[67] | 1 | package regexp2
|
---|
| 2 |
|
---|
| 3 | import (
|
---|
| 4 | "bytes"
|
---|
| 5 | "errors"
|
---|
| 6 |
|
---|
| 7 | "github.com/dlclark/regexp2/syntax"
|
---|
| 8 | )
|
---|
| 9 |
|
---|
| 10 | const (
|
---|
| 11 | replaceSpecials = 4
|
---|
| 12 | replaceLeftPortion = -1
|
---|
| 13 | replaceRightPortion = -2
|
---|
| 14 | replaceLastGroup = -3
|
---|
| 15 | replaceWholeString = -4
|
---|
| 16 | )
|
---|
| 17 |
|
---|
| 18 | // MatchEvaluator is a function that takes a match and returns a replacement string to be used
|
---|
| 19 | type MatchEvaluator func(Match) string
|
---|
| 20 |
|
---|
| 21 | // Three very similar algorithms appear below: replace (pattern),
|
---|
| 22 | // replace (evaluator), and split.
|
---|
| 23 |
|
---|
| 24 | // Replace Replaces all occurrences of the regex in the string with the
|
---|
| 25 | // replacement pattern.
|
---|
| 26 | //
|
---|
| 27 | // Note that the special case of no matches is handled on its own:
|
---|
| 28 | // with no matches, the input string is returned unchanged.
|
---|
| 29 | // The right-to-left case is split out because StringBuilder
|
---|
| 30 | // doesn't handle right-to-left string building directly very well.
|
---|
| 31 | func replace(regex *Regexp, data *syntax.ReplacerData, evaluator MatchEvaluator, input string, startAt, count int) (string, error) {
|
---|
| 32 | if count < -1 {
|
---|
| 33 | return "", errors.New("Count too small")
|
---|
| 34 | }
|
---|
| 35 | if count == 0 {
|
---|
| 36 | return "", nil
|
---|
| 37 | }
|
---|
| 38 |
|
---|
| 39 | m, err := regex.FindStringMatchStartingAt(input, startAt)
|
---|
| 40 |
|
---|
| 41 | if err != nil {
|
---|
| 42 | return "", err
|
---|
| 43 | }
|
---|
| 44 | if m == nil {
|
---|
| 45 | return input, nil
|
---|
| 46 | }
|
---|
| 47 |
|
---|
| 48 | buf := &bytes.Buffer{}
|
---|
| 49 | text := m.text
|
---|
| 50 |
|
---|
| 51 | if !regex.RightToLeft() {
|
---|
| 52 | prevat := 0
|
---|
| 53 | for m != nil {
|
---|
| 54 | if m.Index != prevat {
|
---|
| 55 | buf.WriteString(string(text[prevat:m.Index]))
|
---|
| 56 | }
|
---|
| 57 | prevat = m.Index + m.Length
|
---|
| 58 | if evaluator == nil {
|
---|
| 59 | replacementImpl(data, buf, m)
|
---|
| 60 | } else {
|
---|
| 61 | buf.WriteString(evaluator(*m))
|
---|
| 62 | }
|
---|
| 63 |
|
---|
| 64 | count--
|
---|
| 65 | if count == 0 {
|
---|
| 66 | break
|
---|
| 67 | }
|
---|
| 68 | m, err = regex.FindNextMatch(m)
|
---|
| 69 | if err != nil {
|
---|
| 70 | return "", nil
|
---|
| 71 | }
|
---|
| 72 | }
|
---|
| 73 |
|
---|
| 74 | if prevat < len(text) {
|
---|
| 75 | buf.WriteString(string(text[prevat:]))
|
---|
| 76 | }
|
---|
| 77 | } else {
|
---|
| 78 | prevat := len(text)
|
---|
| 79 | var al []string
|
---|
| 80 |
|
---|
| 81 | for m != nil {
|
---|
| 82 | if m.Index+m.Length != prevat {
|
---|
| 83 | al = append(al, string(text[m.Index+m.Length:prevat]))
|
---|
| 84 | }
|
---|
| 85 | prevat = m.Index
|
---|
| 86 | if evaluator == nil {
|
---|
| 87 | replacementImplRTL(data, &al, m)
|
---|
| 88 | } else {
|
---|
| 89 | al = append(al, evaluator(*m))
|
---|
| 90 | }
|
---|
| 91 |
|
---|
| 92 | count--
|
---|
| 93 | if count == 0 {
|
---|
| 94 | break
|
---|
| 95 | }
|
---|
| 96 | m, err = regex.FindNextMatch(m)
|
---|
| 97 | if err != nil {
|
---|
| 98 | return "", nil
|
---|
| 99 | }
|
---|
| 100 | }
|
---|
| 101 |
|
---|
| 102 | if prevat > 0 {
|
---|
| 103 | buf.WriteString(string(text[:prevat]))
|
---|
| 104 | }
|
---|
| 105 |
|
---|
| 106 | for i := len(al) - 1; i >= 0; i-- {
|
---|
| 107 | buf.WriteString(al[i])
|
---|
| 108 | }
|
---|
| 109 | }
|
---|
| 110 |
|
---|
| 111 | return buf.String(), nil
|
---|
| 112 | }
|
---|
| 113 |
|
---|
| 114 | // Given a Match, emits into the StringBuilder the evaluated
|
---|
| 115 | // substitution pattern.
