Code to convert all types of Word Processing files to plain text.

[msevior physics unimelb edu au]

Hi folks,
         I've been in contact with wmealing about an idea of convert Word
Processor files to plain text files for indexing purposes. I
believe that one of the aims of the dashboard project is locate
files by content. Since a large fraction of the users generate
word processor files, getting the content of these files is
difficult - hence this code.

The attached code defines a simple C-interface into code that controls a
remote AbiWord process that converts files of doc,rtf,abw,wpd,sdx (plus
loads more types I can't remember) to plain text files.

The conversion is very fast as a view of the document is not generated.  A
190 page RTF document full of tables is converted to plain text in about 7
seconds on my 1 Ghz laptop.

There are just two simple functions you need to call.

/*
 * The function is fault tolerant and will report if a requested
 * file is not capable of being coverted.
 *
 * It will also restart AbiWord if it does crash on the next invocation
 * of convertFileToText.
 *
 * Inputs: inFile - The path to the word processor file to be converted.
 *         outFile - The path to the file containing the plain text of the
 *                   word processor file.
 * The function blocks until the conversion is complete or until
 * the conversion fails.
 *
 * It returns 0 upon a successful conversion
 *    and -1 if the conversion fails.
 */
int convertFileToText(const char * inFile, const char * outFile)

and...

/*
 * Call this method after all conversions have completed. Otherwise you'll
 * have a runaway AbiWord-2.2 process.
 */
int finalizeConversions(void)

You need CVS HEAD AbiWord (the 2.1.0 release we'll do in a week or two
will be fine) as well as the AbiCommand plugin to provide the command-line
interface which enables the conversions. The ConvertToText.c and
ConvertToText.h files are maintained in the AbiWord CVS archive in the
abiword-plugins cvs module and the abiword-plugins/tools/abicommand
directory.

I will actively maintain the code and add new features. (For example a
thumb-nailer of wordprocessing documents.)

Best wishes,

Martin Sevior

Attachment: ConvertToText.c
Description: Binary data

Attachment: ConvertToText.h
Description: Binary data