glob/match/btree.go

186 lines
3.7 KiB
Go
Raw Normal View History

2016-01-08 20:14:31 +03:00
package match
import (
"fmt"
2016-01-09 02:34:41 +03:00
"unicode/utf8"
2016-01-08 20:14:31 +03:00
)
type BTree struct {
2016-01-14 21:32:02 +03:00
Value Matcher
Left Matcher
Right Matcher
ValueLengthRunes int
LeftLengthRunes int
RightLengthRunes int
LengthRunes int
2016-01-08 20:14:31 +03:00
}
2016-01-14 18:29:13 +03:00
func NewBTree(Value, Left, Right Matcher) (tree BTree) {
tree.Value = Value
tree.Left = Left
tree.Right = Right
2016-01-08 20:14:31 +03:00
2016-01-14 18:29:13 +03:00
lenOk := true
2016-01-14 21:32:02 +03:00
if tree.ValueLengthRunes = Value.Len(); tree.ValueLengthRunes == -1 {
2016-01-14 18:29:13 +03:00
lenOk = false
}
if Left != nil {
2016-01-14 21:32:02 +03:00
if tree.LeftLengthRunes = Left.Len(); tree.LeftLengthRunes == -1 {
2016-01-14 18:29:13 +03:00
lenOk = false
}
}
2016-01-09 02:34:41 +03:00
2016-01-14 18:29:13 +03:00
if Right != nil {
2016-01-14 21:32:02 +03:00
if tree.RightLengthRunes = Right.Len(); tree.RightLengthRunes == -1 {
2016-01-14 18:29:13 +03:00
lenOk = false
}
2016-01-09 02:34:41 +03:00
}
2016-01-14 18:29:13 +03:00
if lenOk {
2016-01-14 21:32:02 +03:00
tree.LengthRunes = tree.LeftLengthRunes + tree.ValueLengthRunes + tree.RightLengthRunes
2016-01-14 18:29:13 +03:00
} else {
2016-01-14 21:32:02 +03:00
tree.LengthRunes = -1
2016-01-09 02:34:41 +03:00
}
2016-01-14 18:29:13 +03:00
return tree
}
2016-01-09 02:34:41 +03:00
func (self BTree) Len() int {
2016-01-14 21:32:02 +03:00
return self.LengthRunes
2016-01-09 02:34:41 +03:00
}
2016-01-14 18:29:13 +03:00
// todo?
2018-02-09 00:18:42 +03:00
func (self BTree) Index(s string) (index int, segments []int) {
//inputLen := len(s)
//// try to cut unnecessary parts
//// by knowledge of length of right and left part
//offset, limit := self.offsetLimit(inputLen)
//for offset < limit {
// // search for matching part in substring
// vi, segments := self.Value.Index(s[offset:limit])
// if index == -1 {
// return -1, nil
// }
// if self.Left == nil {
// if index != offset {
// return -1, nil
// }
// } else {
// left := s[:offset+vi]
// i := self.Left.IndexSuffix(left)
// if i == -1 {
// return -1, nil
// }
// index = i
// }
// if self.Right != nil {
// for _, seg := range segments {
// right := s[:offset+vi+seg]
// }
// }
// l := s[:offset+index]
// var left bool
// if self.Left != nil {
// left = self.Left.Index(l)
// } else {
// left = l == ""
// }
//}
2016-01-12 14:06:59 +03:00
return -1, nil
}
2016-01-08 20:14:31 +03:00
func (self BTree) Match(s string) bool {
2016-01-09 02:34:41 +03:00
inputLen := len(s)
2016-02-05 17:29:41 +03:00
// try to cut unnecessary parts
// by knowledge of length of right and left part
2018-02-09 00:18:42 +03:00
offset, limit := self.offsetLimit(inputLen)
2016-01-08 20:14:31 +03:00
2016-01-09 02:34:41 +03:00
for offset < limit {
2016-01-14 21:32:02 +03:00
// search for matching part in substring
2016-02-05 17:29:41 +03:00
index, segments := self.Value.Index(s[offset:limit])
2016-01-08 20:14:31 +03:00
if index == -1 {
releaseSegments(segments)
2016-01-08 20:14:31 +03:00
return false
}
2016-01-14 21:32:02 +03:00
l := s[:offset+index]
2016-01-09 02:34:41 +03:00
var left bool
if self.Left != nil {
left = self.Left.Match(l)
} else {
left = l == ""
}
2016-01-08 20:14:31 +03:00
2016-01-09 02:34:41 +03:00
if left {
for i := len(segments) - 1; i >= 0; i-- {
length := segments[i]
2016-01-08 20:14:31 +03:00
2016-01-09 02:34:41 +03:00
var right bool
var r string
// if there is no string for the right branch
if inputLen <= offset+index+length {
r = ""
} else {
r = s[offset+index+length:]
}
if self.Right != nil {
right = self.Right.Match(r)
} else {
right = r == ""
}
2016-01-08 20:14:31 +03:00
2016-01-09 02:34:41 +03:00
if right {
releaseSegments(segments)
2016-01-09 02:34:41 +03:00
return true
}
2016-01-08 20:14:31 +03:00
}
}
2016-01-09 02:34:41 +03:00
_, step := utf8.DecodeRuneInString(s[offset+index:])
offset += index + step
releaseSegments(segments)
2016-01-08 20:14:31 +03:00
}
return false
}
2018-02-09 00:18:42 +03:00
func (self BTree) offsetLimit(inputLen int) (offset int, limit int) {
// self.Length, self.RLen and self.LLen are values meaning the length of runes for each part
// here we manipulating byte length for better optimizations
// but these checks still works, cause minLen of 1-rune string is 1 byte.
if self.LengthRunes != -1 && self.LengthRunes > inputLen {
return 0, 0
}
if self.LeftLengthRunes >= 0 {
offset = self.LeftLengthRunes
}
if self.RightLengthRunes >= 0 {
limit = inputLen - self.RightLengthRunes
} else {
limit = inputLen
}
return offset, limit
}
2016-01-08 20:14:31 +03:00
func (self BTree) String() string {
const n string = "<nil>"
var l, r string
if self.Left == nil {
l = n
} else {
l = self.Left.String()
}
if self.Right == nil {
r = n
} else {
r = self.Right.String()
}
return fmt.Sprintf("<btree:[%s<-%s->%s]>", l, self.Value, r)
2016-01-08 20:14:31 +03:00
}