yq/pkg/yqlib/expression_tokeniser.go
2021-09-12 16:55:55 +10:00

494 lines
17 KiB
Go

package yqlib
import (
"fmt"
"strconv"
"strings"
lex "github.com/timtadh/lexmachine"
"github.com/timtadh/lexmachine/machines"
)
func skip(*lex.Scanner, *machines.Match) (interface{}, error) {
return nil, nil
}
type tokenType uint32
const (
operationToken = 1 << iota
openBracket
closeBracket
openCollect
closeCollect
openCollectObject
closeCollectObject
traverseArrayCollect
)
type token struct {
TokenType tokenType
Operation *Operation
AssignOperation *Operation // e.g. tag (GetTag) op becomes AssignTag if '=' follows it
CheckForPostTraverse bool // e.g. [1]cat should really be [1].cat
Match *machines.Match // match that created this token
}
func (t *token) toString(detail bool) string {
if t.TokenType == operationToken {
if detail {
return fmt.Sprintf("%v (%v)", t.Operation.toString(), t.Operation.OperationType.Precedence)
}
return t.Operation.toString()
} else if t.TokenType == openBracket {
return "("
} else if t.TokenType == closeBracket {
return ")"
} else if t.TokenType == openCollect {
return "["
} else if t.TokenType == closeCollect {
return "]"
} else if t.TokenType == openCollectObject {
return "{"
} else if t.TokenType == closeCollectObject {
return "}"
} else if t.TokenType == traverseArrayCollect {
return ".["
} else {
return "NFI"
}
}
func pathToken(wrapped bool) lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
value := string(m.Bytes)
prefs := traversePreferences{}
if value[len(value)-1:] == "?" {
prefs.OptionalTraverse = true
value = value[:len(value)-1]
}
value = value[1:]
if wrapped {
value = unwrap(value)
}
log.Debug("PathToken %v", value)
op := &Operation{OperationType: traversePathOpType, Value: value, StringValue: value, Preferences: prefs}
return &token{TokenType: operationToken, Operation: op, CheckForPostTraverse: true}, nil
}
}
func opToken(op *operationType) lex.Action {
return opTokenWithPrefs(op, nil, nil)
}
func opAssignableToken(opType *operationType, assignOpType *operationType) lex.Action {
return opTokenWithPrefs(opType, assignOpType, nil)
}
func assignOpToken(updateAssign bool) lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
log.Debug("assignOpToken %v", string(m.Bytes))
value := string(m.Bytes)
op := &Operation{OperationType: assignOpType, Value: assignOpType.Type, StringValue: value, UpdateAssign: updateAssign}
return &token{TokenType: operationToken, Operation: op}, nil
}
}
func multiplyWithPrefs() lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
prefs := multiplyPreferences{}
options := string(m.Bytes)
if strings.Contains(options, "+") {
prefs.AppendArrays = true
}
if strings.Contains(options, "?") {
prefs.TraversePrefs = traversePreferences{DontAutoCreate: true}
}
if strings.Contains(options, "d") {
prefs.DeepMergeArrays = true
}
op := &Operation{OperationType: multiplyOpType, Value: multiplyOpType.Type, StringValue: options, Preferences: prefs}
return &token{TokenType: operationToken, Operation: op}, nil
}
}
func opTokenWithPrefs(op *operationType, assignOpType *operationType, preferences interface{}) lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
log.Debug("opTokenWithPrefs %v", string(m.Bytes))
value := string(m.Bytes)
op := &Operation{OperationType: op, Value: op.Type, StringValue: value, Preferences: preferences}
var assign *Operation
if assignOpType != nil {
assign = &Operation{OperationType: assignOpType, Value: assignOpType.Type, StringValue: value, Preferences: preferences}
}
return &token{TokenType: operationToken, Operation: op, AssignOperation: assign}, nil
}
}
func assignAllCommentsOp(updateAssign bool) lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
log.Debug("assignAllCommentsOp %v", string(m.Bytes))
value := string(m.Bytes)
op := &Operation{
OperationType: assignCommentOpType,
Value: assignCommentOpType.Type,
StringValue: value,
UpdateAssign: updateAssign,
Preferences: commentOpPreferences{LineComment: true, HeadComment: true, FootComment: true},
}
return &token{TokenType: operationToken, Operation: op}, nil
}
}
func literalToken(pType tokenType, checkForPost bool) lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
