open-consul/vendor/k8s.io/client-go/util/jsonpath/parser.go

525 lines
11 KiB
Go

/*
Copyright 2015 The Kubernetes Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
package jsonpath
import (
"errors"
"fmt"
"regexp"
"strconv"
"strings"
"unicode"
"unicode/utf8"
)
const eof = -1
const (
leftDelim = "{"
rightDelim = "}"
)
type Parser struct {
Name string
Root *ListNode
input string
pos int
start int
width int
}
var (
ErrSyntax = errors.New("invalid syntax")
dictKeyRex = regexp.MustCompile(`^'([^']*)'$`)
sliceOperatorRex = regexp.MustCompile(`^(-?[\d]*)(:-?[\d]*)?(:-?[\d]*)?$`)
)
// Parse parsed the given text and return a node Parser.
// If an error is encountered, parsing stops and an empty
// Parser is returned with the error
func Parse(name, text string) (*Parser, error) {
p := NewParser(name)
err := p.Parse(text)
if err != nil {
p = nil
}
return p, err
}
func NewParser(name string) *Parser {
return &Parser{
Name: name,
}
}
// parseAction parsed the expression inside delimiter
func parseAction(name, text string) (*Parser, error) {
p, err := Parse(name, fmt.Sprintf("%s%s%s", leftDelim, text, rightDelim))
// when error happens, p will be nil, so we need to return here
if err != nil {
return p, err
}
p.Root = p.Root.Nodes[0].(*ListNode)
return p, nil
}
func (p *Parser) Parse(text string) error {
p.input = text
p.Root = newList()
p.pos = 0
return p.parseText(p.Root)
}
// consumeText return the parsed text since last cosumeText
func (p *Parser) consumeText() string {
value := p.input[p.start:p.pos]
p.start = p.pos
return value
}
// next returns the next rune in the input.
func (p *Parser) next() rune {
if p.pos >= len(p.input) {
p.width = 0
return eof
}
r, w := utf8.DecodeRuneInString(p.input[p.pos:])
p.width = w
p.pos += p.width
return r
}
// peek returns but does not consume the next rune in the input.
func (p *Parser) peek() rune {
r := p.next()
p.backup()
return r
}
// backup steps back one rune. Can only be called once per call of next.
func (p *Parser) backup() {
p.pos -= p.width
}
func (p *Parser) parseText(cur *ListNode) error {
for {
if strings.HasPrefix(p.input[p.pos:], leftDelim) {
if p.pos > p.start {
cur.append(newText(p.consumeText()))
}
return p.parseLeftDelim(cur)
}
if p.next() == eof {
break
}
}
// Correctly reached EOF.
if p.pos > p.start {
cur.append(newText(p.consumeText()))
}
return nil
}
// parseLeftDelim scans the left delimiter, which is known to be present.
func (p *Parser) parseLeftDelim(cur *ListNode) error {
p.pos += len(leftDelim)
p.consumeText()
newNode := newList()
cur.append(newNode)
cur = newNode
return p.parseInsideAction(cur)
}
func (p *Parser) parseInsideAction(cur *ListNode) error {
prefixMap := map[string]func(*ListNode) error{
rightDelim: p.parseRightDelim,
"[?(": p.parseFilter,
"..": p.parseRecursive,
}
for prefix, parseFunc := range prefixMap {
if strings.HasPrefix(p.input[p.pos:], prefix) {
return parseFunc(cur)
}
}
switch r := p.next(); {
case r == eof || isEndOfLine(r):
return fmt.Errorf("unclosed action")
case r == ' ':
p.consumeText()
case r == '@' || r == '$': //the current object, just pass it
p.consumeText()
case r == '[':
return p.parseArray(cur)
case r == '"' || r == '\'':
return p.parseQuote(cur, r)
case r == '.':
return p.parseField(cur)
case r == '+' || r == '-' || unicode.IsDigit(r):
p.backup()
return p.parseNumber(cur)
case isAlphaNumeric(r):
p.backup()
return p.parseIdentifier(cur)
default:
return fmt.Errorf("unrecognized character in action: %#U", r)
}
return p.parseInsideAction(cur)
}
// parseRightDelim scans the right delimiter, which is known to be present.
