PDF Convert Server Instructions

Overview

The MoneyThumb PDF converter is a Java based server for the conversion and analysis of PDF files

Prerequisites

Linux or Windows server, Java 7 as a minimum version

Minimum of 4GB memory

Installation Package

The installation package includes the following components

jar File

The jar file is available as either of these packages.

MTConvertServerFull.zip containing a single file MTConvertServerFull_versionnumber.jar. This is a runnable jar.
MTConvertServer.zip with just the MoneyThumb jar files and a Pom_dependencies.xml. The two jar files are:

MTConvertServer_versionnumber.jar
pdfocr.jar_versionnumber.jar

The Java environment also needs to have memory sizes for maximum heap size of 1.1GB, and thread stack size of 2 MB. For example:

-Xmx1100m
-Xss2m

For example, in a standard tomcat environment this would be specified in catalina.sh as

export CATALINA_OPTS=”-Xss2m -Xmx1100m”

OCR Module

The OCR module is a native application that needs to be installed.

Download the appropriate file for your platform:

NSOCR_linux.zip
NSOCR_windows32.zip
NSOCR_windows64.zip

The file zip contains all the necessary files. Unpack that to a folder such as

c:\program files\nsocr (for Windows)

or

/usr/local/nsocr (for linux)

The Java invocation then needs to define the path for calling native applications. Add the folder to the Java invocation to specify JNI files with by adding a definition such as the ones below to the Java startup string.

-Djava.library.path=”c:\program files\nsocr”

-Djava.library.path=/usr/local/nsocr/bin

There is no API documentation for the OCR module, it is called as required by the server.

License

The license code is a String approximately 90 characters long and is input as an argument to the conversion routines in PDFConvertServer