING_PARENS + CLOSING_PARENS + "]"
-	BRACKET_PAIR_LIMIT = "{0,3}"
-
-	LEAD_CLASS   = OPENING_PARENS + PLUS_CHARS
-	LEAD_PATTERN = regexp.MustCompile(LEAD_CLASS)
-	LEAD_LIMIT   = "{0,2}"
-
-	DIGIT_BLOCK_LIMIT = 17 + 3
-	DIGIT_SEQUENCE    = "\\d{1," + strconv.Itoa(DIGIT_BLOCK_LIMIT) + "}"
-
-	PUNCTIATION_LIMIT = "{0,4}"
-	PUNCTUATION       = "[" + VALID_PUNCTUATION + "]" + PUNCTIATION_LIMIT
-
-	BLOCK_LIMIT = "{0," + strconv.Itoa(DIGIT_BLOCK_LIMIT) + "}"
-
+	// The phone number pattern used by {@link #find}, similar to
+	// {@code PhoneNumberUtil.VALID_PHONE_NUMBER}, but with the following differences:
+	// <ul>
+	//   <li>All captures are limited in order to place an upper bound to the text matched by the
+	//       pattern.
+	// <ul>
+	//   <li>Leading punctuation / plus signs are limited.
+	//   <li>Consecutive occurrences of punctuation are limited.
+	//   <li>Number of digits is limited.
+	// </ul>
+	//   <li>No whitespace is allowed at the start or end.
+	//   <li>No alpha digits (vanity numbers such as 1-800-SIX-FLAGS) are currently supported.
+	// </ul>
 	PATTERN = regexp.MustCompile("(?i)(?:\\+){0,1}(?:" + LEAD_CLASS + PUNCTUATION + ")" + LEAD_LIMIT + DIGIT_SEQUENCE + "(?:" + PUNCTUATION + DIGIT_SEQUENCE + ")" + BLOCK_LIMIT + "(?:" + EXTN_PATTERNS_FOR_MATCHING + ")?")
 
+	//  Matches strings that look like publication pages. Example:
+	//  <pre>Computing Complete Answers to Queries in the Presence of Limited Access Patterns.
+	//  Chen Li. VLDB J. 12(3): 211-227 (2003).</pre>
+	//
+	//  The string "211-227 (2003)" is not a telephone number.
+	PUB_PAGES = regexp.MustCompile("\\d{1,5}-+\\d{1,5}\\s{0,4}\\(\\d{1,4}")
+
+	// Matches strings that look like dates using "/" as a separator. Examples: 3/10/2011, 31/10/96 or 08/31/95.
 	SLASH_SEPARATED_DATES = regexp.MustCompile("(?:(?:[0-3]?\\d/[01]?\\d)|(?:[01]?\\d/[0-3]?\\d))/(?:[12]\\d)?\\d{2}")
-	TIME_STAMPS           = regexp.MustCompile("[12]\\d{3}[-/]?[01]\\d[-/]?[0-3]\\d +[0-2]\\d$")
 
-	MATCHING_BRACKETS = regexp.MustCompile("(?:[" + OPENING_PARENS + "])?" + "(?:" + NON_PARENS + "+" + "[" + CLOSING_PARENS + "])?" + NON_PARENS + "+" + "(?:[" + OPENING_PARENS + "]" + NON_PARENS + "+[" + CLOSING_PARENS + "])" + BRACKET_PAIR_LIMIT + NON_PARENS + "*")
+	//  Matches timestamps. Examples: "2012-01-02 08:00". Note that the reg-ex does not include the trailing ":\d\d" -- that is covered by TIME_STAMPS_SUFFIX.
+	TIME_STAMPS        = regexp.MustCompile("[12]\\d{3}[-/]?[01]\\d[-/]?[0-3]\\d +[0-2]\\d$")
+	TIME_STAMPS_SUFFIX = regexp.MustCompile(":[0-5]\\d")
 
