Sphider-utf - utf-8 version of sphider here

Hi to all, i make some changes in sphider so it can index multi-lang and utf-8 sites.

[code.google.com]

So here changes:

Shider now can index multy-lang sites, at least it index cp-1251 and utf-8 in 
russian pretty well. It may be some errors and may fail index some pages, but in 
general it work fine.

1. Sphider now use:
- UTF-8 so u must change u database collation to utf8_general_ci
- MySQLi class to interact with mysql
- now use multi-bytes string functions so it work correct with UTF-8 
2. FIXED MySQL server has gone away error
3. couse php limitation and current indexing algoritm shider ignore pages with
size more than 1 megabyte
4. Some changes made to database so make sure use sql/upgrade_to_1.4.sql to 
update u db.
5. Sphider now auto detect site codepage

and many other changes so many i cant remember all.

Current version seems works fine, but need to be tested. It may contains bugs and errors, so i will be glad if u report bugs and errors.

U can get sources from here:

[code.google.com]

p.s. If something dont work it maybe couse i forgot to put some files in rep

[code.google.com] - free for all | support [www.sphider.eu]

My search form does not work until I erase $mysqli_conn->query("SET NAMES 'utf8'"; and when I do that, I can not search utf 8



Edited 1 time(s). Last edit at 10/07/2011 05:20PM by vanja.

Hi, I tried with your sphider utf-8, its working fine for, English, but not for Arabic.

Fixed it,

I put
header("Content-Type: text/html; charset=utf-8";
in the beginning of search.php

Thanks

Hallo SkyRanger,

You said, that you've integrated a file-size limitation. What were exactly the Problems with bigger files?

Because the older Version initialised PDF-files with 3MB and I whink there were no problems.
Maybe, I've overlooked something?

Thanks for a response,

Schmidi

Is this being hosted some where else now?
His google code has been removed.

Just wanted to edit and say. my mistake, the files are not under downloads but under "source>>browse"



Edited 1 time(s). Last edit at 04/02/2012 09:02AM by cladiron.

Can anybody get this to work? Every time I try the search page is blank and does nothing. All the pages on my website are in UTF-8 and I would really like to be able to search for non-standard characters. Can anybody help out?

Thanks

Hi all,

In order to index and search MS Word documents containing non-ascii characters with Sphider, I had to do the following in Fedora Linux (both FC11 & FC17):

1) Modify mysql server configuration in /etc/my.cnf

[mysqld]
.........
skip-character-set-client-handshake
collation-server=utf8_unicode_ci
character-set-server=utf8
..........

2) Modify httpd (Apache) server configuration in /etc/httpd/conf/httpd.conf

AddDefaultCharset UTF-8

3) Modify php configuration in /etc/php.ini

default_charset="UTF-8"

4) Install xpdf package (yum install xpdf) in order to have pdftotext program in /usr/bin/pdftotext and report it in settings/conf.php

$pdftotext_path='/usr/bin/pdftotext';

After all that, French letters é,è,ç,à,ù,ö î etc. show up OK in links, documents and indexes, except in my Sphider admin "Statistics/Search log/Query" report column where they keep garbled (e.g. é instead of é)

This is solved (for php versions <5.4) by updating calls to php "htmlentities" function as follows

replace single-argument call

.htmlentities($word)

by 3-arguments call

.htmlentities($word, ENT_NOQUOTES, "UTF-8"

at lines 1095 and 1114 in admin.php.


I also noted that it is impossible to directly index .doc files without garbling such letters : one must export them from MS Word (and import them in Sphider using "Sites/Add site" either as .pdf files using a PDF converter (such as CutePDF) , or export "Web page" format files (.htm) from MS Word.

Hope it might help other users of standard Sphider struggling with non-ascii documents...

Fedora 11 ; php 5.2.12 ; mysql 14.14 Distrib 5.1.42 ; Apache 2.2.13 ; pdftotext 0.10.7 ; Sphider fresh install



Edited 1 time(s). Last edit at 08/25/2012 02:01PM by grandebou.