|
|
|
@ -2,7 +2,7 @@ package io.legado.app.model.analyzeRule |
|
|
|
|
|
|
|
|
|
//通用的规则切分处理 |
|
|
|
|
|
|
|
|
|
class RuleAnalyzer(data: String) { |
|
|
|
|
class RuleAnalyzer(data: String, code:Boolean = false) { |
|
|
|
|
|
|
|
|
|
private var queue: String = data //被处理字符串 |
|
|
|
|
private var pos = 0 //处理到的位置 |
|
|
|
@ -12,11 +12,13 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
private var startX = 0 //规则的开始 |
|
|
|
|
private var step:Int = 0 //分割字符的长度 |
|
|
|
|
|
|
|
|
|
val chompBalanced = if(code) ::chompCodeBalanced else ::chompRuleBalanced //设置平衡组函数,json或JavaScript时设置成chompCodeBalanced,否则为chompRuleBalanced |
|
|
|
|
|
|
|
|
|
var elementsType = "" |
|
|
|
|
|
|
|
|
|
//当前平衡字段 |
|
|
|
|
fun currBalancedString( stepStart:Int = 1 , stepEnd:Int = 1): String { //stepStart平衡字符的起始分隔字串长度,stepEnd平衡字符的结束分隔字串长度 |
|
|
|
|
return queue.substring(start+stepStart,pos-stepEnd) //当前平衡字段 |
|
|
|
|
return queue.substring(startX+stepStart,pos-stepEnd) //当前平衡字段 |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
|
|
@ -48,14 +50,6 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
val isEmpty: Boolean |
|
|
|
|
get() = queue.length - pos == 0 //是否处理到最后 |
|
|
|
|
|
|
|
|
|
/** |
|
|
|
|
* 检索并返回首字符,但pos不变 |
|
|
|
|
* @return 首字符:若为空则为 0 号字符 |
|
|
|
|
*/ |
|
|
|
|
fun peek(): Char { //检索首字符 |
|
|
|
|
return if (isEmpty) 0.toChar() else queue[pos] |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
/** |
|
|
|
|
* 消耗剩余字串中一个字符。 |
|
|
|
|
* @return 返回剩余字串中的下个字符。 |
|
|
|
@ -191,15 +185,7 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
if (fn) depth++ else depth-- //嵌套或者闭合 |
|
|
|
|
|
|
|
|
|
}else pos++ |
|
|
|
|
// { //转义字符 |
|
|
|
|
// var next = queue[pos++] //拉出被转义字符 |
|
|
|
|
// if(commit && next == 'n') commit = false //匹配单行注释终点。当前为\,下个为n,表示换行 |
|
|
|
|
// else if (!commits && !commit && next == '\\') { |
|
|
|
|
// queue[pos++] //当前为\,下个为\,双重转义中"\\"表示转义字符本身,根据if条件"\\"字串不在注释中,则只能在字串或正则中 |
|
|
|
|
// next = queue[pos++] //拉出下个字符,因为在双重转义的字串或正则中,类似于 \\/ 这样的结构才是转义结构 |
|
|
|
|
// if(next == '\\')queue[pos++] //若为转义字符则继续拉出,因为双重转义中转义字符成对存在,即 \\\\ |
|
|
|
|
// } |
|
|
|
|
// } |
|
|
|
|
|
|
|
|
|
} while (depth > 0 || bracketsDepth >0) //拉出全部符合js语法的字段 |
|
|
|
|
|
|
|
|
|
return if(depth > 0 || bracketsDepth > 0) false else { |
|
|
|
@ -209,9 +195,9 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
/** |
|
|
|
|
* 在双重转义字串中拉出一个规则平衡组 |
|
|
|
|
* 拉出一个代码平衡组,存在转义文本,没有实体字符,通常以{}作为模块 |
|
|
|
|
*/ |
|
|
|
|
fun chompRuleBalanced(open: Char = '[', close: Char = ']',f: ((Char) ->Boolean?)? = null ): Boolean { |
|
|
|
|
fun chompCodeBalanced(open: Char, close: Char): Boolean { |
|
|
|
|
|
|
|
|
|
var pos = pos //声明临时变量记录匹配位置,匹配成功后才同步到类的pos |
|
|
|
|
|
|
|
|
@ -230,35 +216,63 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
|
|
|
|
|
if (inSingleQuote || inDoubleQuote) continue //语法单元未匹配结束,直接进入下个循环 |
|
|
|
|
|
|
|
|
|
if ( c == open )depth++ //开始嵌套一层 |
|
|
|
|
else if ( c== close) depth-- //闭合一层嵌套 |
|
|
|
|
else if(depth == 0 && f != null) { //处于默认嵌套中的非默认字符不需要平衡,仅depth为0时默认嵌套全部闭合,此字符才进行嵌套 |
