Maruan, I finally got back around to trying this new flatten method out. It works fantastic as far as preserving the appearance of the fields. However, if I flatten before a merge, I am still losing data entered into fields.
-----Original Message----- From: Maruan Sahyoun [mailto:[email protected]] Sent: Tuesday, September 15, 2015 4:15 AM To: [email protected] Subject: Re: How flatten without changing appearance Hi, > Am 15.09.2015 um 09:06 schrieb Tilman Hausherr <[email protected]>: > > Am 15.09.2015 um 09:02 schrieb Maruan Sahyoun: >>> Am 15.09.2015 um 08:56 schrieb Tilman Hausherr <[email protected]>: >>> >>> Am 15.09.2015 um 08:48 schrieb Maruan Sahyoun: >>>> Hi Tilman, >>>> >>>>> Am 15.09.2015 um 08:17 schrieb Tilman Hausherr <[email protected]>: >>>>> >>>>> Am 15.09.2015 um 04:47 schrieb Maruan Sahyoun: >>>>>> Hi Kevin, >>>>>> >>>>>> I've created https://issues.apache.org/jira/browse/PDFBOX-2970 >>>>>> <https://issues.apache.org/jira/browse/PDFBOX-2970> for that. >>>>>> >>>>>> In addition you can start playing with that code based on the current >>>>>> PDFBox 2.0.0 snapshot. >>>>>> >>>>>> >>>>>> >>>>>> PDDocument doc = PDDocument.load(new File(...)); >>>>>> PDAcroForm acroForm = doc.getDocumentCatalog().getAcroForm(); >>>>>> // Iterate over all form fields and their widgets and >>>>>> create a >>>>>> // FormXObject at the page content level from that >>>>>> for (PDField field : acroForm.getFieldTree()) >>>>>> { >>>>>> for (PDAnnotationWidget widget : >>>>>> ((PDTerminalField)field).getWidgets()) >>>>>> { >>>>>> PDPage page = widget.getPage(); >>>>>> PDPageContentStream contentStream = new >>>>>> PDPageContentStream(doc, page, true, true); >>>>> I wasn't 100% sure what was meant with "flatten", but now I get it! >>>>> >>>>> I think the first one of the "new PDPageContentStream" should have the >>>>> fifth parameter and be true, in case the existing content stream isn't >>>>> included in q...Q. >>>> thanks for stepping in. >>>> >>>> Is that really needed as we are storing/restoring the current state before >>>> the new content is added. Putting a store/restore around the existing one >>>> - assuming there is only one - would store/restore to the state before >>>> there was any content stream. In addition if there are already multiple >>>> streams before the new one would we need to put a store/restore around >>>> each of them? If there wasn't any that would potentially change the >>>> behavior of the old processing? >>> The constructor identifies if there is already content, and only then, it >>> inserts a saveGraphicsState() before the first existing stream and a >>> restoreGraphicsState() at the end of the new one, so it work both with >>> none, one, and many. >>> >>> Only doing it for the first one is to avoid having too many nesting >>> q...Q >> looking at how Adobe does it each of the new content streams is wrapped in a >> q ...Q - that's why I did the same. > > That isn't what I meant. I meant that only the first one should be > > PDPageContentStream contentStream = new PDPageContentStream(doc, page, > true, true, true); > > instead of > > PDPageContentStream contentStream = new PDPageContentStream(doc, page, > true, true); > > all the rest is fine as it is. > done - thanks for the clarification. Maruan > > Tilman > > > >> >>> What I was thinking of is cases where the existing content stream(s) >>> modifies the CTM without resetting it. This would have the effect that the >>> "fields" would appear at the wrong place / wrong size / outside of the >>> screen. We're having this every few months that somebody wants to add a >>> stamp to a page, it works for most, except for one "special" file. >>> >>> Tilman >>> >>> >>>> BR >>>> Maruan >>>> >>>> >>>> >>>>> Tilman >>>>> >>>>>> PDFormXObject fieldObject = new >>>>>> PDFormXObject(widget.getAppearance().getNormalAppearance().getAppearanceStream().getCOSStream()); >>>>>> Matrix translationMatrix = >>>>>> Matrix.getTranslateInstance(widget.getRectangle().getLowerLeftX(), >>>>>> widget.getRectangle().getLowerLeftY()); >>>>>> contentStream.saveGraphicsState(); >>>>>> contentStream.transform(translationMatrix); >>>>>> contentStream.drawForm(fieldObject); >>>>>> contentStream.restoreGraphicsState(); >>>>>> contentStream.close(); >>>>>> } >>>>>> } >>>>>> >>>>>> // preserve all non widget annotations >>>>>> for (PDPage page : doc.getPages()) >>>>>> { >>>>>> List<PDAnnotation> annotations = new >>>>>> ArrayList<PDAnnotation>(); >>>>>> for (PDAnnotation annotation: >>>>>> page.getAnnotations()) >>>>>> { >>>>>> if (!