前扑后继
太紧
亚克西填词大赛
春天的故事
答记者问－－送点土特产…
感谢你,孔子
海豚湾
复制代码

@echo off
cd.>b.txt
for /f "delims=" %%a in (a.txt) do (
    sed -n "s/.*<a href=\"\^(.*\^)\" target=.*title=\"%%a\".*/\1/p" cs.html | more >>b.txt
)
复制代码

@echo off
cd.>b.txt
for /f "delims=" %%a in (a.txt) do (
    sed -n "s/.*<a href=\"\^(.*\^)\" target=.*title=.*%%a.*/\1/p" cs.html | more >>b.txt
)
复制代码

@ECHO OFF
SETLOCAL ENABLEDELAYEDEXPANSION
FOR /F %%i IN (a.txt) DO (
    FOR /F "delims=" %%j IN ('FINDSTR /R "%%i" cs.html') DO (
        SET html=%%j
        SET html=!html:^<=!
        SET html=!html:^>=!
        FINDSTR /R "%%i" !html! 2> error.txt
        FOR /F "delims=" %%k IN (error.txt) DO (
            SET /A line += 1
            SET arr[!line!]=%%k
            ECHO %%k | FINDSTR /R "%%i" > NUL && ( SET /A line -= 2 & CALL ECHO %%i %%arr[!line!]:~26%% )
        )
    )
)
DEL error.txt & PAUSE
复制代码

Dim regex, fso, vbstr, regstr, vbout
Set fso = CreateObject("scripting.filesystemobject")
vbstr = fso.OpenTextFile("cs.html", 1).ReadAll()
regstr = replace("(" & Replace(fso.OpenTextFile("a.txt", 1).ReadAll(), vbCrLf, "|") & ")", "…", "")
RegExptest vbstr
fso.OpenTextFile("url.txt", 2, 1).Write vbout
Set fso = Nothing

Function RegExptest(rgstr)
  Dim match
  Set regex = New RegExp
  regex.Global = True
  regex.IgnoreCase = True
  regEx.pattern = "<a\s+[.\n]*?href=""(\w+://[^:\s]+)?""[^<>]+""" & regstr & """>"
  For Each match In regex.Execute(rgstr)
    vbout = vbout & match.Submatches(0) & vbCrLf
  Next
  Set regex = Nothing
End Function
复制代码

Dim regex, fso, vbstr, regstr, vbout, oxml, ostream
Set fso = CreateObject("scripting.filesystemobject")
If Not fso.FolderExists("韩寒") Then fso.CreateFolder("韩寒")
Set ohttp = CreateObject("microsoft.xmlhttp")
Set ostream = CreateObject("adodb.stream")
vbstr = fso.OpenTextFile("cs.html", 1).ReadAll()
For Each regstr In  Split(replace(fso.OpenTextFile("a.txt", 1).ReadAll(), "…", ""), vbCrLf)
  RegExptest vbstr
Next
Set fso = Nothing
Set ohttp = Nothing
Set ostream = Nothing
MsgBox "ok"

Function RegExptest(rgstr)
  Dim match
  Set regex = New RegExp
  regex.Global = True
  regex.IgnoreCase = True
  regEx.pattern = "<a\s+[.\n]*?href=""(\w+://[^:\s]+)?""[^<>]*""" & regstr & """>"
  For Each match In regex.Execute(rgstr)
    dowload match.Submatches(0)
  Next
  Set regex = Nothing
End Function

Function dowload(url)
  ohttp.open "get", url, False 
  Do Until ohttp.readyState = 1 : WScript.Sleep 200 : Loop
  ohttp.send()
  ostream.Mode = 3
  ostream.Type = 1
  ostream.Open()
  ostream.Write ohttp.responseBody
  ostream.SaveToFile "韩寒\" & regstr & ".html", 2
  ostream.Close
End Function
复制代码

@ECHO OFF
SETLOCAL ENABLEDELAYEDEXPANSION
FOR /F "delims=" %%i IN ('FINDSTR "articleCell" cs.html') DO (
    SET html=%%i
    SET html=!html:^<=!
    SET html=!html:^>=!
    SET html=!html:"_blank" =!
    SET html=!html: target=!
)
FINDSTR %html% 2> err.txt
FOR /F %%i IN (a.txt) DO (
    FOR /F "tokens=1-2 delims==" %%j IN ('FINDSTR "%%i" err.txt') DO (
        ECHO %%i %%k
    )
)
DEL err.txt & PAUSE
复制代码