2017-08-07 12:07:29 +02:00
|
|
|
package goparsify
|
2017-08-06 06:31:35 +02:00
|
|
|
|
|
|
|
import (
|
2017-08-06 09:28:34 +02:00
|
|
|
"bytes"
|
2017-08-06 06:31:35 +02:00
|
|
|
"fmt"
|
2017-08-06 07:43:23 +02:00
|
|
|
"strings"
|
|
|
|
"unicode/utf8"
|
2017-08-06 06:31:35 +02:00
|
|
|
)
|
|
|
|
|
2017-08-07 10:25:23 +02:00
|
|
|
type Node struct {
|
|
|
|
Token string
|
2017-08-07 13:45:12 +02:00
|
|
|
Children []Node
|
2017-08-07 10:25:23 +02:00
|
|
|
Result interface{}
|
|
|
|
}
|
|
|
|
|
2017-08-07 13:45:12 +02:00
|
|
|
type Parser func(*State) Node
|
2017-08-06 06:31:35 +02:00
|
|
|
|
|
|
|
// Parserish types are any type that can be turned into a Parser by Parsify
|
|
|
|
// These currently include *Parser and string literals.
|
|
|
|
//
|
|
|
|
// This makes recursive grammars cleaner and allows string literals to be used directly in most contexts.
|
|
|
|
// eg, matching balanced paren:
|
|
|
|
// ```go
|
|
|
|
// var group Parser
|
|
|
|
// group = And("(", Maybe(&group), ")")
|
|
|
|
// ```
|
|
|
|
// vs
|
|
|
|
// ```go
|
|
|
|
// var group ParserPtr{}
|
|
|
|
// group.P = And(Exact("("), Maybe(group.Parse), Exact(")"))
|
|
|
|
// ```
|
|
|
|
type Parserish interface{}
|
|
|
|
|
|
|
|
func Parsify(p Parserish) Parser {
|
|
|
|
switch p := p.(type) {
|
2017-08-07 13:38:46 +02:00
|
|
|
case nil:
|
|
|
|
return nil
|
2017-08-07 13:45:12 +02:00
|
|
|
case func(*State) Node:
|
2017-08-07 13:20:30 +02:00
|
|
|
return NewParser("anonymous func", p)
|
2017-08-06 06:31:35 +02:00
|
|
|
case Parser:
|
|
|
|
return p
|
|
|
|
case *Parser:
|
|
|
|
// Todo: Maybe capture this stack and on nil show it? Is there a good error library to do this?
|
2017-08-07 13:45:12 +02:00
|
|
|
return func(ptr *State) Node {
|
2017-08-06 06:31:35 +02:00
|
|
|
return (*p)(ptr)
|
|
|
|
}
|
|
|
|
case string:
|
|
|
|
return Exact(p)
|
|
|
|
default:
|
|
|
|
panic(fmt.Errorf("cant turn a `%T` into a parser", p))
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
func ParsifyAll(parsers ...Parserish) []Parser {
|
|
|
|
ret := make([]Parser, len(parsers))
|
|
|
|
for i, parser := range parsers {
|
|
|
|
ret[i] = Parsify(parser)
|
|
|
|
}
|
|
|
|
return ret
|
|
|
|
}
|
|
|
|
|
2017-08-07 13:20:30 +02:00
|
|
|
func WS() Parser {
|
2017-08-07 13:45:12 +02:00
|
|
|
return NewParser("AutoWS", func(ps *State) Node {
|
2017-08-07 13:20:30 +02:00
|
|
|
ps.WS()
|
2017-08-07 13:45:12 +02:00
|
|
|
return Node{}
|
2017-08-07 13:20:30 +02:00
|
|
|
})
|
|
|
|
}
|
|
|
|
|
2017-08-06 15:32:10 +02:00
|
|
|
func ParseString(parser Parserish, input string) (result interface{}, remaining string, err error) {
|
2017-08-06 07:43:23 +02:00
|
|
|
p := Parsify(parser)
|
2017-08-07 13:20:30 +02:00
|
|
|
ps := InputString(input)
|
|
|
|
|
2017-08-07 10:25:23 +02:00
|
|
|
ret := p(ps)
|
2017-08-07 13:20:30 +02:00
|
|
|
ps.AutoWS()
|
2017-08-06 07:43:23 +02:00
|
|
|
|
2017-08-06 15:32:10 +02:00
|
|
|
if ps.Error.Expected != "" {
|
|
|
|
return nil, ps.Get(), ps.Error
|
2017-08-06 07:43:23 +02:00
|
|
|
}
|
|
|
|
|
2017-08-07 10:25:23 +02:00
|
|
|
return ret.Result, ps.Get(), nil
|
2017-08-06 07:43:23 +02:00
|
|
|
}
|
|
|
|
|
2017-08-06 06:31:35 +02:00
|
|
|
func Exact(match string) Parser {
|
2017-08-07 13:57:12 +02:00
