PDA

View Full Version : Junk in keywords table - Indexing PDF


Bege
04-09-2004, 12:52 AM
I have 1.8.0 installed on Redhat Linux.
When I index my pdf files I get lots of junk in the keywords table. It finds the file ok, but I don't get anything worth any value. Below is a snippit of some of the data. I have used the sample .PDF from this site with no luck. I have read through most of the forums with no luck. I am using pdftotext to create my plaintext file. It dosen't support STDOUT, but does create a txt file that I can open and see that it did parse the file correctly. I have also included a little snippit of my config.php. It almost looks like it is getting the encoding wrong, does anyone have any ideas? Thanks much,
Bege

+----------------+
| keyword |
+----------------+
| 6aeyqo,n |
| E#b5 |
| k��de�� |
| 3Iqha:cp |
+----------------+

define('PHPDIG_INDEX_PDF',true);
define('PHPDIG_PARSE_PDF','/usr/local/bin/pdftotext');
define('PHPDIG_OPTION_PDF','');
define('PHPDIG_PDF_EXTENSION','txt');

Bege
04-09-2004, 01:13 AM
I have looked at the temp files in the text_content dir, and all of the junk that i am getting in the database is in this file. How is the file getting created? When i run the pdftotext in bash everything works just fine, what is the difference?

Charter
04-09-2004, 08:15 AM
Hi. Don't forget the period...

define('PHPDIG_PDF_EXTENSION','.txt');

Try man pdftotext and look for enc to set the encoding if needed.