mirror of
https://shylinux.com/x/ContextOS
synced 2025-04-26 17:24:05 +08:00
441 lines
10 KiB
Go
441 lines
10 KiB
Go
package lex
|
|
|
|
import (
|
|
"contexts/ctx"
|
|
"fmt"
|
|
"strconv"
|
|
"strings"
|
|
)
|
|
|
|
type Seed struct {
|
|
page int
|
|
hash int
|
|
word string
|
|
}
|
|
type State struct {
|
|
star bool
|
|
next int
|
|
hash int
|
|
}
|
|
type Point struct {
|
|
s int
|
|
c byte
|
|
}
|
|
type LEX struct {
|
|
seed []*Seed
|
|
page map[string]int
|
|
hash map[string]int
|
|
|
|
mat []map[byte]*State
|
|
state map[State]*State
|
|
char map[byte][]byte
|
|
|
|
*ctx.Message
|
|
*ctx.Context
|
|
}
|
|
|
|
func (lex *LEX) index(hash string, h string) int {
|
|
which := lex.page
|
|
if hash == "nhash" {
|
|
which = lex.hash
|
|
}
|
|
|
|
if x, e := strconv.Atoi(h); e == nil {
|
|
lex.Assert(hash != "npage" || x < lex.Capi("npage"))
|
|
return x
|
|
}
|
|
|
|
if x, ok := which[h]; ok {
|
|
return x
|
|
}
|
|
|
|
which[h] = lex.Capi(hash, 1)
|
|
lex.Assert(hash != "npage" || lex.Capi("npage") < lex.Capi("nlang"))
|
|
return which[h]
|
|
}
|
|
func (lex *LEX) charset(c byte) []byte {
|
|
if cs, ok := lex.char[c]; ok {
|
|
return cs
|
|
}
|
|
return []byte{c}
|
|
}
|
|
func (lex *LEX) train(page int, hash int, seed []byte) int {
|
|
|
|
ss := []int{page}
|
|
cn := make([]bool, lex.Capi("ncell"))
|
|
cc := make([]byte, 0, lex.Capi("ncell"))
|
|
sn := make([]bool, lex.Capi("nline"))
|
|
|
|
points := []*Point{}
|
|
|
|
for p := 0; p < len(seed); p++ {
|
|
|
|
switch seed[p] {
|
|
case '[':
|
|
set := true
|
|
if p++; seed[p] == '^' {
|
|
set, p = false, p+1
|
|
}
|
|
|
|
for ; seed[p] != ']'; p++ {
|
|
if seed[p] == '\\' {
|
|
p++
|
|
for _, c := range lex.charset(seed[p]) {
|
|
cn[c] = true
|
|
}
|
|
continue
|
|
}
|
|
|
|
if seed[p+1] == '-' {
|
|
begin, end := seed[p], seed[p+2]
|
|
if begin > end {
|
|
begin, end = end, begin
|
|
}
|
|
for c := begin; c <= end; c++ {
|
|
cn[c] = true
|
|
}
|
|
p += 2
|
|
continue
|
|
}
|
|
|
|
cn[seed[p]] = true
|
|
}
|
|
|
|
for c := 0; c < len(cn); c++ {
|
|
if (set && cn[c]) || (!set && !cn[c]) {
|
|
cc = append(cc, byte(c))
|
|
}
|
|
cn[c] = false
|
|
}
|
|
|
|
case '.':
|
|
for c := 0; c < len(cn); c++ {
|
|
cc = append(cc, byte(c))
|
|
}
|
|
|
|
case '\\':
|
|
p++
|
|
for _, c := range lex.charset(seed[p]) {
|
|
cc = append(cc, c)
|
|
}
|
|
default:
|
|
cc = append(cc, seed[p])
|
|
}
|
|
|
|
lex.Log("debug", "page: \033[31m%d %v\033[0m", len(ss), ss)
|
|
lex.Log("debug", "cell: \033[32m%d %v\033[0m", len(cc), cc)
|
|
|
|
flag := '\000'
|
|
if p+1 < len(seed) {
|
|
flag = rune(seed[p+1])
|
|
switch flag {
|
|
case '+', '*', '?':
|
|
p++
|
|
}
|
|
}
|
|
|
|
for _, s := range ss {
|
|
line := 0
|
|
for _, c := range cc {
|
|
|
|
state := &State{}
|
|
if lex.mat[s][c] != nil {
|
|