|
|
|
if len(match) == 1 {
|
|
|
|
matchByte := match[0]
|
|
|
|
return NewParser(match, func(ps *State) Node {
|
|
|
|
ps.AutoWS()
|
2017-08-07 14:38:34 +02:00
|
|
|
if ps.Pos >= len(ps.Input) || ps.Input[ps.Pos] != matchByte {
|
2017-08-07 13:57:12 +02:00
|
|
|
ps.ErrorHere(match)
|
|
|
|
return Node{}
|
|
|
|
}
|
|
|
|
|
|
|
|
ps.Advance(1)
|
|
|
|
|
|
|
|
return Node{Token: match}
|
|
|
|
})
|
|
|
|
}
|
|
|
|
|
2017-08-07 13:45:12 +02:00
|
|
|
return NewParser(match, func(ps *State) Node {
|
2017-08-07 13:20:30 +02:00
|
|
|
ps.AutoWS()
|
2017-08-06 15:32:10 +02:00
|
|
|
if !strings.HasPrefix(ps.Get(), match) {
|
|
|
|
ps.ErrorHere(match)
|
2017-08-07 13:45:12 +02:00
|
|
|
return Node{}
|
2017-08-06 06:31:35 +02:00
|
|
|
}
|
|
|
|
|
2017-08-06 15:32:10 +02:00
|
|
|
ps.Advance(len(match))
|
2017-08-06 07:43:23 +02:00
|
|
|
|
2017-08-07 13:45:12 +02:00
|
|
|
return Node{Token: match}
|
2017-08-07 12:07:29 +02:00
|
|
|
})
|
2017-08-06 06:31:35 +02:00
|
|
|
}
|
|
|
|
|
2017-08-06 15:32:10 +02:00
|
|
|
func parseRepetition(defaultMin, defaultMax int, repetition ...int) (min int, max int) {
|
|
|
|
min = defaultMin
|
|
|
|
max = defaultMax
|
2017-08-06 07:43:23 +02:00
|
|
|
switch len(repetition) {
|
|
|
|
case 0:
|
|
|
|
case 1:
|
|
|
|
min = repetition[0]
|
|
|
|
case 2:
|
|
|
|
min = repetition[0]
|
|
|
|
max = repetition[1]
|
|
|
|
default:
|
|
|
|
panic(fmt.Errorf("Dont know what %d repetion args mean", len(repetition)))
|
|
|
|
}
|
2017-08-06 15:32:10 +02:00
|
|
|
return min, max
|
|
|
|
}
|
2017-08-06 07:43:23 +02:00
|
|
|
|
2017-08-06 15:32:10 +02:00
|
|
|
// parseMatcher turns a string in the format a-f01234A-F into:
|
|
|
|
// - a set string of matches string(01234)
|
|
|
|
// - a set of ranges [][]rune{{'a', 'f'}, {'A', 'F'}}
|
|
|
|
func parseMatcher(matcher string) (matches string, ranges [][]rune) {
|
|
|
|
runes := []rune(matcher)
|
|
|
|
|
|
|
|
for i := 0; i < len(runes); i++ {
|
|
|
|
|
|
|
|
if i+2 < len(runes) && runes[i+1] == '-' {
|
|
|
|
start := runes[i]
|
|
|
|
end := runes[i+2]
|
|
|
|
if start <= end {
|
|
|
|
ranges = append(ranges, []rune{start, end})
|
|
|
|
} else {
|
|
|
|
ranges = append(ranges, []rune{end, start})
|
|
|
|
}
|
|
|
|
} else if i+1 < len(runes) && runes[i] == '\\' {
|
|
|
|
matches += string(runes[i+1])
|
2017-08-06 07:43:23 +02:00
|
|
|
} else {
|
2017-08-06 15:32:10 +02:00
|
|
|
matches += string(runes[i])
|
2017-08-06 06:31:35 +02:00
|
|
|
}
|
2017-08-06 15:32:10 +02:00
|
|
|
|
2017-08-06 06:31:35 +02:00
|
|
|
}
|
|
|
|
|
2017-08-06 15:32:10 +02:00
|
|
|
return matches, ranges
|
|
|
|
}
|
|
|
|
|
|
|
|
func Chars(matcher string, repetition ...int) Parser {
|
2017-08-07 12:07:29 +02:00
|
|
|
return NewParser("["+matcher+"]", charsImpl(matcher, false, repetition...))
|
2017-08-06 15:32:10 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
func NotChars(matcher string, repetition ...int) Parser {
|
2017-08-07 12:07:29 +02:00
|
|
|
return NewParser("!["+matcher+"]", charsImpl(matcher, true, repetition...))
|
2017-08-06 15:32:10 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
func charsImpl(matcher string, stopOn bool, repetition ...int) Parser {
|
|
|
|
min, max := parseRepetition(1, -1, repetition...)