return &token{TokenType: pType, CheckForPostTraverse: checkForPost, Match: m}, nil
}
}
func unwrap(value string) string {
return value[1 : len(value)-1]
}
func numberValue() lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
var numberString = string(m.Bytes)
var number, errParsingInt = strconv.ParseInt(numberString, 10, 64) // nolint
if errParsingInt != nil {
return nil, errParsingInt
}
return &token{TokenType: operationToken, Operation: createValueOperation(number, numberString)}, nil
}
}
func hexValue() lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
var originalString = string(m.Bytes)
var numberString = originalString[2:]
log.Debugf("numberString: %v", numberString)
var number, errParsingInt = strconv.ParseInt(numberString, 16, 64) // nolint
if errParsingInt != nil {
return nil, errParsingInt
}
return &token{TokenType: operationToken, Operation: createValueOperation(number, originalString)}, nil
}
}
func floatValue() lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
var numberString = string(m.Bytes)
var number, errParsingInt = strconv.ParseFloat(numberString, 64) // nolint
if errParsingInt != nil {
return nil, errParsingInt
}
return &token{TokenType: operationToken, Operation: createValueOperation(number, numberString)}, nil
}
}
func booleanValue(val bool) lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
return &token{TokenType: operationToken, Operation: createValueOperation(val, string(m.Bytes))}, nil
}
}
func stringValue(wrapped bool) lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
value := string(m.Bytes)
if wrapped {
value = unwrap(value)
}
value = strings.ReplaceAll(value, "\\\"", "\"")
return &token{TokenType: operationToken, Operation: createValueOperation(value, value)}, nil
}
}
func getVariableOpToken() lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
value := string(m.Bytes)
value = value[1:]
getVarOperation := createValueOperation(value, value)
getVarOperation.OperationType = getVariableOpType
return &token{TokenType: operationToken, Operation: getVarOperation, CheckForPostTraverse: true}, nil
}
}
func envOp(strenv bool) lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
value := string(m.Bytes)
preferences := envOpPreferences{}
if strenv {
// strenv( )
value = value[7 : len(value)-1]
preferences.StringValue = true
} else {
//env( )
value = value[4 : len(value)-1]
}
envOperation := createValueOperation(value, value)
envOperation.OperationType = envOpType
envOperation.Preferences = preferences
return &token{TokenType: operationToken, Operation: envOperation}, nil
}
}
func nullValue() lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
return &token{TokenType: operationToken, Operation: createValueOperation(nil, string(m.Bytes))}, nil
}
}
func selfToken() lex.Action {
return func(s *lex.Scanner, m *machines.Match) (interface{}, error) {
op := &Operation{OperationType: selfReferenceOpType}
return &token{TokenType: operationToken, Operation: op}, nil
}
}
func initLexer() (*lex.Lexer, error) {
lexer := lex.NewLexer()
lexer.Add([]byte(`\(`), literalToken(openBracket, false))
lexer.Add([]byte(`\)`), literalToken(closeBracket, true))
lexer.Add([]byte(`\.\[`), literalToken(traverseArrayCollect, false))
lexer.Add([]byte(`\.\.`), opTokenWithPrefs(recursiveDescentOpType, nil, recursiveDescentPreferences{RecurseArray: true,
TraversePreferences: traversePreferences{DontFollowAlias: true, IncludeMapKeys: false}}))
lexer.Add([]byte(`\.\.\.`), opTokenWithPrefs(recursiveDescentOpType, nil, recursiveDescentPreferences{RecurseArray: true,
TraversePreferences: traversePreferences{DontFollowAlias: true, IncludeMapKeys: true}}))
lexer.Add([]byte(`,`), opToken(unionOpType))
lexer.Add([]byte(`:\s*`), opToken(createMapOpType))
lexer.Add([]byte(`length`), opToken(lengthOpType))
lexer.Add([]byte(`sortKeys`), opToken(sortKeysOpType))
lexer.Add([]byte(`select`), opToken(selectOpType))
lexer.Add([]byte(`has`), opToken(hasOpType))
lexer.Add([]byte(`unique`), opToken(uniqueOpType))
lexer.Add([]byte(`unique_by`), opToken(uniqueByOpType))
lexer.Add([]byte(`explode`), opToken(explodeOpType))
lexer.Add([]byte(`or`), opToken(orOpType))