*state = *lex.mat[s][c]
|
|
} else {
|
|
lex.Capi("nnode", 1)
|
|
}
|
|
lex.Log("debug", "GET(%d,%d): %v", s, c, state)
|
|
|
|
switch flag {
|
|
case '+':
|
|
state.star = true
|
|
case '*':
|
|
state.star = true
|
|
sn[s] = true
|
|
case '?':
|
|
sn[s] = true
|
|
}
|
|
|
|
if state.next == 0 {
|
|
if line == 0 || !lex.Confs("compact") {
|
|
lex.mat = append(lex.mat, make(map[byte]*State))
|
|
line = lex.Capi("nline", 1) - 1
|
|
sn = append(sn, false)
|
|
}
|
|
state.next = line
|
|
}
|
|
sn[state.next] = true
|
|
|
|
lex.mat[s][c] = state
|
|
points = append(points, &Point{s, c})
|
|
lex.Log("debug", "SET(%d,%d): %v(%s,%s)", s, c, state, lex.Cap("nnode"), lex.Cap("nreal"))
|
|
}
|
|
}
|
|
|
|
cc, ss = cc[:0], ss[:0]
|
|
for s, b := range sn {
|
|
if sn[s] = false; b {
|
|
ss = append(ss, s)
|
|
}
|
|
}
|
|
}
|
|
|
|
for _, s := range ss {
|
|
if s < lex.Capi("nlang") || s >= len(lex.mat) {
|
|
continue
|
|
}
|
|
|
|
if len(lex.mat[s]) == 0 {
|
|
lex.Log("debug", "DEL: %d-%d", lex.Capi("nline")-1, lex.Capi("nline", 0, s))
|
|
lex.mat = lex.mat[:s]
|
|
}
|
|
}
|
|
|
|
for _, s := range ss {
|
|
for _, p := range points {
|
|
state := &State{}
|
|
*state = *lex.mat[p.s][p.c]
|
|
|
|
if state.next == s {
|
|
lex.Log("debug", "GET(%d, %d): %v", p.s, p.c, state)
|
|
if state.hash = hash; state.next >= len(lex.mat) {
|
|
state.next = 0
|
|
}
|
|
lex.mat[p.s][p.c] = state
|
|
lex.Log("debug", "SET(%d, %d): %v", p.s, p.c, state)
|
|
}
|
|
|
|
if x, ok := lex.state[*state]; !ok {
|
|
lex.state[*state] = lex.mat[p.s][p.c]
|
|
lex.Capi("nreal", 1)
|
|
} else {
|
|
lex.mat[p.s][p.c] = x
|
|
}
|
|
}
|
|
}
|
|
|
|
return hash
|
|
}
|
|
func (lex *LEX) parse(m *ctx.Message, page int, line []byte) (hash int, rest []byte, word []byte) {
|
|
|
|
pos := 0
|
|
for star, s := 0, page; s != 0 && pos < len(line); pos++ {
|
|
|
|
c := line[pos]
|
|
if c == '\\' && pos < len(line)-1 { //跳过转义
|
|
pos++
|
|
c = lex.charset(line[pos])[0]
|
|
}
|
|
if c > 127 { //跳过中文
|
|
word = append(word, c)
|
|
continue
|
|
}
|
|
|
|
state := lex.mat[s][c]
|
|
lex.Log("debug", "(%d,%d): %v", s, c, state)
|
|
if state == nil {
|
|
s, star, pos = star, 0, pos-1
|
|
continue
|
|
}
|
|
|
|
word = append(word, c)
|
|
|
|
if state.star {
|
|
star = s
|
|
} else if x, ok := lex.mat[star][c]; !ok || !x.star {
|
|
star = 0
|
|
}
|
|
|
|
if s, hash = state.next, state.hash; s == 0 {
|
|
s, star = star, 0
|
|
}
|
|
}
|
|
|
|
if pos == len(line) {
|
|
// hash, pos, word = -1, 0, word[:0]
|
|
} else if hash == 0 {
|
|
pos, word = 0, word[:0]
|
|
}
|
|
rest = line[pos:]
|
|
return
|
|
}
|
|
func (lex *LEX) Spawn(m *ctx.Message, c *ctx.Context, arg ...string) ctx.Server {
|
|
lex.Message = m
|
|
c.Caches = map[string]*ctx.Cache{}
|
|
c.Configs = map[string]*ctx.Config{}
|
|
|
|
s := new(LEX)