|
|
|
|
matches, ranges := parseMatcher(matcher)
|
|
|
|
|
2017-08-07 13:45:12 +02:00
|
|
|
return func(ps *State) Node {
|
2017-08-07 13:20:30 +02:00
|
|
|
ps.AutoWS()
|
2017-08-06 07:43:23 +02:00
|
|
|
matched := 0
|
2017-08-06 15:32:10 +02:00
|
|
|
for ps.Pos+matched < len(ps.Input) {
|
2017-08-06 07:43:23 +02:00
|
|
|
if max != -1 && matched >= max {
|
|
|
|
break
|
|
|
|
}
|
|
|
|
|
2017-08-06 15:32:10 +02:00
|
|
|
r, w := utf8.DecodeRuneInString(ps.Input[ps.Pos+matched:])
|
2017-08-06 07:43:23 +02:00
|
|
|
|
2017-08-06 15:32:10 +02:00
|
|
|
anyMatched := strings.ContainsRune(matches, r)
|
|
|
|
if !anyMatched {
|
|
|
|
for _, rng := range ranges {
|
|
|
|
if r >= rng[0] && r <= rng[1] {
|
|
|
|
anyMatched = true
|
|
|
|
}
|
2017-08-06 07:43:23 +02:00
|
|
|
}
|
|
|
|
}
|
2017-08-06 15:32:10 +02:00
|
|
|
|
|
|
|
if anyMatched == stopOn {
|
2017-08-06 07:43:23 +02:00
|
|
|
break
|
|
|
|
}
|
|
|
|
|
|
|
|
matched += w
|
|
|
|
}
|
|
|
|
|
|
|
|
if matched < min {
|
2017-08-06 15:32:10 +02:00
|
|
|
ps.ErrorHere(matcher)
|
2017-08-07 13:45:12 +02:00
|
|
|
return Node{}
|
2017-08-06 07:43:23 +02:00
|
|
|
}
|
|
|
|
|
2017-08-06 15:32:10 +02:00
|
|
|
result := ps.Input[ps.Pos : ps.Pos+matched]
|
|
|
|
ps.Advance(matched)
|
2017-08-07 13:45:12 +02:00
|
|
|
return Node{Token: result}
|
2017-08-06 07:43:23 +02:00
|
|
|
}
|
2017-08-06 06:31:35 +02:00
|
|
|
}
|
|
|
|
|
2017-08-06 09:28:34 +02:00
|
|
|
func String(quote rune) Parser {
|
2017-08-07 13:45:12 +02:00
|
|
|
return NewParser("string", func(ps *State) Node {
|
2017-08-07 13:20:30 +02:00
|
|
|
ps.AutoWS()
|
2017-08-06 09:28:34 +02:00
|
|
|
var r rune
|
|
|
|
var w int
|
2017-08-06 15:32:10 +02:00
|
|
|
var matched int
|
|
|
|
r, matched = utf8.DecodeRuneInString(ps.Input[ps.Pos:])
|
2017-08-06 09:28:34 +02:00
|
|
|
if r != quote {
|
2017-08-06 15:32:10 +02:00
|
|
|
ps.ErrorHere("\"")
|
2017-08-07 13:45:12 +02:00
|
|
|
return Node{}
|
2017-08-06 09:28:34 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
result := &bytes.Buffer{}
|
|
|
|
|
2017-08-06 15:32:10 +02:00
|
|
|
for ps.Pos+matched < len(ps.Input) {
|
|
|
|
r, w = utf8.DecodeRuneInString(ps.Input[ps.Pos+matched:])
|
2017-08-06 09:28:34 +02:00
|
|
|
matched += w
|
|
|
|
|
|
|
|
if r == '\\' {
|
2017-08-06 15:32:10 +02:00
|
|
|
r, w = utf8.DecodeRuneInString(ps.Input[ps.Pos+matched:])
|
2017-08-06 09:28:34 +02:00
|
|
|
result.WriteRune(r)
|
|
|
|
matched += w
|
|
|
|
continue
|
|
|
|
}
|
|
|
|
|
|
|
|
if r == quote {
|
2017-08-06 15:32:10 +02:00
|
|
|
ps.Advance(matched)
|
2017-08-07 13:45:12 +02:00
|
|
|
return Node{Token: result.String()}
|
2017-08-06 09:28:34 +02:00
|
|
|
}
|
|
|
|
result.WriteRune(r)
|
|
|
|
}
|
|
|
|
|
2017-08-06 15:32:10 +02:00
|
|
|
ps.ErrorHere("\"")
|
2017-08-07 13:45:12 +02:00
|
|
|
return Node{}
|
2017-08-07 12:07:29 +02:00
|
|
|
})
|
2017-08-06 09:28:34 +02:00
|
|
|
}
|