|
|
|
|
val fn = f(c) ?: continue |
|
|
|
|
if (fn) otherDepth++ else otherDepth-- |
|
|
|
|
if ( c == '[' )depth++ //开始嵌套一层 |
|
|
|
|
else if ( c== ']') depth-- //闭合一层嵌套 |
|
|
|
|
else if(depth == 0 ) { |
|
|
|
|
//处于默认嵌套中的非默认字符不需要平衡,仅depth为0时默认嵌套全部闭合,此字符才进行嵌套 |
|
|
|
|
if(c == open)otherDepth++ |
|
|
|
|
else if(c == close)otherDepth-- |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
}else pos++ |
|
|
|
|
// { //转义字符 |
|
|
|
|
// var next = queue[pos++] //拉出被转义字符,匹配\/、\"、\'等 |
|
|
|
|
// if (next == ESC) { |
|
|
|
|
// queue[pos++] //当前为\,下个为\,双重转义中"\\"表示转义字符本身,根据语法特征当前字段在字串或正则中 |
|
|
|
|
// next = queue[pos++] //拉出下个字符,因为在双重转义的字串或正则中,类似于 \\/ 这样的结构才是转义结构 |
|
|
|
|
// if(next == ESC)queue[pos++] //若为转义字符则继续拉出,因为双重转义中转义字符成对存在,即 \\\\ |
|
|
|
|
// } |
|
|
|
|
// } |
|
|
|
|
|
|
|
|
|
} while (depth > 0 || otherDepth > 0) //拉出一个平衡字串 |
|
|
|
|
|
|
|
|
|
return if(depth > 0 || otherDepth > 0) false else { |
|
|
|
|
startX = this.pos //内嵌规则起始 |
|
|
|
|
this.pos = pos //同步位置 |
|
|
|
|
true |
|
|
|
|
} |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
/** |
|
|
|
|
* 拉出一个规则平衡组,经过仔细测试xpath和jsoup中,引号内转义字符无效。 |
|
|
|
|
*/ |
|
|
|
|
fun chompRuleBalanced(open: Char, close: Char): Boolean { |
|
|
|
|
|
|
|
|
|
var pos = pos //声明临时变量记录匹配位置,匹配成功后才同步到类的pos |
|
|
|
|
var depth = 0 //嵌套深度 |
|
|
|
|
var inSingleQuote = false //单引号 |
|
|
|
|
var inDoubleQuote = false //双引号 |
|
|
|
|
|
|
|
|
|
do { |
|
|
|
|
if (pos == queue.length) break |
|
|
|
|
val c = queue[pos++] |
|
|
|
|
if (c == '\'' && !inDoubleQuote) inSingleQuote = !inSingleQuote //匹配具有语法功能的单引号 |
|
|
|
|
else if (c == '"' && !inSingleQuote) inDoubleQuote = !inDoubleQuote //匹配具有语法功能的双引号 |
|
|
|
|
|
|
|
|
|
if (inSingleQuote || inDoubleQuote) continue //语法单元未匹配结束,直接进入下个循环 |
|
|
|
|
else if( c=='\\' ){ //不在引号中的转义字符才将下个字符转义 |
|
|
|
|
pos++ |
|
|
|
|
continue |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
if ( c == open )depth++ //开始嵌套一层 |
|
|
|
|
else if ( c== close) depth-- //闭合一层嵌套 |
|
|
|
|
|
|
|
|
|
} while (depth > 0 ) //拉出一个平衡字串 |
|
|
|
|
|
|
|
|
|
return if(depth > 0) false else { |
|
|
|
|
this.pos = pos //同步位置 |
|
|
|
|
true |
|
|
|
|
} |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
/** |
|
|
|
|
* 不用正则,不到最后不切片也不用中间变量存储,只在序列中标记当前查找字段的开头结尾,到返回时才切片,高效快速准确切割规则 |
|
|
|
|
* 解决jsonPath自带的"&&"和"||"与阅读的规则冲突,以及规则正则或字符串中包含"&&"、"||"、"%%"、"@"导致的冲突 |
|
|
|
|
*/ |
|
|
|
|
tailrec fun splitRule(vararg split: String): Array<String>{ //首段匹配,elementsType为空 |
|
|
|
|
tailrec fun splitRule(vararg split: String): Array<String>{ //首段匹配,elementsType为空 |
|
|
|
|
|
|
|
|
|
if(split.size == 1) { |
|
|
|
|
elementsType = split[0] //设置分割字串 |
|
|
|
@ -322,7 +336,7 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
pos = st //位置推移到筛选器处 |
|
|
|
|
val next = if (queue[pos] == '[') ']' else ')' //平衡组末尾字符 |
|
|
|
|
|
|
|
|
|
if (!chompRuleBalanced(queue[pos], next)) throw Error( |
|
|
|
|
if (!chompBalanced(queue[pos], next)) throw Error( |
|
|
|
|
queue.substring( |
|
|
|
|
0, |
|
|
|
|
start |
|
|
|