|
|
s.Context = c
|
|
return s
|
|
}
|
|
func (lex *LEX) Begin(m *ctx.Message, arg ...string) ctx.Server {
|
|
lex.Message = m
|
|
|
|
lex.Caches["ncell"] = &ctx.Cache{Name: "字符上限", Value: "128", Help: "字符集合的最大数量"}
|
|
lex.Caches["nlang"] = &ctx.Cache{Name: "词法上限", Value: "64", Help: "词法集合的最大数量"}
|
|
|
|
lex.Caches["nseed"] = &ctx.Cache{Name: "种子数量", Value: "0", Help: "词法模板的数量"}
|
|
lex.Caches["npage"] = &ctx.Cache{Name: "集合数量", Value: "0", Help: "词法集合的数量"}
|
|
lex.Caches["nhash"] = &ctx.Cache{Name: "类型数量", Value: "0", Help: "单词类型的数量"}
|
|
|
|
lex.Caches["nline"] = &ctx.Cache{Name: "状态数量", Value: "64", Help: "状态机状态的数量"}
|
|
lex.Caches["nnode"] = &ctx.Cache{Name: "节点数量", Value: "0", Help: "状态机连接的逻辑数量"}
|
|
lex.Caches["nreal"] = &ctx.Cache{Name: "实点数量", Value: "0", Help: "状态机连接的存储数量"}
|
|
|
|
lex.Configs["compact"] = &ctx.Config{Name: "紧凑模式", Value: "true", Help: "词法状态的共用"}
|
|
|
|
if len(arg) > 0 {
|
|
if _, e := strconv.Atoi(arg[0]); lex.Assert(e) {
|
|
lex.Cap("nlang", arg[0])
|
|
lex.Cap("nline", arg[0])
|
|
}
|
|
}
|
|
|
|
lex.page = map[string]int{"nil": 0}
|
|
lex.hash = map[string]int{"nil": 0}
|
|
|
|
lex.mat = make([]map[byte]*State, lex.Capi("nlang"))
|
|
lex.state = make(map[State]*State)
|
|
|
|
lex.char = map[byte][]byte{
|
|
't': []byte{'\t'},
|
|
'n': []byte{'\n'},
|
|
'b': []byte{'\t', ' '},
|
|
's': []byte{'\t', ' ', '\n'},
|
|
'd': []byte{'0', '1', '2', '3', '4', '5', '6', '7', '8', '9'},
|
|
'x': []byte{'0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'a', 'b', 'c', 'd', 'e', 'f', 'A', 'B', 'C', 'D', 'E', 'F'},
|
|
}
|
|
|
|
return lex
|
|
}
|
|
func (lex *LEX) Start(m *ctx.Message, arg ...string) bool {
|
|
lex.Message = m
|
|
return false
|
|
}
|
|
func (lex *LEX) Close(m *ctx.Message, arg ...string) bool {
|
|
switch lex.Context {
|
|
case m.Target():
|
|
case m.Source():
|
|
}
|
|
return true
|
|
}
|
|
|
|
var Index = &ctx.Context{Name: "lex", Help: "词法中心",
|
|
Caches: map[string]*ctx.Cache{
|
|
"nmat": &ctx.Cache{Name: "nmat", Value: "0", Help: "nmat"},
|
|
},
|
|
Configs: map[string]*ctx.Config{},
|
|
Commands: map[string]*ctx.Command{
|
|
"spawn": &ctx.Command{Name: "spawn", Help: "添加词法规则", Hand: func(m *ctx.Message, c *ctx.Context, key string, arg ...string) {
|
|
if _, ok := m.Target().Server.(*LEX); m.Assert(ok) {
|
|
m.Start(fmt.Sprintf("matrix%d", m.Capi("nmat", 1)), "matrix")
|
|
}
|
|
}},
|
|
"train": &ctx.Command{Name: "train seed [hash [page]", Help: "添加词法规则", Hand: func(m *ctx.Message, c *ctx.Context, key string, arg ...string) {
|
|
if lex, ok := m.Target().Server.(*LEX); m.Assert(ok) {
|
|
page, hash := 1, 1
|
|
if len(arg) > 2 {
|
|
page = lex.index("npage", arg[2])
|
|
m.Assert(page < m.Capi("nlang"), "词法集合过多")
|
|
}
|
|
if len(arg) > 1 {
|
|
hash = lex.index("nhash", arg[1])
|
|
}
|
|