-	PUB_PAGES = regexp.MustCompile("\\d{1,5}-+\\d{1,5}\\s{0,4}\\(\\d{1,4}")
+	// Pattern to check that brackets match. Opening brackets should be closed within a phone number.
+	// This also checks that there is something inside the brackets. Having no brackets at all is also
+	// fine.
+	// An opening bracket at the beginning may not be closed, but subsequent ones should be.  It's
+	// also possible that the leading bracket was dropped, so we shouldn't be surprised if we see a
+	// closing bracket first. We limit the sets of brackets in a phone number to four.
+	MATCHING_BRACKETS = regexp.MustCompile("(?:[" + OPENING_PARENS + "])?" + "(?:" + NON_PARENS + "+" + "[" + CLOSING_PARENS + "])?" + NON_PARENS + "+" + "(?:[" + OPENING_PARENS + "]" + NON_PARENS + "+[" + CLOSING_PARENS + "])" + BRACKET_PAIR_LIMIT + NON_PARENS + "*")
 
-	/**
-	 * Patterns used to extract phone numbers from a larger phone-number-like pattern. These are
-	 * ordered according to specificity. For example, white-space is last since that is frequently
-	 * used in numbers, not just to separate two numbers. We have separate patterns since we don't
-	 * want to break up the phone-number-like text on more than one different kind of symbol at one
-	 * time, although symbols of the same type (e.g. space) can be safely grouped together.
-	 *
-	 * Note that if there is a match, we will always check any text found up to the first match as
-	 * well.
-	 */
+	// Patterns used to extract phone numbers from a larger phone-number-like pattern. These are
+	// ordered according to specificity. For example, white-space is last since that is frequently
+	// used in numbers, not just to separate two numbers. We have separate patterns since we don't
+	// want to break up the phone-number-like text on more than one different kind of symbol at one
+	// time, although symbols of the same type (e.g. space) can be safely grouped together.
+	//
+	// Note that if there is a match, we will always check any text found up to the first match as
+	// well.
 	INNER_MATCHES = []*regexp.Regexp{
 		// Breaks on the slash - e.g. "651-234-2345/332-445-1234"
 		regexp.MustCompile("/+(.*)"),
@@ -84,8 +90,49 @@ var (
 		// Breaks on space - e.g. "3324451234 8002341234"
 		regexp.MustCompile("\\p{Z}+(\\P{Z}+)"),
 	}
+
+	//  Punctuation that may be at the start of a phone number - brackets and plus signs.
+	LEAD_CLASS   = OPENING_PARENS + PLUS_CHARS
+	LEAD_PATTERN = regexp.MustCompile(LEAD_CLASS)
+
+	// Builds the MATCHING_BRACKETS and PATTERN regular expressions. The building blocks below exist to make the pattern more easily understood.
+	OPENING_PARENS = "\\(\\[\uFF08\uFF3B"
+	CLOSING_PARENS = "\\)\\]\uFF09\uFF3D"
+	NON_PARENS     = "[^" + OPENING_PARENS + CLOSING_PARENS + "]"
+
+	// Limit on the number of pairs of brackets in a phone number.
+	BRACKET_PAIR_LIMIT = "{0,3}"
+
+	// Limit on the number of leading (plus) characters.
+	LEAD_LIMIT = "{0,2}"
+
+	// Limit on the number of consecutive punctuation characters.
+	PUNCTIATION_LIMIT = "{0,4}"
+
+	// The maximum number of digits allowed in a digit-separated block. As we allow all digits in a
+	//single block, set high enough to accommodate the entire national number and the international
+	//country code.
+	DIGIT_BLOCK_LIMIT = 17 + 3
+
+	// Limit on the number of blocks separated by punctuation. Uses digitBlockLimit since some
+	// formats use spaces to separate each digit.
+	BLOCK_LIMIT = "{0," + strconv.Itoa(DIGIT_BLOCK_LIMIT) + "}"
+
+	// A punctuation sequence allowing white space.
+	PUNCTUATION = "[" + VALID_PUNCTUATION + "]" + PUNCTIATION_LIMIT
+
+	// A digits block without punctuation.
+	DIGIT_SEQUENCE = "\\d{1," + strconv.Itoa(DIGIT_BLOCK_LIMIT) + "}"
 )
 
+// Creates a new instance.
+//
+// Arguments:
+// text -- The character sequence that we will search
+// country -- The country to assume for phone numbers not written in
+//            international format (with a leading plus, or with the
+//            international dialing prefix of the specified region). May be
+//            "ZZ" if only numbers with a leading plus should be considered.
 func NewPhoneNumberMatcher(text string, region string) PhoneNumberMatcher {
 	m := PhoneNumberMatcher{
 		text:            text,
@@ -100,6 +147,8 @@ func NewPhoneNumberMatcher(text string, region string) PhoneNumberMatcher {
 	return m
 }
 
