Have a look at PdfConverter

[ylussaud]

If it works fine, it would be nice to generate a docx or a pdf according to the file extension of the output document.

[ylussaud]

It is not part of the POI project and need new dependencies:

fr.opensagres.xdocreport fr.opensagres.poi.xwpf.converter.core 2.0.2

[ylussaud]

The converter uses iText which is LGPL that can be an other problem.

[ejuliot]

POI already has a built-in support for DOCX to PDF conversion. Loot at https://stackoverflow.com/questions/43363624/converting-docx-into-pdf-in-java (org.apache.poi.xwpf.converter.pdf.PdfConverter)

[ylussaud]

As stated above PdfConverter is not part of apache POI but fr.opensagres.poi.xwpf.converter.core that support apache POI 4.0.1. M2Doc is using apache POI 4.1.0 and will move to next versions.

[ylussaud]

The LGPL licence is not an issue, there is LGPL code in the Orbit update site. At the moment both M2Doc and fr.opensagres.poi.xwpf.converter.pdf 2.0.0 depend on POI 5.2.3 so I was able to tests the pdf conversion.

There are the following issues:

• when a table is present it sometimes throws an NPE:

  fr.opensagres.poi.xwpf.converter.core.XWPFConverterException: java.lang.NullPointerException: Cannot invoke "org.openxmlformats.schemas.wordprocessingml.x2006.main.CTTblGrid.getGridColList()" because "grid" is null
  at fr.opensagres.poi.xwpf.converter.pdf.PdfConverter.doConvert(PdfConverter.java:71)
  at fr.opensagres.poi.xwpf.converter.pdf.PdfConverter.doConvert(PdfConverter.java:39)
  at fr.opensagres.poi.xwpf.converter.core.AbstractXWPFConverter.convert(AbstractXWPFConverter.java:42)

  To solve this issue we need to add a CTTblGrid tot the created XWPFTable. This implies knowing the width of each column. A width has been added to the MCell (see #472) but I'm not sure we will be able to compute a width when importing from HTML. I'm opening this issue #525.

• with asImage test :

  java.lang.StackOverflowError
  at java.base/java.lang.StringBuffer.<init>(StringBuffer.java:133)
  at com.lowagie.text.pdf.BidiLine.createArrayOfPdfChunks(Unknown Source)
  at com.lowagie.text.pdf.BidiLine.createArrayOfPdfChunks(Unknown Source)
  at com.lowagie.text.pdf.BidiLine.processLine(Unknown Source)
  at com.lowagie.text.pdf.ColumnText.go(Unknown Source)
  at com.lowagie.text.pdf.ColumnText.goComposite(Unknown Source)
  at com.lowagie.text.pdf.ColumnText.go(Unknown Source)
  at com.lowagie.text.pdf.ColumnText.go(Unknown Source)
  at com.lowagie.text.pdf.PdfPRow.writeCells(Unknown Source)
  at com.lowagie.text.pdf.PdfPTable.writeSelectedRows(Unknown Source)
  at com.lowagie.text.pdf.PdfPTable.writeSelectedRows(Unknown Source)
  at com.lowagie.text.pdf.PdfPTable.writeSelectedRows(Unknown Source)
  at com.lowagie.text.pdf.ColumnText.goComposite(Unknown Source)
  at com.lowagie.text.pdf.ColumnText.go(Unknown Source)
  at com.lowagie.text.pdf.ColumnText.go(Unknown Source)
  at com.lowagie.text.pdf.PdfDocument.addPTable(Unknown Source)
  at com.lowagie.text.pdf.PdfDocument.add(Unknown Source)
  at com.lowagie.text.Document.add(Unknown Source)
  at fr.opensagres.xdocreport.itext.extension.ExtendedDocument.add(ExtendedDocument.java:114)
  at fr.opensagres.poi.xwpf.converter.pdf.internal.elements.StylableDocument.flushTable(StylableDocument.java:374)
  at fr.opensagres.poi.xwpf.converter.pdf.internal.elements.StylableDocument.pageBreak(StylableDocument.java:141)
  at fr.opensagres.poi.xwpf.converter.pdf.internal.elements.StylableDocument.columnBreak(StylableDocument.java:120)
  at fr.opensagres.poi.xwpf.converter.pdf.internal.elements.StylableDocument.simulateText(StylableDocument.java:230)
  at fr.opensagres.poi.xwpf.converter.pdf.internal.elements.StylableDocument.pageBreak(StylableDocument.java:160)
  at fr.opensagres.poi.xwpf.converter.pdf.internal.elements.StylableDocument.columnBreak(StylableDocument.java:120)

• some differences between the word document and the pdf document:
  • some bullets from bullet list are missing (HTML ul test)
  • ...

Overall the output pdf is pretty close to the word document if it don't use MTable.