Tesseract - ERROR net.sourceforge.tess4j.Tesseract - null

Question

Created a java application that uses Tesseract in order to convert a given image or pdf to a string format, when running it on my machine as a unit test using junit it runs grea

Answer 1

Resources I used: Windows 10 (tried on Windows Server 2016 as well), JAVA, MAVEN

Status: Working good on my local as well as VM 

1. Download  Tess4J-3.4.8  from here http://tess4j.sourceforge.net/  and set your ENV variable path under Advance System Setting 
2. Get repo from MAVEN - 

<dependency>
<groupId>net.sourceforge.tess4j</groupId>
<artifactId>tess4j</artifactId>
<version>4.5.1</version>
</dependency>
<dependency>
<groupId>org.ghost4j</groupId>
<artifactId>ghost4j</artifactId>
<version>1.0.1</version>
</dependency>
<dependency>
<groupId>net.sourceforge.lept4j</groupId>
<artifactId>lept4j</artifactId>
<version>1.7.0</version>
</dependency>

3. Get libtesseract302.dll and copy to "C:\Windows\System32" folder 
from here http://api.256file.com/libtesseract302.dll/en-download-56466.html
do not forget to set your ENV variable path under Advance System Setting  

4. Download and install Visual C++ 2015 Redistributable or VC++ 2017 Redistributable (I installed both )
from here https://programmer.help/blogs/net.sourceforge.tess4j.tesseractexception-java.lang.nullpointerexception.html 

then do restart your PC 

5. on Safer side can have some Jar files if you dont have already in local - Please see image

do not forget to set your ENV variable path for JARs under Advance System Setting 

Answer 2

As @Piotr R mentioned the error was ghostscriptException.getCause() is null and the reason for that is that the path configured in the file object sent to Tesseract was not a valid one, now the definition of valid for Tesseract is a bit different then yours, he consider only a local address as valid, so when setting a file located on AWS S3 even if it's public it will throw an error. The solution was saving it locally and deleting it after Tesseract is done.

Answer 3

My guess is that there is GhostscriptException which is not logged properly, and this is causing NullPointerException:

https://github.com/nguyenq/tess4j/blob/212d72bc2ec8b3a4d4f5a18f1eb01a0622fc5521/src/main/java/net/sourceforge/tess4j/util/PdfUtilities.java#L107

106        } catch (GhostscriptException e) {
107            logger.error(e.getCause().toString(), e);
108        } finally {

In line 107 - e.getCause() is (probably) null, calling null.toString() throws NPE.

(from the specs - getCause can be null: https://docs.oracle.com/javase/7/docs/api/java/lang/Throwable.html#getCause(), GhostscriptException is also allowing the cause to be null: http://grepcode.com/file/repo1.maven.org/maven2/org.ghost4j/ghost4j/1.0.0/org/ghost4j/GhostscriptException.java)

To verify this answer (without recompiling the whole tess4j) you could start your program in the debug mode and put a breakpoint at line 107. This will give you information about the real Exception.