Sphinx中文分词在discuz 中的应用

推荐原创

coralzd 2014-12-23 16:13:20 博主文章分类：database ©著作权

©著作权归作者所有：来自51CTO博客作者coralzd的原创作品，请联系作者获取转载授权，否则将追究法律责任

Sphinx中文分词在discuz 中的应用。

Sphinx-for-chinese是一款专注于中文搜索的全文检索软件，在sphinx的基础上添加了中文处理模块并优化了中文搜索效果。相比而言，以某某论坛几千万数据环境来看，比Coreseek封装的sphinx中文分词+mmseg3组合要好一点

1.准备环境

cd /var/tmp/
wget http://www.sphinx-search.com/downloads/sphinx-for-chinese-2.2.1-dev-r4311.tar.gz
wget http://docs.zjyxh.com/discuzx/sphinx-for-chinese/xdict
mkdir -p /data0/dzbbs/indexdata/
mkdir -p /data1/dzbbs/indexdata/

2.安装 sphinx for chinese

CXX=gcc
CFLAGS=" -O3 -fomit-frame-pointer -pipe -march=nocona -mfpmath=sse -m128bit-long-double -mmmx -msse -msse2 -maccumulate-outgoing-args -m64 -ftree-loop-linear -fprefetch-l
oop-arrays -freg-struct-return -fgcse-sm -fgcse-las -frename-registers -fforce-addr -fivopts -ftree-vectorize -ftracer -frename-registers -minline-all-stringops -fbranch-target-
load-optimize2"
CXXFLAGS="${CFLAGS}"
export CXX CXXFLAGS CFLAGS
tar zxf sphinx-for-chinese-2.2.1-dev-r4311.tar.gz
cd sphinx-for-chinese-2.2.1-dev-r4311
./configure --prefix=/usr/local/sfc22/ --with-mysql=/usr/local/mysql/ --enable-id64
make -j8 install

3.配置 sphinx for chinese

cd /usr/local/sfc22/etc/

mv xdict xdictbak

cp -a /var/tmp/xdict .

vim sphinx.conf

source pre_forum_thread

{

type = mysql

sql_host = localhost

sql_user = sphinx

sql_pass = sphinx2013

sql_db = discuzx

sql_port = 3306

sql_query_pre = SET NAMES UTF8

sql_query_pre = SET SESSION query_cache_type=OFF

sql_query_pre = REPLACE INTO pre_common_sphinxcounter SELECT 1, MAX(tid) FROM pre_forum_thread

sql_query = SELECT t.tid as id,t.tid,t.subject,t.digest,t.displayorder,t.authorid,t.lastpost,t.special \

FROM pre_forum_thread AS t \

WHERE t.tid>=$start AND t.tid<=$end

sql_query_range = SELECT (SELECT MIN(tid) FROM pre_forum_thread),maxid FROM pre_common_sphinxcounter WHERE indexid=1

sql_range_step = 5000

sql_attr_uint = tid

sql_attr_uint = digest

sql_attr_uint = displayorder

sql_attr_uint = authorid

sql_attr_uint = special

sql_attr_timestamp =lastpost

sql_query_info = SELECT * FROM pre_forum_thread WHERE tid=$id

}

#threads

index pre_forum_thread

{

source = pre_forum_thread

path = /data0/dzbbs/indexdata/pre_forum_thread #windows下最好用全路径

docinfo = extern

mlock = 0

morphology = none

#charset_dictpath = /usr/local/mmseg32/etc/ #BSD、Linux环境下设置，/符号结尾

#charset_dictpath= etc/ #Windows环境下设置，/符号结尾

#charset_debug = 0

#### 索引的词最小长度

min_word_len = 2

#charset_type = zh_cn.utf-8

html_strip = 1

##### 字符表，注意：如使用这种方式，则sphinx会对中文进行单字切分，

##### 即进行字索引，若要使用中文分词，必须使用其他分词插件如 coreseek，sfc

charset_type = utf-8

chinese_dictionary = /usr/local/sfc22/etc/xdict

min_prefix_len = 0

min_infix_len = 1

ngram_len = 0

}

#threads_minute

source pre_forum_thread_minute : pre_forum_thread

{

#sql_query_pre =

sql_query_pre = SET NAMES UTF8

sql_query_pre = SET SESSION query_cache_type=OFF

sql_query_range = SELECT maxid-1,(SELECT MAX(tid) FROM pre_forum_thread) FROM pre_common_sphinxcounter WHERE indexid=1

}

#threads_minute

index pre_forum_thread_minute : pre_forum_thread

{

source = pre_forum_thread_minute

path = /data0/dzbbs/indexdata/pre_forum_thread_minute #windows下最好用全路径

}

#posts

source pre_forum_post : pre_forum_thread

{

type = mysql

sql_query_pre =

sql_query_pre = SET NAMES UTF8

sql_query_pre = SET SESSION query_cache_type=OFF

sql_query_pre = REPLACE INTO pre_common_sphinxcounter SELECT 2, MAX(pid) FROM pre_forum_post

sql_query = SELECT p.pid AS id,p.tid,p.subject,p.message,t.digest,t.displayorder,t.authorid,t.lastpost,t.special \

FROM pre_forum_post AS p LEFT JOIN pre_forum_thread AS t USING(tid) where p.pid >=$start and p.pid <=$end \

