Pdf text extraction failure

suviswan · August 22, 2014, 3:57am

Hi

I am using Aspose Java Pdf 9.3.0. In one of our machine, the text extraction from pdf fails. The is the stack trace is at the end of this post.

We are using java version “1.7.0_60” and it runs on ubuntu "Linux xxxx 3.2.0-58-generic #88-Ubuntu SMP Tue Dec 3 17:37:58 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux"

But the same code works in Mac OS X and CentOs. How to debug this problem ?

Exception in thread "main" java.lang.NullPointerException

at com.aspose.pdf.internal.p426.z12.m1(Unknown Source)

at com.aspose.pdf.internal.p426.z12.m7(Unknown Source)

at com.aspose.pdf.internal.p426.z14.m1(Unknown Source)

at com.aspose.pdf.internal.p426.z14.m6(Unknown Source)

at com.aspose.pdf.internal.p426.z14.(Unknown Source)

at com.aspose.pdf.TextAbsorber.visit(Unknown Source)

at com.aspose.pdf.facades.PdfExtractor.extractTextInternal(Unknown Source)

at com.aspose.pdf.facades.PdfExtractor.extractText(Unknown Source)

tilal.ahmad · August 22, 2014, 10:25am

Hi Surendra,

Thanks for your inquiry. We will appreciate it if you please share your sample PDF document and code here. We will test the scenario at our end on Ubuntu server and will update you our findings accordingly.

We are sorry for the inconvenience caused.

Best Regards,

suviswan · August 22, 2014, 10:37pm

The below is the piece of code that extracts text. The actual pdf is failing. I tried 3 more pdfs and all are failing.

PdfExtractor extractor = new PdfExtractor();

extractor.bindPdf(inputStream);

extractor.extractText(Charset.forName("UTF-8"));

ByteArrayOutputStream baos = new ByteArrayOutputStream();

extractor.getText(baos);

return new String(baos.toByteArray(), "UTF-8");

tilal.ahmad · August 25, 2014, 12:17am

Hi Surendra,

Thanks for sharing additional information. We are trying to replicate your issue over a Ubuntu server VM and will update you our findings soon.

Best Regards,

tilal.ahmad · August 25, 2014, 12:24pm

Hi Surendra,

Thanks for your patience. We have tested the scenario using Ubuntu 13.10 server with Aspose.Pdf for Java 9.3.1 and unable to replicate the issue. Please download and try latest version of Aspose.Pdf for Java 9.3.1. Hopefully it will resolve the issue. If issue persist then please share your Ubuntu version, So we will try to replicate the issue on specific version.

We are sorry for the inconvenience caused.

Best Regards,

suviswan · August 25, 2014, 10:32pm

Hi

Thanks for the update. Our Ubuntu version is 12.04.

lsb_release -a

No LSB modules are available.

Distributor ID: Ubuntu

Description: Ubuntu 12.04.4 LTS

Release: 12.04

Codename: precise

In the mean time we will try with 9.3.1

Thanks

Suren

suviswan · August 25, 2014, 10:57pm

Hi

I tried 9.3.1 and it’s also failing with the same error.

Exception in thread “main” java.lang.NullPointerException

at com.aspose.pdf.internal.p428.z12.m1(Unknown Source)

at com.aspose.pdf.internal.p428.z12.m7(Unknown Source)

at com.aspose.pdf.internal.p428.z14.m1(Unknown Source)

at com.aspose.pdf.internal.p428.z14.m6(Unknown Source)

at com.aspose.pdf.internal.p428.z14.(Unknown Source)

at com.aspose.pdf.TextAbsorber.visit(Unknown Source)

at com.aspose.pdf.facades.PdfExtractor.extractTextInternal(Unknown Source)

at com.aspose.pdf.facades.PdfExtractor.extractText(Unknown Source)

Thanks

Suren

tilal.ahmad · August 27, 2014, 12:47am

Hi Suren,

Thanks for your feedback. I am preparing the
required platform to simulate the environment as of yours. As soon as
everything is setup, I will test the issue on my end and will suggest you
accordingly.

We are sorry for the inconvenience faced.