+// Trims away any characters after the first match of pattern in
+// candidate, returning the trimmed version.
 func (*PhoneNumberMatcher) trimAfterFirstMatch(pattern *regexp.Regexp, candidate string) string {
 	trailingCharsMatch := pattern.FindStringIndex(candidate)
 	if trailingCharsMatch != nil {
@@ -112,20 +161,39 @@ func (*PhoneNumberMatcher) isInvalidPunctuationSymbol(char rune) bool {
 	return char == '%' || unicode.In(char, unicode.Sc)
 }
 
+// Parses a phone number from the candidate using phonenumberutil.parse and
+// verifies it matches the requested leniency. If parsing and verification succeed, a
+// corresponding PhoneNumberMatch is returned, otherwise this method returns None.
+//
+// Arguments:
+//
+// candidate -- The candidate match.
+//
+// offset -- The offset of candidate within self.text.
+//
+// Returns the parsed and validated phone number match, or None.
 func (p *PhoneNumberMatcher) parseAndVerify(candidate string, offset int) (*PhoneNumberMatch, error) {
+	// Check the candidate doesn't contain any formatting which would
+	// indicate that it really isn't a phone number.
 	if MATCHING_BRACKETS.FindStringIndex(candidate) == nil || PUB_PAGES.FindStringIndex(candidate) != nil {
 		return nil, nil
 	}
 
+	// If leniency is set to VALID or stricter, we also want to skip
+	// numbers that are surrounded by Latin alphabetic characters, to
+	// skip cases like abc8005001234 or 8005001234def.
 	if p.leniency >= VALID {
 		if offset > 0 && LEAD_PATTERN.FindStringIndex(candidate) == nil {
+			// If the candidate is not at the start of the text, and does
+			// not start with phone-number punctuation, check the previous
+			// character
 			previousChar := p.text[offset-1]
-			for i := 0; i < 4; i++ {
-				if utf8.Valid([]byte{previousChar}) {
-					break
-				}
-				previousChar = p.text[offset+i]
-			}
+			//for i := 0; i < 4; i++ {
+			//	if utf8.Valid([]byte{previousChar}) {
+			//		break
+			//	}
+			//	previousChar = p.text[offset+i]
+			//}
 			if p.isInvalidPunctuationSymbol(rune(previousChar)) || unicode.IsLetter(rune(previousChar)) {
 				return nil, nil
 			}
@@ -133,12 +201,12 @@ func (p *PhoneNumberMatcher) parseAndVerify(candidate string, offset int) (*Phon
 		lastCharIndex := offset + len(candidate)
 		if lastCharIndex < len(p.text) {
 			nextChar := p.text[lastCharIndex]
-			for i := 1; i < 5; i++ {
-				if utf8.Valid([]byte{nextChar}) {
-					break
-				}
-				nextChar = p.text[lastCharIndex-i]
-			}
+			//for i := 1; i < 5; i++ {
+			//	if utf8.Valid([]byte{nextChar}) {
+			//		break
+			//	}
+			//	nextChar = p.text[lastCharIndex-i]
+			//}
 			if p.isInvalidPunctuationSymbol(rune(nextChar)) || unicode.IsLetter(rune(nextChar)) {
 				return nil, nil
 			}
@@ -165,7 +233,10 @@ func (p *PhoneNumberMatcher) extractMatch(candidate string, offset int) *PhoneNu
 	}
 
 	if TIME_STAMPS.FindStringIndex(candidate) != nil {
-		return nil
+		followingText := p.text[offset+len(candidate):]
+		if TIME_STAMPS_SUFFIX.FindStringIndex(followingText) != nil {
+			return nil
+		}
 	}
 
 	match, _ := p.parseAndVerify(candidate, offset)

From c11d83c36ae3681bbdea2009ea342cadc73cfe71 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Thu, 25 Mar 2021 15:47:19 +0100
Subject: [PATCH 25/30] Added comments

