featurebase/pql/parser.go
2017-04-28 14:22:03 -05:00

319 lines
7.5 KiB
Go

// Copyright 2017 Pilosa Corp.
//
// 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 pql
import (
"fmt"
"io"
"strconv"
"strings"
)
// TimeFormat is the go-style time format used to parse string dates.
const TimeFormat = "2006-01-02T15:04"
// Parser represents a parser for the PQL language.
type Parser struct {
scanner *bufScanner
}
// NewParser returns a new instance of Parser.
func NewParser(r io.Reader) *Parser {
return &Parser{
scanner: newBufScanner(r),
}
}
// ParseString parses s into a query.
func ParseString(s string) (*Query, error) {
return NewParser(strings.NewReader(s)).Parse()
}
// Parse parses the next node in the query.
func (p *Parser) Parse() (*Query, error) {
q := &Query{}
for {
call, err := p.parseCall()
if err == io.EOF {
break
} else if err != nil {
return nil, err
}
q.Calls = append(q.Calls, call)
}
// Require at least one call.
if len(q.Calls) == 0 {
return nil, io.ErrUnexpectedEOF
}
return q, nil
}
// parseCall parses the next function call.
func (p *Parser) parseCall() (*Call, error) {
var c Call
// Read call name.
tok, pos, lit := p.scanIgnoreWhitespace()
if tok == EOF {
return nil, io.EOF
} else if tok != IDENT {
return nil, &ParseError{Message: fmt.Sprintf("expected identifier, found: %s", lit), Pos: pos}
}
c.Name = lit
// Scan opening parenthesis.
if err := p.expect(LPAREN); err != nil {
return nil, err
}
// Parse children first.
children, err := p.parseChildren()
if err != nil {
return nil, err
}
c.Children = children
// If next token is a closing paren then exit.
if tok, pos, lit := p.scanIgnoreWhitespace(); tok == RPAREN {
return &c, nil
} else if tok == IDENT {
p.unscan(1)
} else if tok != COMMA {
return nil, parseErrorf(pos, "expected comma, right paren, or identifier, found %q", lit)
}
// Parse key/value arguments.
args, err := p.parseArgs()
if err != nil {
return nil, err
}
c.Args = args
// Scan closing parenthesis.
if err := p.expect(RPAREN); err != nil {
return nil, err
}
return &c, nil
}
// parseChildren parses call children.
func (p *Parser) parseChildren() ([]*Call, error) {
var offset int
var children []*Call
for {
// Ensure next two tokens are IDENT+LPAREN.
if tok, _, _ := p.scanIgnoreWhitespace(); tok != IDENT {
p.unscanIgnoreWhitespace(1 + offset)
return children, nil
}
if tok, _, _ := p.scan(); tok != LPAREN {
p.unscanIgnoreWhitespace(2 + offset)
return children, nil
}
// Push tokens back on scanner and parse as a call.
p.unscan(2)
child, err := p.parseCall()
if err != nil {
return nil, err
}
children = append(children, child)
// Exit if closing paren.
if tok, pos, lit := p.scanIgnoreWhitespace(); tok == RPAREN {
p.unscan(1)
return children, nil
} else if tok != COMMA {
return nil, parseErrorf(pos, "expected comma or right paren, found %q", lit)
}
// Make sure comma is unscanned.
offset = 1
}
}
// parseArgs parses key/value arguments.
func (p *Parser) parseArgs() (map[string]interface{}, error) {
args := make(map[string]interface{})
for {
// Parse key.
tok, pos, lit := p.scanIgnoreWhitespace()
if tok == RPAREN {
p.unscan(1)
return args, nil
} else if tok != IDENT {
return nil, parseErrorf(pos, "expected argument key, found %q", lit)
}
key := lit
// Expect '=' next.
if tok, pos, lit := p.scanIgnoreWhitespace(); tok != EQ {
return nil, parseErrorf(pos, "expected equals sign, found %q", lit)
}
// Parse value.