lexer.Add([]byte(`and`), opToken(andOpType))
lexer.Add([]byte(`not`), opToken(notOpType))
lexer.Add([]byte(`ireduce`), opToken(reduceOpType))
lexer.Add([]byte(`;`), opToken(blockOpType))
lexer.Add([]byte(`\/\/`), opToken(alternativeOpType))
lexer.Add([]byte(`documentIndex`), opToken(getDocumentIndexOpType))
lexer.Add([]byte(`di`), opToken(getDocumentIndexOpType))
lexer.Add([]byte(`splitDoc`), opToken(splitDocumentOpType))
lexer.Add([]byte(`join`), opToken(joinStringOpType))
lexer.Add([]byte(`sub`), opToken(subStringOpType))
lexer.Add([]byte(`match`), opToken(matchOpType))
lexer.Add([]byte(`capture`), opToken(captureOpType))
lexer.Add([]byte(`test`), opToken(testOpType))
lexer.Add([]byte(`any`), opToken(anyOpType))
lexer.Add([]byte(`any_c`), opToken(anyConditionOpType))
lexer.Add([]byte(`all`), opToken(allOpType))
lexer.Add([]byte(`all_c`), opToken(allConditionOpType))
lexer.Add([]byte(`split`), opToken(splitStringOpType))
lexer.Add([]byte(`keys`), opToken(keysOpType))
lexer.Add([]byte(`style`), opAssignableToken(getStyleOpType, assignStyleOpType))
lexer.Add([]byte(`tag`), opAssignableToken(getTagOpType, assignTagOpType))
lexer.Add([]byte(`anchor`), opAssignableToken(getAnchorOpType, assignAnchorOpType))
lexer.Add([]byte(`alias`), opAssignableToken(getAliasOptype, assignAliasOpType))
lexer.Add([]byte(`filename`), opToken(getFilenameOpType))
lexer.Add([]byte(`fileIndex`), opToken(getFileIndexOpType))
lexer.Add([]byte(`fi`), opToken(getFileIndexOpType))
lexer.Add([]byte(`path`), opToken(getPathOpType))
lexer.Add([]byte(`to_entries`), opToken(toEntriesOpType))
lexer.Add([]byte(`from_entries`), opToken(fromEntriesOpType))
lexer.Add([]byte(`with_entries`), opToken(withEntriesOpType))
lexer.Add([]byte(`lineComment`), opTokenWithPrefs(getCommentOpType, assignCommentOpType, commentOpPreferences{LineComment: true}))
lexer.Add([]byte(`headComment`), opTokenWithPrefs(getCommentOpType, assignCommentOpType, commentOpPreferences{HeadComment: true}))
lexer.Add([]byte(`footComment`), opTokenWithPrefs(getCommentOpType, assignCommentOpType, commentOpPreferences{FootComment: true}))
lexer.Add([]byte(`comments\s*=`), assignAllCommentsOp(false))
lexer.Add([]byte(`comments\s*\|=`), assignAllCommentsOp(true))
lexer.Add([]byte(`collect`), opToken(collectOpType))
lexer.Add([]byte(`\s*==\s*`), opToken(equalsOpType))
lexer.Add([]byte(`\s*!=\s*`), opToken(notEqualsOpType))
lexer.Add([]byte(`\s*=\s*`), assignOpToken(false))
lexer.Add([]byte(`del`), opToken(deleteChildOpType))
lexer.Add([]byte(`\s*\|=\s*`), assignOpToken(true))
lexer.Add([]byte("( |\t|\n|\r)+"), skip)
lexer.Add([]byte(`\."[^ "]+"\??`), pathToken(true))
lexer.Add([]byte(`\.[^ \}\{\:\[\],\|\.\[\(\)=\n]+\??`), pathToken(false))
lexer.Add([]byte(`\.`), selfToken())
lexer.Add([]byte(`\|`), opToken(pipeOpType))
lexer.Add([]byte(`0[xX][0-9A-Fa-f]+`), hexValue())
lexer.Add([]byte(`-?\d+(\.\d+)`), floatValue())
lexer.Add([]byte(`-?[1-9](\.\d+)?[Ee][-+]?\d+`), floatValue())
lexer.Add([]byte(`-?\d+`), numberValue())
lexer.Add([]byte(`[Tt][Rr][Uu][Ee]`), booleanValue(true))
lexer.Add([]byte(`[Ff][Aa][Ll][Ss][Ee]`), booleanValue(false))
lexer.Add([]byte(`[Nn][Uu][Ll][Ll]`), nullValue())
lexer.Add([]byte(`~`), nullValue())
lexer.Add([]byte(`"([^"\\]*(\\.[^"\\]*)*)"`), stringValue(true))
lexer.Add([]byte(`strenv\([^\)]+\)`), envOp(true))
lexer.Add([]byte(`env\([^\)]+\)`), envOp(false))
lexer.Add([]byte(`\[`), literalToken(openCollect, false))
lexer.Add([]byte(`\]\??`), literalToken(closeCollect, true))
lexer.Add([]byte(`\{`), literalToken(openCollectObject, false))
lexer.Add([]byte(`\}`), literalToken(closeCollectObject, true))
lexer.Add([]byte(`\*[\+|\?d]*`), multiplyWithPrefs())
lexer.Add([]byte(`\+`), opToken(addOpType))
lexer.Add([]byte(`\+=`), opToken(addAssignOpType))
lexer.Add([]byte(`\-`), opToken(subtractOpType))
lexer.Add([]byte(`\-=`), opToken(subtractAssignOpType))
lexer.Add([]byte(`\$[a-zA-Z_-0-9]+`), getVariableOpToken())
lexer.Add([]byte(`as`), opTokenWithPrefs(assignVariableOpType, nil, assignVarPreferences{}))
lexer.Add([]byte(`ref`), opTokenWithPrefs(assignVariableOpType, nil, assignVarPreferences{IsReference: true}))