---
 phonenumbermatcher.go | 55 +++++++++++++++++++++++++++++++++----------
 1 file changed, 43 insertions(+), 12 deletions(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 289c4f4..1fc81d0 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -188,12 +188,8 @@ func (p *PhoneNumberMatcher) parseAndVerify(candidate string, offset int) (*Phon
 			// not start with phone-number punctuation, check the previous
 			// character
 			previousChar := p.text[offset-1]
-			//for i := 0; i < 4; i++ {
-			//	if utf8.Valid([]byte{previousChar}) {
-			//		break
-			//	}
-			//	previousChar = p.text[offset+i]
-			//}
+			// We return nil if it is a latin letter or an invalid
+			// punctuation symbol
 			if p.isInvalidPunctuationSymbol(rune(previousChar)) || unicode.IsLetter(rune(previousChar)) {
 				return nil, nil
 			}
@@ -201,12 +197,6 @@ func (p *PhoneNumberMatcher) parseAndVerify(candidate string, offset int) (*Phon
 		lastCharIndex := offset + len(candidate)
 		if lastCharIndex < len(p.text) {
 			nextChar := p.text[lastCharIndex]
-			//for i := 1; i < 5; i++ {
-			//	if utf8.Valid([]byte{nextChar}) {
-			//		break
-			//	}
-			//	nextChar = p.text[lastCharIndex-i]
-			//}
 			if p.isInvalidPunctuationSymbol(rune(nextChar)) || unicode.IsLetter(rune(nextChar)) {
 				return nil, nil
 			}
@@ -219,6 +209,11 @@ func (p *PhoneNumberMatcher) parseAndVerify(candidate string, offset int) (*Phon
 	}
 
 	if p.leniency.Verify(number, candidate) {
+		//  We used parse(keep_raw_input=True) to create this number,
+		//  but for now we don't return the extra values parsed.
+		//  TODO: stop clearing all values here and switch all users
+		//  over to using raw_input rather than the raw_string of
+		//  PhoneNumberMatch.
 		match := NewPhoneNumberMatch(offset, candidate, *number)
 
 		return &match, nil
@@ -227,11 +222,23 @@ func (p *PhoneNumberMatcher) parseAndVerify(candidate string, offset int) (*Phon
 	return nil, nil
 }
 
+// Attempts to extract a match from a candidate string.
+//
+//  Arguments:
+//
+//  candidate -- The candidate text that might contain a phone number.
+//
+//  offset -- The offset of candidate within self.text
+//
+//  Returns the match found, None if none can be found
 func (p *PhoneNumberMatcher) extractMatch(candidate string, offset int) *PhoneNumberMatch {
+	// Skip a match that is more likely a publication page reference or a
+	// date.
 	if SLASH_SEPARATED_DATES.FindStringIndex(candidate) != nil {
 		return nil
 	}
 
+	// Skip potential time-stamps.
 	if TIME_STAMPS.FindStringIndex(candidate) != nil {
 		followingText := p.text[offset+len(candidate):]
 		if TIME_STAMPS_SUFFIX.FindStringIndex(followingText) != nil {
@@ -239,14 +246,27 @@ func (p *PhoneNumberMatcher) extractMatch(candidate string, offset int) *PhoneNu
 		}
 	}
 
+	// Try to come up with a valid match given the entire candidate.
 	match, _ := p.parseAndVerify(candidate, offset)
 	if match != nil {
 		return match
 	}
 
+	// If that failed, try to find an "inner match" -- there might be a
+	// phone number within this candidate.
 	return p.extractInnerMatch(candidate, offset)
 }
 
+// Attempts to extract a match from candidate if the whole candidate
+// does not qualify as a match.
+//
+//  Arguments:
+//
+//  candidate -- The candidate text that might contain a phone number
+//
+//  offset -- The current offset of candidate within text
+//
+//  Returns the match found, None if none can be found
 func (p *PhoneNumberMatcher) extractInnerMatch(candidate string, offset int) *PhoneNumberMatch {
 	for _, possibleInnerMatch := range INNER_MATCHES {
 		groupMatch := possibleInnerMatch.FindStringIndex(candidate)
@@ -257,6 +277,7 @@ func (p *PhoneNumberMatcher) extractInnerMatch(candidate string, offset int) *Ph
 				break
 			}
 			if isFirstMatch {
+				// We should handle any group before this one too.
 				group := p.trimAfterFirstMatch(UNWANTED_END_CHAR_PATTERN, candidate[:groupMatch[0]])
 				match, _ := p.parseAndVerify(group, offset)
 				if match != nil {
@@ -282,6 +303,14 @@ func (p *PhoneNumberMatcher) extractInnerMatch(candidate string, offset int) *Ph
 	return nil
 }
 