var value interface{}
tok, pos, lit = p.scanIgnoreWhitespace()
switch tok {
case IDENT:
if lit == "true" {
value = true
} else if lit == "false" {
value = false
} else if lit == "null" {
value = nil
} else {
value = lit
}
case STRING:
value = lit
case INTEGER:
v, err := strconv.ParseInt(lit, 10, 64)
if err != nil {
return nil, err
}
value = v
case FLOAT:
v, err := strconv.ParseFloat(lit, 64)
if err != nil {
return nil, err
}
value = v
case LBRACK:
v, err := p.parseList()
if err != nil {
return nil, err
}
value = v
default:
return nil, parseErrorf(pos, "invalid argument value: %q", lit)
}
// Ensure key doesn't already exist.
if _, ok := args[key]; ok {
return nil, parseErrorf(pos, "argument key already used: %s", key)
}
// Add key/value pair to arguments.
args[key] = value
// Exit if closing paren.
if tok, pos, lit := p.scanIgnoreWhitespace(); tok == RPAREN {
p.unscan(1)
return args, nil
} else if tok != COMMA {
return nil, parseErrorf(pos, "expected comma or right paren, found %q", lit)
}
}
}
// parseList parses a list of primitives. This is used by the TopN() filters.
func (p *Parser) parseList() ([]interface{}, error) {
var values []interface{}
for {
// Read next value.
tok, pos, lit := p.scanIgnoreWhitespace()
switch tok {
case IDENT:
if lit == "true" {
values = append(values, true)
} else if lit == "false" {
values = append(values, false)
} else {
values = append(values, lit)
}
case STRING:
values = append(values, lit)
case INTEGER:
v, err := strconv.ParseInt(lit, 10, 64)
if err != nil {
return nil, err
}
values = append(values, v)
default:
return nil, parseErrorf(pos, "invalid list value: %q", lit)
}
// Expect a comma or closing bracket next.
if tok, pos, lit := p.scanIgnoreWhitespace(); tok == RBRACK {
break
} else if tok != COMMA {
return nil, parseErrorf(pos, "expected comma, found %q", lit)
}
}
return values, nil
}
// scan returns the next token from the scanner.
func (p *Parser) scan() (tok Token, pos Pos, lit string) { return p.scanner.Scan() }
// scanIgnoreWhitespace returns the next non-whitespace token from the scanner.
func (p *Parser) scanIgnoreWhitespace() (tok Token, pos Pos, lit string) {
tok, pos, lit = p.scan()
if tok == WS {
tok, pos, lit = p.scan()
}
return
}
// unscan returns the last n tokens back to the scanner.
func (p *Parser) unscan(n int) {
for i := 0; i < n; i++ {
p.scanner.unscan()
}
}
// unscanIgnoreWhitespace returns the last n non-WS tokens back to the scanner.
func (p *Parser) unscanIgnoreWhitespace(n int) {
for i := 0; i < n; {
p.scanner.unscan()
if tok, _, _ := p.scanner.curr(); tok != WS {
i++
}
}
}
// expect returns an error if the next token is not exp.
func (p *Parser) expect(exp Token) error {
if tok, pos, lit := p.scan(); tok != exp {
return parseErrorf(pos, "expected %s, found %q", exp.String(), lit)
}
return nil
}
// pos returns the current position.
func (p *Parser) pos() Pos { return p.scanner.pos() }
// ParseError represents an error that occurred while parsing a PQL query.
type ParseError struct {
Message string
Pos Pos
}
// Error returns a string representation of e.
func (e *ParseError) Error() string {
return fmt.Sprintf("%s occurred at line %d, char %d", e.Message, e.Pos.Line+1, e.Pos.Char+1)
}
// parseErrorf returns a formatted parse error.
func parseErrorf(pos Pos, format string, args ...interface{}) *ParseError {
return &ParseError{
Message: fmt.Sprintf(format, args...),
Pos: pos,
}
}