func (p *Parser) parseRightDelim(cur *ListNode) error {
p.pos += len(rightDelim)
p.consumeText()
return p.parseText(p.Root)
}
// parseIdentifier scans build-in keywords, like "range" "end"
func (p *Parser) parseIdentifier(cur *ListNode) error {
var r rune
for {
r = p.next()
if isTerminator(r) {
p.backup()
break
}
}
value := p.consumeText()
if isBool(value) {
v, err := strconv.ParseBool(value)
if err != nil {
return fmt.Errorf("can not parse bool '%s': %s", value, err.Error())
}
cur.append(newBool(v))
} else {
cur.append(newIdentifier(value))
}
return p.parseInsideAction(cur)
}
// parseRecursive scans the recursive desent operator ..
func (p *Parser) parseRecursive(cur *ListNode) error {
p.pos += len("..")
p.consumeText()
cur.append(newRecursive())
if r := p.peek(); isAlphaNumeric(r) {
return p.parseField(cur)
}
return p.parseInsideAction(cur)
}
// parseNumber scans number
func (p *Parser) parseNumber(cur *ListNode) error {
r := p.peek()
if r == '+' || r == '-' {
p.next()
}
for {
r = p.next()
if r != '.' && !unicode.IsDigit(r) {
p.backup()
break
}
}
value := p.consumeText()
i, err := strconv.Atoi(value)
if err == nil {
cur.append(newInt(i))
return p.parseInsideAction(cur)
}
d, err := strconv.ParseFloat(value, 64)
if err == nil {
cur.append(newFloat(d))
return p.parseInsideAction(cur)
}
return fmt.Errorf("cannot parse number %s", value)
}
// parseArray scans array index selection
func (p *Parser) parseArray(cur *ListNode) error {
Loop:
for {
switch p.next() {
case eof, '\n':
return fmt.Errorf("unterminated array")
case ']':
break Loop
}
}
text := p.consumeText()
text = text[1 : len(text)-1]
if text == "*" {
text = ":"
}
//union operator
strs := strings.Split(text, ",")
if len(strs) > 1 {
union := []*ListNode{}
for _, str := range strs {
parser, err := parseAction("union", fmt.Sprintf("[%s]", strings.Trim(str, " ")))
if err != nil {
return err
}
union = append(union, parser.Root)
}
cur.append(newUnion(union))
return p.parseInsideAction(cur)
}
// dict key
value := dictKeyRex.FindStringSubmatch(text)
if value != nil {
parser, err := parseAction("arraydict", fmt.Sprintf(".%s", value[1]))
if err != nil {
return err
}
for _, node := range parser.Root.Nodes {
cur.append(node)
}
return p.parseInsideAction(cur)
}
//slice operator
value = sliceOperatorRex.FindStringSubmatch(text)
if value == nil {
return fmt.Errorf("invalid array index %s", text)
}
value = value[1:]
params := [3]ParamsEntry{}
for i := 0; i < 3; i++ {
if value[i] != "" {
if i > 0 {
value[i] = value[i][1:]
}
if i > 0 && value[i] == "" {
params[i].Known = false
} else {
var err error
params[i].Known = true
params[i].Value, err = strconv.Atoi(value[i])
if err != nil {
return fmt.Errorf("array index %s is not a number", value[i])
}
}
} else {
if i == 1 {
params[i].Known = true
params[i].Value = params[0].Value + 1
params[i].Derived = true
} else {
params[i].Known = false
params[i].Value = 0
}
}
}
cur.append(newArray(params))
return p.parseInsideAction(cur)
}
// parseFilter scans filter inside array selection
func (p *Parser) parseFilter(cur *ListNode) error {
p.pos += len("[?(")
p.consumeText()
begin := false
end := false
var pair rune
Loop:
for {
r := p.next()
switch r {
case eof, '\n':
return fmt.Errorf("unterminated filter")
case '"', '\'':
if begin == false {
//save the paired rune
begin = true
pair = r
continue
}
//only add when met paired rune
if p.input[p.pos-2] != '\\' && r == pair {
end = true
}
case ')':