err := lexer.CompileNFA()
if err != nil {
return nil, err
}
return lexer, nil
}
type expressionTokeniser interface {
Tokenise(expression string) ([]*token, error)
}
type expressionTokeniserImpl struct {
lexer *lex.Lexer
}
func newExpressionTokeniser() expressionTokeniser {
var lexer, err = initLexer()
if err != nil {
panic(err)
}
return &expressionTokeniserImpl{lexer}
}
func (p *expressionTokeniserImpl) Tokenise(expression string) ([]*token, error) {
scanner, err := p.lexer.Scanner([]byte(expression))
if err != nil {
return nil, fmt.Errorf("Parsing expression: %v", err)
}
var tokens []*token
for tok, err, eof := scanner.Next(); !eof; tok, err, eof = scanner.Next() {
if tok != nil {
currentToken := tok.(*token)
log.Debugf("Tokenising %v", currentToken.toString(true))
tokens = append(tokens, currentToken)
}
if err != nil {
return nil, fmt.Errorf("Parsing expression: %v", err)
}
}
var postProcessedTokens = make([]*token, 0)
skipNextToken := false
for index := range tokens {
if skipNextToken {
skipNextToken = false
} else {
postProcessedTokens, skipNextToken = p.handleToken(tokens, index, postProcessedTokens)
}
}
return postProcessedTokens, nil
}
func (p *expressionTokeniserImpl) handleToken(tokens []*token, index int, postProcessedTokens []*token) (tokensAccum []*token, skipNextToken bool) {
skipNextToken = false
currentToken := tokens[index]
log.Debug("processing %v", currentToken.toString(true))
if currentToken.TokenType == traverseArrayCollect {
//need to put a traverse array then a collect currentToken
// do this by adding traverse then converting currentToken to collect
if index == 0 || tokens[index-1].TokenType != operationToken ||
tokens[index-1].Operation.OperationType != traversePathOpType {
log.Debug(" adding self")
op := &Operation{OperationType: selfReferenceOpType, StringValue: "SELF"}
postProcessedTokens = append(postProcessedTokens, &token{TokenType: operationToken, Operation: op})
}
log.Debug(" adding traverse array")
op := &Operation{OperationType: traverseArrayOpType, StringValue: "TRAVERSE_ARRAY"}
postProcessedTokens = append(postProcessedTokens, &token{TokenType: operationToken, Operation: op})
currentToken = &token{TokenType: openCollect}
}
if index != len(tokens)-1 && currentToken.AssignOperation != nil &&
tokens[index+1].TokenType == operationToken &&
tokens[index+1].Operation.OperationType == assignOpType {
log.Debug(" its an update assign")
currentToken.Operation = currentToken.AssignOperation
currentToken.Operation.UpdateAssign = tokens[index+1].Operation.UpdateAssign
skipNextToken = true
}
log.Debug(" adding token to the fixed list")
postProcessedTokens = append(postProcessedTokens, currentToken)
if index != len(tokens)-1 &&
((currentToken.TokenType == openCollect && tokens[index+1].TokenType == closeCollect) ||
(currentToken.TokenType == openCollectObject && tokens[index+1].TokenType == closeCollectObject)) {
log.Debug(" adding empty")
op := &Operation{OperationType: emptyOpType, StringValue: "EMPTY"}
postProcessedTokens = append(postProcessedTokens, &token{TokenType: operationToken, Operation: op})
}
if index != len(tokens)-1 && currentToken.CheckForPostTraverse &&
tokens[index+1].TokenType == operationToken &&
tokens[index+1].Operation.OperationType == traversePathOpType {
log.Debug(" adding pipe because the next thing is traverse")
op := &Operation{OperationType: shortPipeOpType, Value: "PIPE"}
postProcessedTokens = append(postProcessedTokens, &token{TokenType: operationToken, Operation: op})
}
if index != len(tokens)-1 && currentToken.CheckForPostTraverse &&
tokens[index+1].TokenType == openCollect {
// if tokens[index].TokenType == closeCollect {
// log.Debug(" adding pipe because next is opencollect")
// op := &Operation{OperationType: shortPipeOpType, Value: "PIPE"}
// postProcessedTokens = append(postProcessedTokens, &token{TokenType: operationToken, Operation: op})
// }
log.Debug(" adding traverArray because next is opencollect")
op := &Operation{OperationType: traverseArrayOpType}
postProcessedTokens = append(postProcessedTokens, &token{TokenType: operationToken, Operation: op})
}
return postProcessedTokens, skipNextToken
}