+// Attempts to find the next subsequence in the searched sequence on or after index
+// that represents a phone number. Returns the next match, None if none was found.
+//
+// Arguments:
+//
+// index -- The search index to start searching at.
+//
+// Returns the phone number match found, None if none can be found.
 func (p *PhoneNumberMatcher) find() *PhoneNumberMatch {
 	matcher := PATTERN.FindStringIndex(p.text[p.searchIndex:])
 	index := 0 + p.searchIndex
@@ -310,6 +339,7 @@ func (p *PhoneNumberMatcher) find() *PhoneNumberMatch {
 	return nil
 }
 
+// Indicates whether there is another match available
 func (p *PhoneNumberMatcher) hasNext() bool {
 	if p.state == notReady {
 		p.lastMatch = p.find()
@@ -323,6 +353,7 @@ func (p *PhoneNumberMatcher) hasNext() bool {
 	return p.state == ready
 }
 
+// Return the next match; raises Exception if no next match available
 func (p *PhoneNumberMatcher) Next() (*PhoneNumberMatch, error) {
 	if !p.hasNext() {
 		return nil, io.EOF

From a46b3749bf04d3108cffd33ab625c183e03a091b Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Thu, 25 Mar 2021 15:52:19 +0100
Subject: [PATCH 26/30] Reset unused files

---
 cmd/phoneserver/go.mod  |   2 +-
 cmd/phoneserver/main.go | 111 +++++++++++++++++++---------------------
 2 files changed, 55 insertions(+), 58 deletions(-)

diff --git a/cmd/phoneserver/go.mod b/cmd/phoneserver/go.mod
index fc7907f..c3fa7cb 100644
--- a/cmd/phoneserver/go.mod
+++ b/cmd/phoneserver/go.mod
@@ -7,4 +7,4 @@ replace github.com/nyaruka/phonenumbers => ../../
 require (
 	github.com/aws/aws-lambda-go v1.13.1
 	github.com/nyaruka/phonenumbers v0.0.0-00010101000000-000000000000
-)
+)
\ No newline at end of file
diff --git a/cmd/phoneserver/main.go b/cmd/phoneserver/main.go
index d057f36..2af237e 100644
--- a/cmd/phoneserver/main.go
+++ b/cmd/phoneserver/main.go
@@ -1,16 +1,13 @@
 package main
 
-//import (
-//	"encoding/json"
-//	"net/http"
-//
-//	"github.com/nyaruka/phonenumbers"
-//)import (
-//	"encoding/json"
-//	"net/http"
-//
-//	"github.com/nyaruka/phonenumbers"
-//)
+import (
+	"encoding/json"
+	"net/http"
+
+	"github.com/aws/aws-lambda-go/events"
+	"github.com/aws/aws-lambda-go/lambda"
+	"github.com/nyaruka/phonenumbers"
+)
 
 var Version = "dev"
 
@@ -29,49 +26,49 @@ type successResponse struct {
 	Version                string `json:"version"`
 }
 
