[文本处理] 怎么用bat提取html文件中pre标签内的部分

Set fso = CreateObject("scripting.filesystemobject")
Set ws = CreateObject("WScript.shell")
vbstr = Replace(fso.OpenTextFile("temp.html", 1, 1).ReadAll(), "<pre>", "<pre id=text>")
fso.OpenTextFile("temp.html", 2, 1).Write vbstr
GetStr ws.CurrentDirectory & "\temp.html"
Set fso = Nothing
Set ws = nothing
MsgBox "ok"

Function GetStr(pathfile)
  Set oDOM = GetObject(pathfile, "htmlfile")
  Do Until oDOM.readyState="complete":WScript.Sleep 200:Loop
  txt = oDOM.getElementByid("text").innertext
  fso.OpenTextFile("temp.txt", 2, 1).Write txt
  Set oDOM = Nothing
End Function
复制代码

@echo off
for /f "tokens=1,2 delims=:" %%a in ('findstr /n "<pre>" *.html') do call>>"%%~na.txt" :pre %%a %%b
pause&exit
:pre
for /f "skip=%2 delims=" %%a in (%~s1) do if "%%a"=="</pre>" (exit /b) else echo %%a
复制代码

Set fso = CreateObject("scripting.filesystemobject")
Set ws = CreateObject("WScript.shell")
For Each file In fso.GetFolder(ws.CurrentDirectory & "\").files
  If InStr (LCase(file.Name), ".htm") Then GetStr file
next
Set fso = Nothing
Set ws = Nothing
MsgBox "ok"

Function GetStr(pathfile)
  Set oDOM = GetObject(pathfile, "htmlfile")
  Do Until oDOM.readyState="complete":WScript.Sleep 200:Loop
  For Each vbstr In oDOM.getElementsBytagname("pre")
    txt = txt & vbstr.innertext & vbCrLf & vbCrLf
  Next
  fso.OpenTextFile("temp.txt", 8, 1).Write txt & vbCrLf & vbCrLf
  Set oDOM = Nothing
End Function
复制代码

@echo off&setlocal enabledelayedexpansion
for %%a in (*.html) do (
for /f "delims=" %%b in (%%a) do (
	if "%%a" neq "!last!" (
		set /a "n=1","t=0"
		set last=%%a
		cd.>"%%~na.txt"
	)
	set tmp=$%%b
	(if !t!==1 for /f "tokens=1* delims=$" %%b in ("!tmp:</pre>=$!") do echo;%%b
	if "!tmp:<pre>=!!tmp:</pre>=!" neq "!tmp!!tmp!" (
		set /a "t=^!t"
		if "!tmp:<pre>!" neq "!tmp!" (
			if !t!==1 echo;!tmp:*^<pre^>=!
		)
	))>>"%%~na.txt"
))
pause
::这个可以处理与标签同行的内容
复制代码

@echo off&setlocal enabledelayedexpansion 2>nul 3>nul
for /f "tokens=1,2* delims=:" %%a in ('findstr /v /n "<pre> </pre>" *.html') do (
	if "%%a" neq "!last!" set /a "n=1","t=0"&set last=%%a&cd.>%%~na.txt
	if "%%b"=="!n!" set /a n+=1,"t=^!t" else (
		if "!t!"=="0" echo;>>%%~na.txt %%c
		set /a n=%%b+2
	)
)
pause
复制代码