(annotation instanceof PDAnnotationWidget)) >>>>>> { >>>>>> annotations.add(annotation); >>>>>> } >>>>>> } >>>>>> page.setAnnotations(annotations); >>>>>> } >>>>>> // remove the fields >>>>>> acroForm.setFields(Collections.<PDField>emptyList()); >>>>>> doc.save(...); >>>>>> doc.close(); >>>>>> >>>>>> >>>>>> BR >>>>>> Maruan >>>>>> >>>>>> >>>>>> >>>>>>> Am 14.09.2015 um 17:57 schrieb Maruan Sahyoun <[email protected]>: >>>>>>> >>>>>>> >>>>>>> >>>>>>>> Am 14.09.2015 um 16:53 schrieb Kevin Ternes <[email protected]>: >>>>>>>> >>>>>>>> I am trying to refactor for 2.0.0-SNAPSHOT now. >>>>>>>> >>>>>>>> Can someone tell me how to flatten a PDDocument in PDFBox 2 ? >>>>>>>> All the examples seem to pertain to v1.x >>>>>>> there are non but I've started to get that into the project >>>>>>> >>>>>>>> -----Original Message----- >>>>>>>> From: Maruan Sahyoun [mailto:[email protected]] >>>>>>>> Sent: Saturday, September 12, 2015 1:55 AM >>>>>>>> To: [email protected] >>>>>>>> Subject: Re: How flatten without changing appearance >>>>>>>> >>>>>>>> Hi, >>>>>>>> >>>>>>>>> Am 12.09.2015 um 00:08 schrieb Kevin Ternes <[email protected]>: >>>>>>>>> >>>>>>>>> Maruan, >>>>>>>>> That would be great. Please have a look at: >>>>>>>>> https://onedrive.live.com/redir?resid=9CCA324BE57ADA7!76929&au >>>>>>>>> thkey=!A KE0x0fh5QDkIIw&ithint=file%2czip This should >>>>>>>>> demonstrate the problem by processing several files. >>>>>>>> I took a quick look. Which version of PDFBox are you using? Would it >>>>>>>> be possible to go to (the yet to be released) 2.0.0 version? >>>>>>>> >>>>>>>> BR >>>>>>>> Maruan >>>>>>>> >>>>>>>> >>>>>>>>> The users complaint is: Boxes are being greyed out and font sizes are >>>>>>>>> being changed by the flatten. >>>>>>>>> My flatten is done by the PDFBoxUtils.flatten() method. >>>>>>>>> >>>>>>>>> Thanks, >>>>>>>>> -Kevin >>>>>>>>> >>>>>>>>> -----Original Message----- >>>>>>>>> From: Maruan Sahyoun [mailto:[email protected]] >>>>>>>>> Sent: Friday, September 11, 2015 12:17 PM >>>>>>>>> To: [email protected] >>>>>>>>> Subject: Re: How to merge forms where there is >>>>>>>>> >>>>>>>>> how urgent is it for you? If you could share a sample file and your >>>>>>>>> code for flattening I could take a look. >>>>>>>>> >>>>>>>>> BR >>>>>>>>> Maruan >>>>>>>>> >>>>>>>>>> Am 11.09.2015 um 18:23 schrieb Kevin Ternes <[email protected]>: >>>>>>>>>> >>>>>>>>>> I have not been able to successfully flatten a document without >>>>>>>>>> affecting the formatting of the field. >>>>>>>>>> There are some example codes out there but none of them work >>>>>>>>>> correctly for me. >>>>>>>>>> >>>>>>>>>> -----Original Message----- >>>>>>>>>> From: Maruan Sahyoun [mailto:[email protected]] >>>>>>>>>> Sent: Friday, September 11, 2015 3:15 AM >>>>>>>>>> To: [email protected] >>>>>>>>>> Subject: Re: How to merge forms where there is >>>>>>>>>> >>>>>>>>>> Hi Kevin >>>>>>>>>> None of these field values will need to be changed afterward the >>>>>>>>>> merge. They are set to read-only. >>>>>>>>>>> I tried flattening the source fields but most of these documents >>>>>>>>>>> rely on field annotation for the field value formatting and this is >>>>>>>>>>> not to be changed. >>>>>>>>>> when a field is flattened correctly the formatting of the >>>>>>>>>> annotations visually representing the field become part of the page >>>>>>>>>> content stream. So if you don't need the fields at all - as there is >>>>>>>>>> no further input - you could flatten the source documents prior to >>>>>>>>>> merging them. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
