1、按官方的标准安装，我的系统是CentOS(64位)，默认安装了Python 2.6.6，如果没有安装还需要先安装Python 2.6.6，版本不要错；官方安装地址：http://www.coreseek.cn/products-install/install_on_bsd_linux/ 安装时注意在./config的时候加上对python的支持，因为我们要使用的是Oracle数据源

相关代码：

$ cd csft-3.2.14
$ ./configure –prefix=/usr/local/coreseek  –without-unixodbc –with-mmseg –with-mmseg-includes=/usr/local/mmseg3/include/mmseg/ –with-mmseg-libs=/usr/local/mmseg3/lib/ –with-python
$ make && make install

2、安装cx_Oracle:下载cx_Oracle-5.1.2-11g-py26-1.x86_64.rpm

执行安装包
rpm -ivh cx_Oracle-5.1.2-11g-py26-1.x86_64.rpm

3、下载安装instantclient-basic-linux.x64-11.2.0.3.0.zip，我下载的是zip包，当然我认为也可以下载rpm包，而且不需要自己配置会更合适一些

4、打开~下的.bash_profile，添加下面两句（如果下载的rpm包，则不需要做如下操作

增加以下两行代码
LD_LIBRARY_PATH=/root/instantclient_11_2;
export LD_LIBRARY_PATH

5、在python下测试import cx_Oracle，如没有报错，则说明python连接oracle已经没有错误了

6、配置csft的config文件/root/coreseek-3.2.14/testpack/etc/csft_demo_python_pyoracle.conf

文件如下：
python
{
path = /root/coreseek-3.2.14/testpack/etc/pysource
path = /root/coreseek-3.2.14/testpack/etc/pysource/csft_demo_pyoracle
}
#源定义
source python_oracle
{
type = python
name = csft_demo_pyoracle.MainSource
}
#index定义
index python_oracle
{
source            = python_oracle
path            = /root/coreseek-3.2.14/testpack/var/data/python_oracle
docinfo            = extern
mlock            = 0
morphology        = none
min_word_len        = 1
html_strip                = 0
#中文分词配置，详情请查看：http://www.coreseek.cn/products-install/coreseek_mmseg/
    charset_dictpath = /usr/local/mmseg3/etc/
charset_type        = zh_cn.utf-8
}
#全局index定义
indexer
{
mem_limit            = 128M
}
#searchd服务定义
searchd
{
listen                  =   9312
read_timeout        = 5
max_children        = 30
max_matches            = 1000
seamless_rotate        = 0
preopen_indexes        = 0
unlink_old            = 1
pid_file = /root/coreseek-3.2.14/testpack/var/log/searchd_python_oracle.pid
log = /root/coreseek-3.2.14/testpack/var/log/searchd_python_oracle.log
query_log = /root/coreseek-3.2.14/testpack/var/log/query_python_oracle.log
}

7、设置/root/coreseek-3.2.14/testpack/etc/pysource/csft_demo_pyoracle/__init__.py，

文件内容如下
# -*- coding:utf-8 -*-

from os import path
import os
os.environ[‘NLS_LANG’] = ‘SIMPLIFIED CHINESE_CHINA.UTF8’
import sys
import cx_Oracle

class MainSource(object):
def __init__(self, conf):
self.conf =  conf
self.idx = 0
self.data = []
self.conn = None
self.cur = None

    def GetScheme(self):  #获取结构，docid、文本、整数
return [
(‘id’ , {‘docid’:True, } ),
(‘title’, { ‘type’:’text’} ),
(‘abstract’, { ‘type’:’text’} ),
(‘search_type’, {‘type’:’integer’} ),
]
def GetFieldOrder(self): #字段的优先顺序
return [(‘title’, ‘abstract’)]

def Connected(self):   #如果是数据库，则在此处做数据库连接
if self.conn==None:
self.conn = cx_Oracle.connect(‘yzjs/a123456@10.166.166.222/genome’)
self.cur = self.conn.cursor()
sql = ‘SELECT t.id as id, t.title, t.abstract, t.search_type as search_type from V_SPHINX_INDEX t where t.title is not null and t.abstract is not null’
self.cur.execute(sql)
for rows in self.cur:
item = []
item.append(rows[0])
item.append(rows[1])
item.append(rows[2].read())
item.append(rows[3])
self.data.append(item)
pass

    def NextDocument(self):   #取得每一个文档记录的调用
if self.idx < len(self.data):
item = self.data[self.idx]
self.docid = self.id = item[0] #’docid’:True
self.title = item[1]#.decode(“GBK”).encode(“UTF-8”)
self.abstract = item[2]#.decode(“GBK”).encode(“UTF-8”)
self.search_type = item[3]
self.idx += 1
return True
else:
return False

if __name__ == “__main__”:    #直接访问演示部分
conf = {}
source = MainSource(conf)
source.Connected()
while source.NextDocument():
print “id=%d, subject=%s” % (source.id, source.abstract)#.decode(“UTF-8”))
pass
#eof

8、然后根据下面三段脚本，可以方便索引，启动，热索引数据

索引
/usr/local/coreseek/bin/indexer -c /root/coreseek-3.2.14/testpack/etc/csft_demo_python_pyoracle.conf –all
启动
/usr/local/coreseek/bin/searchd -c /root/coreseek-3.2.14/testpack/etc/csft_demo_python_pyoracle.conf
停止
/usr/local/coreseek/bin/searchd -c /root/coreseek-3.2.14/testpack/etc/csft_demo_python_pyoracle.conf –stop
重编译
/usr/local/coreseek/bin/indexer -c  /root/coreseek-3.2.14/testpack/etc/csft_demo_python_pyoracle.conf –all –rotate

 

9、引入sphinxapi.php，写如下代码测试

$cl = new SphinxClient ();
$cl -> SetServer ( ‘192.168.42.148’, 9312);
$cl -> SetConnectTimeout ( 3 );
$cl -> SetArrayResult ( true );
$cl -> SetMatchMode ( SPH_MATCH_ANY);
$data[‘res’] = $res = $cl -> Query ( $keyword, ‘*’ );