sys_TextExtraction and 6.6

Rhythmyx Technical Discussion
1

Is sys_TextExtraction still going to work if we upgrade to 6.6? I ask because I notice in the documentation it says that this extension requires full-text search, and I know that the searching technology is different in 6.6.

Thanks,
April

2

April,

Text extraction still works, with two caveats:

First, we had to implement a new text extraction extension, sys_textExtractor for the Lucene search engine. The extension used with the Convera search engine is obsolete and should be replaced with the new extension.

Second, the list of supported file types for text extraction is much smaller.

For details, see the upgrade document appropriate to the version from which you are upgrading.

RLJII

3

Hi Robert

This list of input parameters in the document does not match those of the sys_textextractor edit. The exit only has three params(Source, OutputParam and ErrorMessageParam) whilst the documentation has five (Source, OutputParam, FileTypeParam, ErrorMessageParamOutputEncoding)

When I use the sys_textextractor as described in the document I get the following error message when open the content editor:

0 An error occurred while pre-processing the request submitted by the host (-not available-).
1 Java/global/percussion/contenteditor/sys_textExtractor: First two params are required.

My input params are as follows:

I’ve tried this:

Source = PSXSingleHTMLParam/item_file_attachment
OutputParam = PSXSingleHTMLParam/extracted_text
ErrorMessageParam = Literal error_message

and this:

Source = PSXSingleHTMLParam/item_file_attachment
OutputParam = Literal extracted_text
ErrorMessageParam = Literal error_message

Just so you know the system is 6.6.1 upgraded from 6.5.2 upgraded from 5.71.

Cheers
James

4

James,

On investigation, you are correct; the extension only has three parameters. I have updated our source and the corrected information will be included next time we generate the Technical Reference.

As to your problem, try using the literal name of the field rather than the the field in all three extension parameters, rather than the SingleHTMLParameter. The developer informs me that specifying the field name as a literal is the correct approach. I have updated the source to address this problem as well.

RLJII

5

Hi Robert

Add all values as literals and it works a treat.

Cheers
James