-//func writeResponse(status int, body interface{}) (events.APIGatewayProxyResponse, error) {
-//	js, err := json.MarshalIndent(body, "", "    ")
-//	if err != nil {
-//		return events.APIGatewayProxyResponse{
-//			StatusCode: 500,
-//			Body:       err.Error(),
-//		}, nil
-//	}
-//
-//	return events.APIGatewayProxyResponse{
-//		StatusCode: 200,
-//		Body:       string(js),
-//		Headers:    map[string]string{"Content-Type": "application/json"},
-//	}, nil
-//}
-//
-//func parse(request events.APIGatewayProxyRequest) (events.APIGatewayProxyResponse, error) {
-//	phone := request.QueryStringParameters["phone"]
-//
-//	// required phone number
-//	if phone == "" {
-//		return writeResponse(http.StatusBadRequest, errorResponse{"missing body", "missing 'phone' parameter"})
-//	}
-//
-//	// optional country code
-//	country := request.QueryStringParameters["country"]
-//
-//	metadata, err := phonenumbers.Parse(phone, country)
-//	if err != nil {
-//		return writeResponse(http.StatusBadRequest, errorResponse{"error parsing phone", err.Error()})
-//	}
-//
-//	return writeResponse(http.StatusOK, successResponse{
-//		NationalNumber:         *metadata.NationalNumber,
-//		CountryCode:            *metadata.CountryCode,
-//		IsPossible:             phonenumbers.IsPossibleNumber(metadata),
-//		IsValid:                phonenumbers.IsValidNumber(metadata),
-//		NationalFormatted:      phonenumbers.Format(metadata, phonenumbers.NATIONAL),
-//		InternationalFormatted: phonenumbers.Format(metadata, phonenumbers.INTERNATIONAL),
-//		Version:                Version,
-//	})
-//}
-//
-//func main() {
-//	lambda.Start(parse)
-//}
+func writeResponse(status int, body interface{}) (events.APIGatewayProxyResponse, error) {
+	js, err := json.MarshalIndent(body, "", "    ")
+	if err != nil {
+		return events.APIGatewayProxyResponse{
+			StatusCode: 500,
+			Body:       err.Error(),
+		}, nil
+	}
+
+	return events.APIGatewayProxyResponse{
+		StatusCode: 200,
+		Body:       string(js),
+		Headers:    map[string]string{"Content-Type": "application/json"},
+	}, nil
+}
+
+func parse(request events.APIGatewayProxyRequest) (events.APIGatewayProxyResponse, error) {
+	phone := request.QueryStringParameters["phone"]
+
+	// required phone number
+	if phone == "" {
+		return writeResponse(http.StatusBadRequest, errorResponse{"missing body", "missing 'phone' parameter"})
+	}
+
+	// optional country code
+	country := request.QueryStringParameters["country"]
+
+	metadata, err := phonenumbers.Parse(phone, country)
+	if err != nil {
+		return writeResponse(http.StatusBadRequest, errorResponse{"error parsing phone", err.Error()})
+	}
+
+	return writeResponse(http.StatusOK, successResponse{
+		NationalNumber:         *metadata.NationalNumber,
+		CountryCode:            *metadata.CountryCode,
+		IsPossible:             phonenumbers.IsPossibleNumber(metadata),
+		IsValid:                phonenumbers.IsValidNumber(metadata),
+		NationalFormatted:      phonenumbers.Format(metadata, phonenumbers.NATIONAL),
+		InternationalFormatted: phonenumbers.Format(metadata, phonenumbers.INTERNATIONAL),
+		Version:                Version,
+	})
+}
+
+func main() {
+	lambda.Start(parse)
+}

From 1832dc624d090ca036752bed747d07960d62fd96 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Thu, 25 Mar 2021 15:54:54 +0100
Subject: [PATCH 27/30] Set mod to nyaruka

---
 go.mod | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/go.mod b/go.mod
index 24c16e7..fad8daa 100644
--- a/go.mod
+++ b/go.mod
@@ -1,4 +1,4 @@
-module github.com/mauritsderuiter95/phonenumbers
+module github.com/nyaruka/phonenumbers
 
 require (
 	github.com/golang/protobuf v1.3.2

From 30bc7151ca7feaf7c1287026d9fd5f7690caaabf Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Thu, 25 Mar 2021 15:56:25 +0100
Subject: [PATCH 28/30] Set go version to 1.13

---
 go.mod | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/go.mod b/go.mod
index fad8daa..b3c5356 100644
--- a/go.mod
+++ b/go.mod
@@ -1,7 +1,5 @@
 module github.com/nyaruka/phonenumbers
 
-require (
-	github.com/golang/protobuf v1.3.2
-)
+require github.com/golang/protobuf v1.3.2
 
-go 1.16
+go 1.13

From 7592927b95b9be1b4b7bac2c7c2e76d6fe5ec759 Mon Sep 17 00:00:00 2001
From: mauritsderuiter <maurits@leadinfo.com>
Date: Fri, 26 Mar 2021 13:23:02 +0100
Subject: [PATCH 29/30] Ran go mod tidy, reset cmd/phonnumbers/go.mod

