以编程方式捕获HTML文件中的样式并使用PowerShell转换为Word文档

Question

我正在制作一个脚本，它将在两年一度的印刷会议上提取30多个谈话的HTML文件。已选择Word作为打印工具。会谈有大胆，斜体，轮廓，标题和其他与风格相关的元素。我需要将所有这些样式捕获到word文档中，而无需人工干预。我尝试了两种解决这个问题的方法。

1. 针对每个谈话并挑选包含相关信息（标题，演讲者，段落等）的元素。这很漂亮，但风格缺失。有没有办法用这种方法保留样式信息？ 下面的完整代码示例显示了所有协同工作：

   $primaryURL = "https://www.domain.tld/folder?lang=eng"
   $baseURL = $primaryURL.Split("`?")[0]
   $talkArray = @()
   $content = (Invoke-WebRequest -Uri $primaryURL).Content
   $links = (Invoke-WebRequest -Uri $primaryURL).Links
   $talkLinks = $links | Where {$_.outerHTML -like "*/folder/$year/$month*"} | Select -expand href
   $talkInfo = $links | Where {$_.outerHTML -like "*/folder/$year/$month*"} | Select -expand outerText
   $domainName = $baseUrl.Split("/")[2]
   Foreach ($talk in $talkInfo)
   {
       $title = ($talk.Split("`n`r") -replace "`#.*", "$([char]0)" -replace "#.*" -replace "$([char]"0", "#" -replace "^\s*" -replace "\s*$")" | ? { $_; })[1]
       $speaker = ($talk.Split("`n`r") -replace "`#.*", "$([char]0)" -replace "#.*" -replace "$([char]"0", "#" -replace "^\s*" -replace "\s*$")" | ? { $_; })[2]
       $link = "https://" + $domainName + $talkLinks[$count]
   
       $obj = New-Object PSObject
       $obj | Add-Member -MemberType NoteProperty -Name "Title" -Value $title
       $obj | Add-Member -MemberType NoteProperty -Name "Speaker" -Value $speaker
       $obj | Add-Member -MemberType NoteProperty -Name "url" -Value $link
       $talkArray += $obj
       $count++
   }
   $wordDocPath = "U:\" + $month + " " + $year + " Conference.docx"
   $word = New-Object -ComObject word.application
   $word.Visible = $false
   $doc = $word.documents.add()
   $doc.Styles["No Spacing"].ParagraphFormat.SpaceAfter = 0
   $doc.Styles["No Spacing"].ParagraphFormat.SpaceBefore = 0
   $margin = 36 # 0.5 inches
   $doc.PageSetup.LeftMargin = $margin
   $doc.PageSetup.RightMargin = $margin
   $doc.PageSetup.TopMargin = $margin
   $doc.PageSetup.BottomMargin = $margin
   $selection = $word.Selection
   
   $outputPath = $wordDocPath
   $doc.SaveAs($outputPath)
   $doc.Close()
   $word.Quit()
   $word = $null
   
   $objWord = New-Object -comobject Word.Application
   $objWord.Visible = $false 
   $objDoc = $objWord.Documents.Open($wordDocPath)
   $objSelection = $objWord.Selection
   $trash = $objSelection.EndKey(6, 0)
   $objSelection.PageSetup.TextColumns.SetCount(2)
   
   ForEach ($talk in $talkArray)
   {
       $talkTitle = $talk.Title
       $a, $p, $results, $wrap, $n = $null
       $talkContent = Invoke-WebRequest -Uri $talk.url
       $speakerPic = "https:" + ($talkContent.images | Where {$_.alt -like "$talkTitle"}).src
       $picName = (($talkContent.images | Where {$_.alt -like "$talkTitle"}).src).Split("/")[-1]
       $kicker = ($talkContent.ParsedHTML.GetElementByID('kicker1')).innerHTML
       $count = 3
       $line = 1
   
       # - Download the picture
       $picPath = $tempDir + "\" + $picName
       Download-File -url $speakerPic -file $picPath
   
       # - Write the header information
       $text = $talkTitle
       $objSelection.Style="Heading 1"
       $objSelection.Font.Name = "Times New Roman"
       $objSelection.TypeText($text)
       $objSelection.TypeParagraph()
       $text = $talk.Speaker + "`r`n"
       $objSelection.Font.Bold = 1
       $objSelection.Font.Name = "Times New Roman"
       $objSelection.TypeText("by $text")
       $objSelection.Font.Bold = 0
       $objSelection.TypeParagraph()
       $objShape = $objDoc.Shapes
       $wrap = [Microsoft.Office.Interop.Word.WdWrapType]::wdWrapSquare # WdWrapTopBottom
       $objShape.AddPicture($picPath) | Out-Null
       $objShape.Range(1).WrapFormat.Type = $wrap
       $objSelection.Endkey()
       $objSelection.Font.Bold = 1
       $objSelection.Font.Italic = 1
       $objSelection.Font.Name = "Times New Roman"
       $objSelection.TypeText($kicker)
       $objSelection.TypeParagraph()
       $objSelection.Font.Bold = 0
       $objSelection.Font.Italic = 0
   