Best Regards,

suviswan · August 27, 2014, 2:19am

Hi

Thanks for the update. Is there anyway i can provide the debug logs ? Basically we have struck in this issue and unable to proceed and our deadlines are fast approaching

Thanks

Suren

tilal.ahmad · August 28, 2014, 12:59am

Hi Suren,

Thanks for your inquiry. Yes you can share the debug logs in this thread, either copy logs in body of message or attach to thread. We will look into it and guide you accordingly.

We are sorry for the inconvenience caused.

Best Regards,

suviswan · August 28, 2014, 4:25am

Hi

My question about debug logs were vague in the previous post. My actual question is how do i take debug logs for this problem ?

Thanks

Suren

tilal.ahmad · August 29, 2014, 2:56am

Hi Suren,

Thanks for your feedback and sorry for the confusion. Yes Aspose.PDF for Java jar is obfuscated so you can not get any further logs from Aspose.Pdf for Java jar. We are investigated the issue and will keep you updated.

Best Regards,

suviswan · September 1, 2014, 10:32pm

Hi

Were you able to reproduce the problem ? Kindly let us know the progress and the fix date. This will be a critical aspect in buying the aspose license.

Thanks

Suren

tilal.ahmad · September 2, 2014, 9:59pm

Hi Suren,

We are sorry for the inconvenience. I am afraid the requied VM is not prepared, we are working over it and will update you soon.

Meanwhile can you please install Microsoft fonts and set default locale to English before instantiating Aspose.Pdf obect as following and share the results. As recent past we have noticed some issues with fonts/locale setting.

sudo apt-get install ttf-mscorefonts-installer

Locale.setDefault(Locale.ENGLISH);

…

Best Regards,

suviswan · September 3, 2014, 5:11am

Hi

I tried this and it’s not working. The same exception is thrown.

Thanks

Suren

tilal.ahmad · September 4, 2014, 3:17am

Hi Suren,

Thanks for your feedback. We are working to replicate the issue and will get back to you soon.

We are sorry for the inconvenience caused.

Best Regards,

joopduim · September 4, 2014, 7:47am

Hi suviswan,

I had exactly the same problem. Had to do with the Microsoft True Type fonts not being installed on our Linux server. We are running JBoss application server under Suse Linux. However it looks like not all MTT fonts are needed. I ended up with putting only one font file (copied from C:\windows\fonts) under our JBoss conf directory in a newly created fonts directory. Looks like that is enough for the PDF conversion to start. I have tested several PDF files and they all are correctly converted.

The extra folder was added with the following code:

Document.addLocalFontPath(System.getProperty(“jboss.server.home.dir”) + File.separator + “conf” + File.separator + “fonts”);

Regards,

Joop

suviswan · September 4, 2014, 10:43am

Hi Joop

Thanks for the response ! Do you remember which font is it ? I have installed fonts package suggested by “Tilal Ahmad”. But it didn’t solve the problem. If you can remember the specific font, i can try the solution suggested by you.

Thanks

Suren

tilal.ahmad · September 5, 2014, 5:54am

Hi Suren,

Thanks for your inquiry. You can reference installed Microsoft fonts folder in Linux as following. Hopefully it will help you to accomplish the task

// Set font folder path… folder
path of mscorefonts<o:p></o:p>

String path = "/usr/share/fonts/truetype/msttcorefonts/";

// Adding a single font directory

com.aspose.pdf.Document.addLocalFontPath(path);

// seting the user list for standard font directories

//java.util.List list = com.aspose.pdf.Document.getLocalFontPaths();

//list.add(path);

com.aspose.pdf.Document pdfDocument = new com.aspose.pdf.Document(inputStream);

PdfExtractor extractor = new PdfExtractor();

extractor.bindPdf(pdfDocument);

extractor.extractText(Charset.forName("UTF-8"));

ByteArrayOutputStream baos = new ByteArrayOutputStream();

extractor.getText(baos);

return new String(baos.toByteArray(), "UTF-8");

Please feel free to contact us for any further assistance.

Best Regards,

suviswan · September 5, 2014, 8:24am

Hi

Thanks ! This solved the problem. I will do more testing on our use cases and get back. Thanks again for the help.

Cheers

Suren