AND p.first=1

sql_query_range = SELECT (SELECT MIN(pid) FROM pre_forum_post),maxid FROM pre_common_sphinxcounter WHERE indexid=2

sql_range_step = 5000

sql_attr_uint = tid

sql_attr_uint = digest

sql_attr_uint = displayorder

sql_attr_uint = authorid

sql_attr_uint = special

sql_attr_timestamp = lastpost

sql_query_info = SELECT * FROM pre_forum_post WHERE pid=$id

}

#posts

index pre_forum_post

{

source = pre_forum_post

path = /data1/dzbbs/indexdata/pre_forum_post #windows下最好用全路径

docinfo = extern

mlock = 0

morphology = none

#charset_dictpath = /usr/local/mmseg32/etc/ #BSD、Linux环境下设置，/符号结尾

#charset_dictpath= etc/ #Windows环境下设置，/符号结尾

#charset_debug = 0

#### 索引的词最小长度

min_word_len = 2

#charset_type = zh_cn.utf-8

html_strip = 0

##### 字符表，注意：如使用这种方式，则sphinx会对中文进行单字切分，

##### 即进行字索引，若要使用中文分词，必须使用其他分词插件如 coreseek，sfc

charset_type = utf-8

chinese_dictionary = /usr/local/sfc22/etc/xdict

min_prefix_len = 0

min_infix_len = 1

ngram_len = 0

}

#pre_forum_post_minute

source pre_forum_post_minute : pre_forum_post

{

sql_query_pre = SET NAMES UTF8

sql_query_pre = SET SESSION query_cache_type=OFF

sql_query_range = SELECT maxid-1,(SELECT MAX(pid) FROM pre_forum_post) FROM pre_common_sphinxcounter WHERE indexid=2

}

#pre_forum_post_minute

index pre_forum_post_minute : pre_forum_post

{

source = pre_forum_post

path = /data0/dzbbs/indexdata/pre_forum_post #windows下最好用全路径

}

#全局indexer定义

indexer

{

mem_limit = 2047M

write_buffer = 64M

}

#searchd服务定义

searchd

{

listen = 3312

read_timeout = 5

max_children = 30

max_matches = 500

listen_backlog = 20

seamless_rotate = 0

preopen_indexes = 0

max_filter_values = 16384

max_filters = 1024

mva_updates_pool = 16M

max_packet_size = 32M

read_buffer = 1M

unlink_old = 1

pid_file = /usr/local/sfc22/var/log/searchd_discuzx.pid #windows下最好用全路径

log = /usr/local/sfc22/var/log/searchd_discuzx.log #windows下最好用全路径

query_log = /usr/local/sfc22/var/log/query_discuzx.log #windows下最好用全路径

}

4 测试sfc 中文分词效果。

/usr/local/sfc22/bin/search -c /usr/local/sfc22/etc/sphinx.conf 分享精彩

words:

1. '分享': 194 documents, 266 hits

2. '精彩': 368 documents, 425 hits

3. '精彩视频': 2 documents, 2 hits

4. '精彩节目': 2 documents, 2 hits

5. '精彩绝伦': 2 documents, 2 hits

6. '精彩纷呈': 7 documents, 7 hits

7. '精彩瞬间': 5 documents, 5 hits

8. '精彩回顾': 2 documents, 3 hits

可以看出sfc 中文分词效果还是比较准确的。

5 启动sfc

/usr/local/sfc22/bin/searchd -c /usr/local/sfc22/etc/sphinx.conf

6 编写主索引和增量索引脚本

1）编写主题表主索引和增量索引以及帖子表增量索引，每20分钟运行一次。

vim /usr/local/bin/bbssearch.sh

#!/bin/bash
#desc: this scripts for sphinx for chinese index
#date:2014.12.23
#testd in CentOS 6.5 x86_64
#saved in /usr/local/bin/bbssearch.sh
#written by coralzd@gmail.com coralzd.blog.51cto.com blog.zjyxh.com
export PATH=/usr/local/sfc22/bin/:/sbin/:$PATH
Y=$(date +%Y)
m=$(date +%m)
d=$(date +%d)
# create thread merge index
echo "-- thread main index start `date +${Y}-${m}-${d}.%H:%M:%S` --"  >> /var/log/sphinx.log
indexer pre_forum_thread_minute --rotate
echo "-- thread merge index end `date +${Y}-${m}-${d}.%H:%M:%S` --"  >> /var/log/sphinx.log
# create post merge index
echo "-- post main index start `date +${Y}-$m-$d.%H:%M:%S` --"  >> /var/log/sphinx.log
indexer pre_forum_post_minute --rotate
echo "-- post merge index end `date +${Y}-$m-$d.%H:%M:%S` --"  >> /var/log/sphinx.log
# main  index + merge index thread
echo "-- thread main merge + index start `date +${Y}-${m}-${d}.%H:%M:%S` --"  >> /var/log/sphinx.log
indexer --merge pre_forum_thread pre_forum_thread_minute --merge-dst-range deleted 0 0 --rotate
echo "-- thread main merge + index end `date +${Y}-${m}-${d}.%H:%M:%S` --"  >> /var/log/sphinx.log

2）编写帖子表主索引，每周运行一次。

vim /usr/local/bin/bbssearch2.sh

#!/bin/bash
#desc: this scripts for sphinx for chinese index
#date:2014.12.23
#testd in CentOS 6.5 x86_64
#saved in /usr/local/bin/bbssearch2.sh
#written by coralzd@gmail.com coralzd.blog.51cto.com blog.zjyxh.com
export PATH=/usr/local/sfc22/bin/:/sbin/:$PATH
Y=$(date +%Y)
m=$(date +%m)
d=$(date +%d)
# create post merge index
echo "-- post main index start `date +${Y}-$m-$d.%H:%M:%S` --"  >> /var/log/sphinx.log
indexer pre_forum_post --rotate
echo "-- post merge index end `date +${Y}-$m-$d.%H:%M:%S` --"  >> /var/log/sphinx.log

将其放到crontab 计划任务。