       # - Collect the paragraphs
       $paragraphs = @()
       do
       {
           [string]$p = ($talkContent.ParsedHTML.getElementById("p$count")).InnerText
           if ($p) {
               $obj = New-Object PSObject
               $obj | Add-Member -MemberType NoteProperty -Name "p" -Value $p
               $obj | Add-Member -MemberType NoteProperty -Name "line" -Value $line
               $paragraphs += $obj
               $line = $line + 1
           }
           Else {
           }
           $count = $count + 1
       } while ($p)
   
       # - Write the paragraphs
       $paragraphs | % {
           $results = [regex]::Split($_.p,'(?<![\d\s])(?<![\(\s])(?<![\:\s])(?<![\-\s])(\d{1,2})(?!\d)')
           $line = $_.line
   
           $objSelection.Style="Normal"
           $objSelection.Font.SuperScript = 0
           $objSelection.Font.Size = 9
           $objSelection.Font.Name = "Times New Roman"
           $objSelection.TypeText("$line. ")
           foreach ($item in $results)
           {
               $text = $item
               if ($text -match '(?<![\d\s])(?<![\(\s])(?<![\:\s])(?<![\-\s])(\d{1,2})(?!\d)')
               {
                   $objSelection.Font.SuperScript = 1
                   $objSelection.TypeText("$text")
                   $objSelection.Font.SuperScript = 0
               }
               Else
               {
                   if ($text -ne " ")
                   {
                       #$objSelection.Style="Normal"
                       $objSelection.Font.SuperScript = 0
                       $objSelection.Font.Size = 9
                       $objSelection.Font.Name = "Times New Roman"
                       $objSelection.TypeText("$text")
                   }
               }
   
           }
           $objSelection.TypeParagraph()
       }
   
       $count = 1
       # - Work on the notes
       $notes = @()
       $line = 1
       do
       {
           [string]$n = ($talkContent.ParsedHTML.getElementById("note$count")).InnerText
           if ($n) 
           {
               $obj = New-Object PSObject
               $obj | Add-Member -MemberType NoteProperty -Name "n" -Value $n -Force
               $obj | Add-Member -MemberType NoteProperty -Name "line" -Value $line -Force
               $notes += $obj
               $line = $line + 1
           }
           Else {
           }
           $count = $count + 1
       } while ($n)
       $text = "`vNOTES:"
       $objSelection.TypeText("$text")
       $objSelection.TypeParagraph()
   
       $notes | % {
           $note = $_.n
           $line = $_.line
           $text = $note
           $objSelection.Font.Size = 9
           $objSelection.Font.Name = "Times New Roman"
           $objSelection.TypeText("$line. $text")
           $objSelection.TypeText("`r`n")
       }
       $objSelection.InsertNewPage()
   }
   
   $objDoc.saveas($wordDocPath)
   $objDoc.close()
   $objWord.quit()
   $objWord = $null

来自html的示例段落（带有样式元素）：

    <p id="p1">This is an example with <em>italic</em> text and <strong>bold</strong> text.</p>

2. 转储每个HTML文件，清理它，然后转换为Word。这样可以保留这种风格，但是去除我几乎放弃的所有垃圾（顶部的网页标题，链接和菜单，页脚底部的链接和导航按钮）变得非常麻烦。也许有一种更简单的方法？代码示例如下：

   # Code snipped. This version performs similar functions to collect information on each talk, creates the temp folder and the word doc. Then:
   
   [ref]$SaveFormat = "microsoft.office.interop.word.WdSaveFormat" -as [type]
   $word = New-Object -ComObject word.application
   $word.visible = $false
   $document = $word.documents.open($htmlPath)
   $selection = $word.Selection
   $paras = $document.Paragraphs
   foreach ($para in $paras) 
   { 
       $text = $para.Range.Text
       If ($text -like "$title*")
       {
           break
       }
       Else
       {
           $para.Range.Select()
           $selection.Cut()
       }
   }
   $document.saveas([ref] $wordDocPath, [ref]$SaveFormat::wdFormatDocumentDefault)
   $document.close()
   $word.Quit() 
   $word = $null

这不是剥离违规元素的完整代码，但它首先在talk title元素之前清除所有内容。在此之后还有其他行必须删除，然后正如我所提到的，在谈话段落之后的几个元素。在此之后，我将需要处理Word中的每个段落以匹配最终产品的字体，大小，列，标题等。

问题：

1. 你会采用哪种方法？
2. 如果方法1，如何将段落标记中的样式元素转换为Word文档？
3. 如果方法2，是否有更好的方法来处理文档，以便我可以归零我正在寻找的元素？

提前致谢。

Answer 1

我用这种方法。

1. 刮掉并保存你想要的部分作为html_file。
2. 使用Range.Insertfile(html_file)将保留html中的所有样式。