|
---|
| 116 | func replacementImpl(data *syntax.ReplacerData, buf *bytes.Buffer, m *Match) {
|
---|
| 117 | for _, r := range data.Rules {
|
---|
| 118 |
|
---|
| 119 | if r >= 0 { // string lookup
|
---|
| 120 | buf.WriteString(data.Strings[r])
|
---|
| 121 | } else if r < -replaceSpecials { // group lookup
|
---|
| 122 | m.groupValueAppendToBuf(-replaceSpecials-1-r, buf)
|
---|
| 123 | } else {
|
---|
| 124 | switch -replaceSpecials - 1 - r { // special insertion patterns
|
---|
| 125 | case replaceLeftPortion:
|
---|
| 126 | for i := 0; i < m.Index; i++ {
|
---|
| 127 | buf.WriteRune(m.text[i])
|
---|
| 128 | }
|
---|
| 129 | case replaceRightPortion:
|
---|
| 130 | for i := m.Index + m.Length; i < len(m.text); i++ {
|
---|
| 131 | buf.WriteRune(m.text[i])
|
---|
| 132 | }
|
---|
| 133 | case replaceLastGroup:
|
---|
| 134 | m.groupValueAppendToBuf(m.GroupCount()-1, buf)
|
---|
| 135 | case replaceWholeString:
|
---|
| 136 | for i := 0; i < len(m.text); i++ {
|
---|
| 137 | buf.WriteRune(m.text[i])
|
---|
| 138 | }
|
---|
| 139 | }
|
---|
| 140 | }
|
---|
| 141 | }
|
---|
| 142 | }
|
---|
| 143 |
|
---|
| 144 | func replacementImplRTL(data *syntax.ReplacerData, al *[]string, m *Match) {
|
---|
| 145 | l := *al
|
---|
| 146 | buf := &bytes.Buffer{}
|
---|
| 147 |
|
---|
| 148 | for _, r := range data.Rules {
|
---|
| 149 | buf.Reset()
|
---|
| 150 | if r >= 0 { // string lookup
|
---|
| 151 | l = append(l, data.Strings[r])
|
---|
| 152 | } else if r < -replaceSpecials { // group lookup
|
---|
| 153 | m.groupValueAppendToBuf(-replaceSpecials-1-r, buf)
|
---|
| 154 | l = append(l, buf.String())
|
---|
| 155 | } else {
|
---|
| 156 | switch -replaceSpecials - 1 - r { // special insertion patterns
|
---|
| 157 | case replaceLeftPortion:
|
---|
| 158 | for i := 0; i < m.Index; i++ {
|
---|
| 159 | buf.WriteRune(m.text[i])
|
---|
| 160 | }
|
---|
| 161 | case replaceRightPortion:
|
---|
| 162 | for i := m.Index + m.Length; i < len(m.text); i++ {
|
---|
| 163 | buf.WriteRune(m.text[i])
|
---|
| 164 | }
|
---|
| 165 | case replaceLastGroup:
|
---|
| 166 | m.groupValueAppendToBuf(m.GroupCount()-1, buf)
|
---|
| 167 | case replaceWholeString:
|
---|
| 168 | for i := 0; i < len(m.text); i++ {
|
---|
| 169 | buf.WriteRune(m.text[i])
|
---|
| 170 | }
|
---|
| 171 | }
|
---|
| 172 | l = append(l, buf.String())
|
---|
| 173 | }
|
---|
| 174 | }
|
---|
| 175 |
|
---|
| 176 | *al = l
|
---|
| 177 | }
|
---|