wikilinks-parser/lexer.go

package wikilink

import (
	"fmt"
	"strings"
	"unicode/utf8"
)

const (
	eof rune = -1
)

const (
	openLink  = "[["
	closeLink = "]]"
	alias     = "|"
	blockRef  = "#^"
)

func lex(name, input string) *lexer {
	l := &lexer{
		name:  name,
		input: input,
		state: lexText,
		items: make(chan item, 2),
	}
	go l.run()
	return l
}

func (l *lexer) nextItem() item {
	for {
		select {
		case item := <-l.items:
			return item
		default:
			l.state = l.state(l)
		}
	}
}

func (l *lexer) ignore() {
	l.start = l.pos
}

func (l *lexer) backup() {
	l.pos -= l.width
}

type lexer struct {
	name, input       string
	start, pos, width int
	state             stateFn
	items             chan item
}

func (l *lexer) peek() rune {
	r := l.next()
	l.backup()
	return r
}

func (l *lexer) accept(valid string) bool {
	if strings.IndexRune(valid, l.next()) >= 0 {
		return true
	}
	l.backup()
	return false
}

func (l *lexer) acceptRun(valid string) {
	for strings.IndexRune(valid, l.next()) >= 0 {
	}
	l.backup()
}

func (l *lexer) emit(t itemType) {
	l.items <- item{t, l.input[l.start:l.pos]}
	l.start = l.pos
}

func (l *lexer) errorf(format string, args ...interface{}) stateFn {
	l.items <- item{
		itemError,
		fmt.Sprintf(format, args...),
	}
	return nil
}

func lexFragment(l *lexer) stateFn {
	return l.errorf("lexFragment not implemented")
}

func lexAlias(l *lexer) stateFn {
	return l.errorf("lexAlias not implemented")
}

func lexInsideLink(l *lexer) stateFn {
	for {
		if strings.HasPrefix(l.input[l.pos:], closeLink) {
			return lexCloseLink
		}

		r := l.next()

		switch {
		case r == eof:
		case r == '\n':
			return l.errorf("unclosed link")
		case r == '#':
			return lexFragment
		case r == '|':
			return lexAlias
		}
	}
}

func lexOpenLink(l *lexer) stateFn {
	l.pos += len(openLink)
	l.emit(itemOpenLink)
	return lexInsideLink
}

func lexCloseLink(l *lexer) stateFn {
	l.pos += len(closeLink)
	l.emit(itemCloseLink)
	return lexText
}

func lexText(l *lexer) stateFn {
	for {
		if strings.HasPrefix(l.input[l.pos:], openLink) {
			if l.pos > l.start {
				l.emit(itemText)
			}
			return lexOpenLink
		}

		if l.next() == eof {
			break
		}
		if l.pos > l.start {
			l.emit(itemText)
		}
		l.emit(itemEOF)
		return nil
	}
	return nil
}

func (l *lexer) next() rune {
	var r rune
	if l.pos >= len(l.input) {
		l.width = 0
		return eof
	}
	r, l.width = utf8.DecodeRuneInString(l.input[l.pos:])
	l.pos += l.width
	return r
}

func (l *lexer) run() {
	for state := lexText; state != nil; {
		state = state(l)
	}
	close(l.items)
}

type stateFn func(*lexer) stateFn

type itemType int

const (
	itemError itemType = iota
	itemEOF
	itemText
	itemOpenLink
	itemCloseLink
	itemBlockRef
	itemAlias
)

type item struct {
	typ itemType
	val string
}

func (i item) String() string {
	switch i.typ {

	case itemEOF:
		return "EOF"
	case itemError:
		return i.val
	}
	if len(i.val) > 10 {
		return fmt.Sprintf("%.10q...", i.val)
	}
	return fmt.Sprintf("%q...", i.val)
}