@ -334,10 +348,10 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
start = pos //设置开始查找筛选器位置的起始位置 |
|
|
|
|
|
|
|
|
|
return splitRule(* split) //递归调用首段匹配 |
|
|
|
|
} |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
@JvmName("splitRuleNext") |
|
|
|
|
private tailrec fun splitRule(): Array<String>{ //二段匹配被调用,elementsType非空(已在首段赋值),直接按elementsType查找,比首段采用的方式更快 |
|
|
|
|
@JvmName("splitRuleNext") |
|
|
|
|
private tailrec fun splitRule(): Array<String>{ //二段匹配被调用,elementsType非空(已在首段赋值),直接按elementsType查找,比首段采用的方式更快 |
|
|
|
|
|
|
|
|
|
val end = pos //记录分隔位置 |
|
|
|
|
pos = start //重回开始,启动另一种查找 |
|
|
|
@ -383,7 +397,7 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
pos = st //位置推移到筛选器处 |
|
|
|
|
val next = if (queue[pos] == '[') ']' else ')' //平衡组末尾字符 |
|
|
|
|
|
|
|
|
|
if (!chompRuleBalanced(queue[pos], next)) throw Error( |
|
|
|
|
if (!chompBalanced(queue[pos], next)) throw Error( |
|
|
|
|
queue.substring( |
|
|
|
|
0, |
|
|
|
|
start |
|
|
|
@ -399,10 +413,10 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
rule |
|
|
|
|
}else splitRule() //递归匹配 |
|
|
|
|
|
|
|
|
|
} |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/** |
|
|
|
|
/** |
|
|
|
|
* 替换内嵌规则 |
|
|
|
|
* @param inner 起始标志,如{$. 或 {{ |
|
|
|
|
* @param startStep 不属于规则部分的前置字符长度,如{$.中{不属于规则的组成部分,故startStep为1 |
|
|
|
@ -410,7 +424,7 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
* @param fr 查找到内嵌规则时,用于解析的函数 |
|
|
|
|
* |
|
|
|
|
* */ |
|
|
|
|
fun innerRule( inner:String,startStep:Int = 1,endStep:Int = 1,fr:(String)->String?): String { |
|
|
|
|
fun innerRule( inner:String,startStep:Int = 1,endStep:Int = 1,fr:(String)->String?): String { |
|
|
|
|
|
|
|
|
|
val start0 = pos //规则匹配前起点 |
|
|
|
|
|
|
|
|
@ -418,16 +432,9 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
|
|
|
|
|
while (!isEmpty && consumeTo(inner)) { //拉取成功返回true,ruleAnalyzes里的字符序列索引变量pos后移相应位置,否则返回false,且isEmpty为true |
|
|
|
|
val posPre = pos //记录上次结束位置 |
|
|
|
|
if (chompRuleBalanced {//拉出一个以[]为默认嵌套、以{}为补充嵌套的平衡字段 |
|
|
|
|
when (it) { |
|
|
|
|
'{' -> true |
|
|
|
|
'}' -> false |
|
|
|
|
else -> null |
|
|
|
|
} |
|
|
|
|
}) { |
|
|
|
|
if (chompBalanced('{', '}')) { |
|
|
|
|
val frv= fr(currBalancedString(startStep,endStep)) |
|
|
|
|
if(frv != null) { |
|
|
|
|
|
|
|
|
|
st.append(queue.substring(start,posPre)+frv) //压入内嵌规则前的内容,及内嵌规则解析得到的字符串 |
|
|
|
|
continue //获取内容成功,继续选择下个内嵌规则 |
|
|
|
|
|
|
|
|
@ -443,7 +450,7 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
st.append(remainingString()) //压入剩余字符串 |
|
|
|
|
st.toString() |
|
|
|
|
} |
|
|
|
|
} |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
// /** |
|
|
|
|
// * 匹配并返回标签中的属性键字串(字母、数字、-、_、:) |
|
|
|
@ -478,7 +485,7 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
// return query |
|
|
|
|
// } |
|
|
|
|
|
|
|
|
|
companion object { |
|
|
|
|
companion object { |
|
|
|
|
/** |
|
|
|
|
* 转义字符 |
|
|
|
|
*/ |
|
|
|
@ -557,5 +564,4 @@ class RuleAnalyzer(data: String) { |
|
|
|
|
*/ |
|
|
|
|
val splitListDefault = arrayOf("-",".","!","@","@@") |
|
|
|
|
|
|
|
|
|
} |
|
|
|
|
} |
|
|
|
|
}} |