---
 cmd/phoneserver/go.mod |  2 +-
 go.sum                 | 17 -----------------
 2 files changed, 1 insertion(+), 18 deletions(-)

diff --git a/cmd/phoneserver/go.mod b/cmd/phoneserver/go.mod
index c3fa7cb..fc7907f 100644
--- a/cmd/phoneserver/go.mod
+++ b/cmd/phoneserver/go.mod
@@ -7,4 +7,4 @@ replace github.com/nyaruka/phonenumbers => ../../
 require (
 	github.com/aws/aws-lambda-go v1.13.1
 	github.com/nyaruka/phonenumbers v0.0.0-00010101000000-000000000000
-)
\ No newline at end of file
+)
diff --git a/go.sum b/go.sum
index 83d6e04..9a46bf9 100644
--- a/go.sum
+++ b/go.sum
@@ -1,19 +1,2 @@
-github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
-github.com/aws/aws-lambda-go v1.23.0 h1:Vjwow5COkFJp7GePkk9kjAo/DyX36b7wVPKwseQZbRo=
-github.com/aws/aws-lambda-go v1.23.0/go.mod h1:jJmlefzPfGnckuHdXX7/80O3BvUUi12XOkbv4w9SGLU=
-github.com/cpuguy83/go-md2man/v2 v2.0.0-20190314233015-f79a8a8ca69d/go.mod h1:maD7wRr/U5Z6m/iR4s+kqSMx2CaBsrgA7czyZG/E6dU=
-github.com/cpuguy83/go-md2man/v2 v2.0.0/go.mod h1:maD7wRr/U5Z6m/iR4s+kqSMx2CaBsrgA7czyZG/E6dU=
-github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
-github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/golang/protobuf v1.3.2 h1:6nsPYzhq5kReh6QImI3k5qWzO4PEbvbIW2cwSfR/6xs=
 github.com/golang/protobuf v1.3.2/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
-github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
-github.com/russross/blackfriday/v2 v2.0.1/go.mod h1:+Rmxgy9KzJVeS9/2gXHxylqXiyQDYRxCVz55jmeOWTM=
-github.com/shurcooL/sanitized_anchor_name v1.0.0/go.mod h1:1NzhyTcUVG4SuEtjjoZeVRXNmyL/1OwPU0+IJeTBvfc=
-github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
-github.com/stretchr/testify v1.6.1/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
-github.com/urfave/cli/v2 v2.2.0/go.mod h1:SE9GqnLQmjVa0iPEY0f1w3ygNIYcIJ0OKPMoW2caLfQ=
-gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
-gopkg.in/yaml.v2 v2.2.2/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
-gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
-gopkg.in/yaml.v3 v3.0.0-20200615113413-eeeca48fe776/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=

From 0e9208758775143175f8c1d9cee3ad7b47487d49 Mon Sep 17 00:00:00 2001
From: Sami Kammoun <kammounsami77@gmail.com>
Date: Sat, 27 Jul 2024 10:57:07 +0100
Subject: [PATCH 30/30] allow gradual non-breaking releases

---
 phonenumbermatcher.go | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/phonenumbermatcher.go b/phonenumbermatcher.go
index 1fc81d0..086f46a 100644
--- a/phonenumbermatcher.go
+++ b/phonenumbermatcher.go
@@ -125,6 +125,11 @@ var (
 	DIGIT_SEQUENCE = "\\d{1," + strconv.Itoa(DIGIT_BLOCK_LIMIT) + "}"
 )
 
+func NewPhoneNumberMatcher(seq string) *PhoneNumberMatcher {
+	// TODO: Add region parameter and implement in the next major release (2.0)
+	return nil
+}
+
 // Creates a new instance.
 //
 // Arguments:
@@ -133,7 +138,8 @@ var (
 //            international format (with a leading plus, or with the
 //            international dialing prefix of the specified region). May be
 //            "ZZ" if only numbers with a leading plus should be considered.
-func NewPhoneNumberMatcher(text string, region string) PhoneNumberMatcher {
+func NewPhoneNumberMatcherForRegion(text string, region string) PhoneNumberMatcher {
+	// TODO: Deprecate in the next major release (2.0) and remove in the following major release (3.0).
 	m := PhoneNumberMatcher{
 		text:            text,
 		preferredRegion: region,