Tesseract full integration - CE3.4b

As we transition to the new Hyland Community, some post links may be temporarily disrupted as we finalize redirects. We apologize for the inconvenience and are working to resolve this as quickly as possible. Details.

normando

Hello all.

I was writed a how to in spanish forums to integrate tesseract OCR into alfresco. I repeat the post here because that will be useful for somebody.

Sorry for my bad language. I will try to do my best.

Before beging the howto, it is importante to make a few considerations:

1- Tesseract only support tif files.
2- Tesseract 2.x do not work with files ended with .tiff extension. Only work with .tif extensions. Versión 3.x can
3- Because alfresco store the temporary files with an extension as your own mime type (ex. if you upload a .tif file, alfresco make a node for that file with the .tiff extension), we need to make a wrapper to deal with this issue (copy the node to .tif extension)
4- Tesseract append to the output file the .txt extension, so if you use:

tesseract input_file.tif output_file.txt‍

you will get a file output_file.txt.txt

So we will not add the .txt extension.

Before anything, ensure you have installed tesseract, and the dictionaries for the language you want to made OCR. Also add the english too.

The first step to do is test tesseract from the command line:

tesseract input_file.tif output_file -l eng‍

The syntax parameter -l indicate the tesseract will use the english dictionary. You can modify this to spa for spanish, and so on. If the above command produce a text file with the content of the OCR, then we can continue.

Create a file extension, named ocrtiff-transform-context.xml into /tomcat/shared/classes/alfresco/extenssion with the following content:

    <?xml version='1.0' encoding='UTF-8'?>
    <!DOCTYPE beans PUBLIC '-//SPRING//DTD BEAN//EN' 'http://www.springframework.org/dtd/spring-beans.dtd'>

    <beans>
        <bean id="transformer.worker.ocr.tiff" class="org.alfresco.repo.content.transform.RuntimeExecutableContentTransformerWorker">

            <property name="mimetypeService">
                <ref bean="mimetypeService" />
            </property>

              <property name="checkCommand">
                 <bean class="org.alfresco.util.exec.RuntimeExec">
                    <property name="commandsAndArguments">
                        <map>
                            <entry key=".*">
                                <list>
    <!–                            <value>tesseract</value> –>
                                    <value>/opt/alfresco/ocr</value>
                                </list>
                            </entry>
                        </map>
                    </property>
                    <property name="errorCodes">
                       <value>2</value>
                    </property>
                 </bean>
              </property>

              <property name="transformCommand">
                 <bean class="org.alfresco.util.exec.RuntimeExec">
                    <property name="commandsAndArguments">
                        <map>
                            <entry key=".*">
                                <list>
    <!–                            <value>tesseract</value>
                                    <value>${source}</value>
                                    <value>${target}</value>
                                    <value>-l</value>
                                    <value>eng</value> –>
                                    <value>/opt/alfresco/ocr</value>
                                    <value>${source}</value>
                                    <value>${target}</value>
                                </list>
                            </entry>
                        </map>
                    </property>
                    <property name="errorCodes">
                       <value>1,2</value>
                    </property>
                 </bean>
              </property>

              <property name="explicitTransformations">
                 <list>
                    <bean class="org.alfresco.repo.content.transform.ExplictTransformationDetails">
                        <property name="sourceMimetype"><value>image/tiff</value></property>
                        <property name="targetMimetype"><value>text/plain</value></property>
                    </bean>
                 </list>
              </property>
        </bean>

        <bean id="transformer.ocr.tiff" class="org.alfresco.repo.content.transform.ProxyContentTransformer" parent="baseContentTransformer">
            <property name="worker">
                <ref bean="transformer.worker.ocr.tiff" />
            </property>
        </bean>
    </beans>
‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍

Then create a wrapper file, named "ocr" and put into the alfresco root directory. In my case I put into /opt/alfresco.

    #!/bin/bash
    # save arguments to variables
    SOURCE=$1
    TARGET=$2
    TMPDIR=/tmp
    FILENAME=`basename $SOURCE`
    OCRFILE=$FILENAME.tif

    # to see what happens
    #echo "from $SOURCE to $TARGET" >>/tmp/ocrtransform.log

    cp -f $SOURCE $TMPDIR/$OCRFILE

    # call tesseract and redirect output to $TARGET
    tesseract $TMPDIR/$OCRFILE ${TARGET%\.*} -l eng
    rm -f $TMPDIR/$OCRFILE
‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍

Made then as executable file (chown 755 ocr)

Restart alfresco, and upload a tif file. Then if you are into alfresco explorer, click on the "I" (info) icon to see the strings that tesseract can extract from the fif file. Also you can made a search into alfresco explorer, or share (in this last, you need to upload the file under your site).

Enjoy

PD: If you can improve this how to, let me know.

fatal

Hi,

I think they are a problem with

Made then as executable file (chown 755 ocr)

I have another problem, i have create my extension but OCR don't start when i upload my .tif
it's possible to create screenshoot ? ty

Sorry my english is very bad (3.4.B version)