批处理之家 - Powered by Discuz! Board

import java.io.BufferedReader;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.io.OutputStreamWriter;
import java.io.PrintWriter;
import java.net.URL;
import java.net.URLConnection;
import java.util.regex.*;
public class Test {
public static void main(String[] args) {
try {
URL url = new URL("http://127.0.0.1/html/10.248.1.68(80).html");
URLConnection conn = url.openConnection();
conn.setDoOutput(true);
InputStream in = null;
in = url.openStream();
String content = pipe(in,"utf-8");
//System.out.println(content);
String Regex=">IP地址</th><td>(.*?)</td></tr>(.*?)>安全评估分</th><td>(.*?)</td></tr>(.*?)src=\"images/blank.gif\"(.*?)</td><td colspan='1' rowspan='1' width='40px'>(.*?)</td><td colspan='1' rowspan='1' width='40px'>(.*?)</td><td colspan='1' rowspan='1' width='30px'>(.*?)</th><td colspan='1' rowspan='1' width='150px'>(.*?)</td><td>(.*?)";
String html=content;
Pattern p1=Pattern.compile(Regex);
Matcher m=p1.matcher(html);
while(m.find())
{
//System.out.println(m.group());
System.out.println("IP地址:"+m.group(1)+"\r\n检查分:"+m.group(3)+"\r\n基线项:"+m.group(5)+"\r\n风险等级:"+m.group(7)+"\r\n检查点:"+m.group(9));
}
} catch (Exception e) {
e.printStackTrace();
}
}
static String pipe(InputStream in,String charset) throws IOException {
StringBuffer s = new StringBuffer();
if(charset==null||"".equals(charset)){
charset="utf-8";
}
String rLine = null;
BufferedReader bReader = new BufferedReader(new InputStreamReader(in,charset));
PrintWriter pw = null;
FileOutputStream fo = new FileOutputStream("test.html");
OutputStreamWriter writer = new OutputStreamWriter(fo, "utf-8");
pw = new PrintWriter(writer);
while ( (rLine = bReader.readLine()) != null) {
String tmp_rLine = rLine;
int str_len = tmp_rLine.length();
if (str_len > 0) {
s.append(tmp_rLine);
pw.println(tmp_rLine);
pw.flush();
}
tmp_rLine = null;
}
in.close();
pw.close();
return s.toString();
}
}

复制代码

;-----------------------------------------
/*
[config]
ahk_version=Unicode
icon=
bin=
*/

;@Ahk2Exe-SetVersion 1.0.0.0
;-----------------------------------------
#NoEnv
#NoTrayIcon
#SingleInstance Force
SetWorkingDir %A_ScriptDir%
SetBatchLines -1
ListLines Off

gui_title := "提取评估报告 v1.00"

; 命令行参数
;
If 1 in /?,-h,--help
{
MsgBox, % "用法: " A_ScriptName " [-dir html目录] [-o 输出文件]"
ExitApp
}

arg := ParamToObj()
arg.o := (arg.o = "") ? "评估报告_最终结果.xls" : arg.o

; 开始提取
;
FinalResult := "序号 IP地址操作系统检查分数主机是否达标基线项风险等级检查点实际值标准值配置方法结果`r`n"
global indexCount := 0

SetWorkingDir, % arg.dir
Loop, *.html
{
ExtractInfo(A_LoopFileFullPath, result)
FinalResult .= result
}

; 保存结果
FileOpen(arg.o, "w").Write(FinalResult)
MsgBox, 处理完毕！
Return

; ==============================================================================================
ExtractInfo(HtmlFile, ByRef ResultTable) {
ResultTable := ""

; 以 UTF-8 编码读取 html 文件内容
FileRead, data, % "*P65001 " HtmlFile

; 提取 IP地址、操作系统、检查分数
RegExMatch(data, "i)IP地址</th><td>(.*?)<.*?配置模板</th><td>(\w+).*?安全评估分</th><td>(.*?)<", m)
devInfo := m1 A_Tab m2 A_Tab m3 A_Tab ( m3>=95 ? "达标" : "不达标" )

; 将搜索范围缩小至“2.检查结果”
RegExMatch(data, "2\.检查结果</div>.*?3\.辅助信息检查结果</div>", data)

; 取得所有红色的“基线项”的出现位置
posList  := [], startPos := 1
s := "<tr class='unaccord'><td colspan='1' rowspan='1' width='100px'>"
While, pos := InStr(data, s,, startPos)
startPos := pos + StrLen(s), posList.Insert(startPos)

; 取得每个红色“基线项”下面的红色信息
For i, startPos in posList
{
endPos  := (i = posList.MaxIndex()) ? StrLen(data) : posList[i+1]
subData := SubStr(data, startPos, endPos-startPos)

; 获取“基线项”和“风险等级”
RegExMatch(subData, "i)/blank\.gif.*?<td>(.*?)</td><.*?>(.*?)</td>", m)
baseInfo := m1 A_Tab m2

; 获取“检查点”...
_pos := 1, _m := ""
_regEx := "i)<tr class='unaccord'><td>(.*?)</td><td>(.*?)</td><td>.*?<td>(.*?)</td><td>(.*?)</td>"
While, _pos := RegExMatch( subData, _regEx, _m, _pos+StrLen(_m) ) {
baseInfo_child := _m1 A_Tab _m2 A_Tab _m3 A_Tab _m4
indexCount ++
ResultTable .= indexCount A_Tab devInfo A_Tab baseInfo A_Tab baseInfo_child A_Tab "不合规" "`r`n"
}
}

ResultTable := (ResultTable = "") ? (devInfo "`r`n") : ResultTable
}

ParamToObj() {
global 0
obj := {}

Loop, %0%
{
param := %A_Index%

If RegExMatch(param, "^(/|-)+(\w+)(\W(.+))?$", match) {
obj[match2] := match4
key       := match2
}
Else If (key != "")
obj[key] := param, key := ""
}

Return obj
}