|
|
if lex.mat[page] == nil {
|
|
lex.mat[page] = map[byte]*State{}
|
|
}
|
|
|
|
m.Result(0, lex.train(page, hash, []byte(arg[0])))
|
|
lex.seed = append(lex.seed, &Seed{page, hash, arg[0]})
|
|
lex.Cap("stream", fmt.Sprintf("%d,%s,%s", lex.Capi("nseed", 1), lex.Cap("npage"), lex.Cap("nhash")))
|
|
}
|
|
}},
|
|
"parse": &ctx.Command{Name: "parse line [page]", Help: "解析单词", Hand: func(m *ctx.Message, c *ctx.Context, key string, arg ...string) {
|
|
if lex, ok := m.Target().Server.(*LEX); m.Assert(ok) {
|
|
page := 1
|
|
if len(arg) > 1 {
|
|
page = lex.index("npage", arg[1])
|
|
}
|
|
|
|
hash, rest, word := lex.parse(m, page, []byte(arg[0]))
|
|
m.Result(0, hash, string(rest), string(word))
|
|
}
|
|
}},
|
|
"show": &ctx.Command{Name: "show info", Help: "查看信息", Hand: func(m *ctx.Message, c *ctx.Context, key string, arg ...string) {
|
|
if lex, ok := m.Target().Server.(*LEX); m.Assert(ok) {
|
|
switch arg[0] {
|
|
case "seed":
|
|
for _, v := range lex.seed {
|
|
m.Add("append", "page", fmt.Sprintf("%d", v.page))
|
|
m.Add("append", "hash", fmt.Sprintf("%d", v.hash))
|
|
m.Add("append", "word", fmt.Sprintf("%s", strings.Replace(strings.Replace(v.word, "\n", "\\n", -1), "\t", "\\t", -1)))
|
|
}
|
|
m.Table()
|
|
case "page":
|
|
for k, v := range lex.page {
|
|
m.Add("append", "page", k)
|
|
m.Add("append", "code", fmt.Sprintf("%d", v))
|
|
}
|
|
m.Sort("code", "int").Table()
|
|
case "hash":
|
|
for k, v := range lex.hash {
|
|
m.Add("append", "hash", k)
|
|
m.Add("append", "code", fmt.Sprintf("%d", v))
|
|
}
|
|
m.Table()
|
|
case "mat":
|
|
for _, v := range lex.mat {
|
|
for j := byte(0); j < byte(m.Capi("ncell")); j++ {
|
|
s := v[j]
|
|
if s == nil {
|
|
m.Add("append", fmt.Sprintf("%c", j), "")
|
|
} else {
|
|
// m.Add("append", fmt.Sprintf("%c", j), fmt.Sprintf("(%t,%d,%d)", s.star, s.next, s.hash))
|
|
star := 0
|
|
if s.star {
|
|
star = 1
|
|
}
|
|
m.Add("append", fmt.Sprintf("%c", j), fmt.Sprintf("%d,%d,%d", star, s.next, s.hash))
|
|
}
|
|
}
|
|
}
|
|
|
|
ncol := len(m.Meta["append"])
|
|
nrow := len(m.Meta[m.Meta["append"][0]])
|
|
for i := 0; i < ncol-1; i++ {
|
|
for j := i + 1; j < ncol; j++ {
|
|
same := true
|
|
for n := 0; n < nrow; n++ {
|
|
if m.Meta[m.Meta["append"][i]][n] != m.Meta[m.Meta["append"][j]][n] {
|
|
same = false
|
|
break
|
|
}
|
|
}
|
|
|
|
if same {
|
|
key = m.Meta["append"][i] + m.Meta["append"][j]
|
|
m.Meta[key] = m.Meta[m.Meta["append"][i]]
|
|
m.Meta["append"][i] = key
|
|
for k := j; k < ncol-1; k++ {
|
|
m.Meta["append"][k] = m.Meta["append"][k+1]
|
|
}
|
|
ncol--
|
|
j--
|
|
}
|
|
}
|
|
}
|
|
m.Meta["append"] = m.Meta["append"][:ncol]
|
|
m.Table()
|
|
}
|
|
}
|
|
}},
|
|
},
|
|
}
|
|
|
|
func init() {
|
|
lex := &LEX{}
|
|
lex.Context = Index
|
|
ctx.Index.Register(Index, lex)
|
|
}
|