//in rightParser below quotes only appear zero or once
//and must be paired at the beginning and end
if begin == end {
break Loop
}
}
}
if p.next() != ']' {
return fmt.Errorf("unclosed array expect ]")
}
reg := regexp.MustCompile(`^([^!<>=]+)([!<>=]+)(.+?)$`)
text := p.consumeText()
text = text[:len(text)-2]
value := reg.FindStringSubmatch(text)
if value == nil {
parser, err := parseAction("text", text)
if err != nil {
return err
}
cur.append(newFilter(parser.Root, newList(), "exists"))
} else {
leftParser, err := parseAction("left", value[1])
if err != nil {
return err
}
rightParser, err := parseAction("right", value[3])
if err != nil {
return err
}
cur.append(newFilter(leftParser.Root, rightParser.Root, value[2]))
}
return p.parseInsideAction(cur)
}
// parseQuote unquotes string inside double or single quote
func (p *Parser) parseQuote(cur *ListNode, end rune) error {
Loop:
for {
switch p.next() {
case eof, '\n':
return fmt.Errorf("unterminated quoted string")
case end:
//if it's not escape break the Loop
if p.input[p.pos-2] != '\\' {
break Loop
}
}
}
value := p.consumeText()
s, err := UnquoteExtend(value)
if err != nil {
return fmt.Errorf("unquote string %s error %v", value, err)
}
cur.append(newText(s))
return p.parseInsideAction(cur)
}
// parseField scans a field until a terminator
func (p *Parser) parseField(cur *ListNode) error {
p.consumeText()
for p.advance() {
}
value := p.consumeText()
if value == "*" {
cur.append(newWildcard())
} else {
cur.append(newField(strings.Replace(value, "\\", "", -1)))
}
return p.parseInsideAction(cur)
}
// advance scans until next non-escaped terminator
func (p *Parser) advance() bool {
r := p.next()
if r == '\\' {
p.next()
} else if isTerminator(r) {
p.backup()
return false
}
return true
}
// isTerminator reports whether the input is at valid termination character to appear after an identifier.
func isTerminator(r rune) bool {
if isSpace(r) || isEndOfLine(r) {
return true
}
switch r {
case eof, '.', ',', '[', ']', '$', '@', '{', '}':
return true
}
return false
}
// isSpace reports whether r is a space character.
func isSpace(r rune) bool {
return r == ' ' || r == '\t'
}
// isEndOfLine reports whether r is an end-of-line character.
func isEndOfLine(r rune) bool {
return r == '\r' || r == '\n'
}
// isAlphaNumeric reports whether r is an alphabetic, digit, or underscore.
func isAlphaNumeric(r rune) bool {
return r == '_' || unicode.IsLetter(r) || unicode.IsDigit(r)
}
// isBool reports whether s is a boolean value.
func isBool(s string) bool {
return s == "true" || s == "false"
}
//UnquoteExtend is almost same as strconv.Unquote(), but it support parse single quotes as a string
func UnquoteExtend(s string) (string, error) {
n := len(s)
if n < 2 {
return "", ErrSyntax
}
quote := s[0]
if quote != s[n-1] {
return "", ErrSyntax
}
s = s[1 : n-1]
if quote != '"' && quote != '\'' {
return "", ErrSyntax
}
// Is it trivial? Avoid allocation.
if !contains(s, '\\') && !contains(s, quote) {
return s, nil
}
var runeTmp [utf8.UTFMax]byte
buf := make([]byte, 0, 3*len(s)/2) // Try to avoid more allocations.
for len(s) > 0 {
c, multibyte, ss, err := strconv.UnquoteChar(s, quote)
if err != nil {
return "", err
}
s = ss
if c < utf8.RuneSelf || !multibyte {
buf = append(buf, byte(c))
} else {
n := utf8.EncodeRune(runeTmp[:], c)
buf = append(buf, runeTmp[:n]...)
}
}
return string(buf), nil
}
func contains(s string, c byte) bool {
for i := 0; i < len(s); i++ {
if s[i] == c {
return true
}
}
return false
}