初探站内搜索（中）
接上篇
首先我们配置global.asax.在Application_Start中添加如下代码。创建日志，和定时建立索引任务。
1
protected
void
Application_Start(
object
sender, EventArgs e)
2
{
3
log4net.Config.XmlConfigurator.Configure();
//
log4net起初配置，注意web.config里面改
4
Segment.Init(HttpContext.Current.Server.MapPath(
"
~/PanGu.xml
"
));
//
采用配置文件来初始化
5
6
ISchedulerFactory sf
=
new
StdSchedulerFactory();
//
创建计划工厂
7
IScheduler sched
=
sf.GetScheduler();
//
创建任务的执行者
8
JobDetail job
=
new
JobDetail(
"
job1
"
,
"
group1
"
,
typeof
(Index_Job));
//
IndexJob为实现了IJob接口的类
9
DateTime ts
=
TriggerUtils.GetNextGivenSecondDate(
null
,
1
);
//
1秒后开始执行
10
TimeSpan interval
=
TimeSpan.FromHours (
1
);
//
每隔1小时运行一次，这里的方法是可以改的
11
Trigger trigger
=
new
SimpleTrigger(
"
trigger1
"
,
"
group1
"
,
"
job1
"
,
"
group1
"
, ts,
null
,
12
SimpleTrigger.RepeatIndefinitely, interval);
//
创建触发器
13
sched.AddJob(job,
true
);
14
sched.ScheduleJob(trigger);
15
sched.Start();
//
启动后任务开始计划
16
17
}
然后我们进入Index_Job类中写它的定时任务代码：注意继承IJob接口把计划任务的执行代码写到Execute方法里面。
索引完成后需要索引库的文件都被lucene.net扔到Index文件夹下。
1
public
class
Index_Job:IJob
2
{
3
private
static
ILog log
=
LogManager.GetLogger(
typeof
(Index_Job));
4
5
#region
IJob 成员
6
7
public
void
Execute(JobExecutionContext context)
8
{
9
//
string indexPath = HttpContext.Current.Server.MapPath("~/Index");
10
11
string
indexPath
=
HostingEnvironment.MapPath(
"
~/Index
"
);
12
log.Debug(
"
开始创建索引,索引目录:
"
+
indexPath);
13
FSDirectory directory
=
FSDirectory.Open(
new
DirectoryInfo(indexPath),
new
NativeFSLockFactory());
14
bool
isUpdate
=
IndexReader.IndexExists(directory);
15
log.Debug(
"
索引目录存在状态:
"
+
isUpdate);
16
if
(isUpdate)
17
{
18
if
(IndexWriter.IsLocked(directory))
19
{
20
log.Debug(
"
解锁索引库
"
);
21
IndexWriter.Unlock(directory);
22
}
23
}
24
log.Debug(
"
开始爬文章
"
);
25
IndexWriter writer
=
new
IndexWriter(directory,
new
PanGuAnalyzer(),
!
isUpdate, Lucene.Net.Index.IndexWriter.MaxFieldLength.UNLIMITED);
26
string
siteURL
=
ConfigurationManager.AppSettings[
"
SiteURL
"
];
27
//
for (int i = 1; i <= GetMaxId(siteURL); i++)
28
29
for
(
int
i
=
900
; i
<=
1000
; i
++
)
30
{
31
log.Debug(
"
开始爬编号为
"
+
i.ToString()
+
"
的帖子
"
);
32
try
33
{
34
WebClient wc
=
new
WebClient();
35
wc.Encoding
=
Encoding.UTF8;
//
36
string
url
=
siteURL
+
"
showtopic-
"
+
i
+
"
.aspx
"
;
37
//
string url = "
http://localhost
:8081/showtopic-" + i + ".aspx";
38
string
txt
=
wc.DownloadString(url);
39
HTMLDocumentClass htmldoc
=
new
HTMLDocumentClass();
40
htmldoc.designMode
=
"
on
"
;
//
这样就不解析javascript了
41
htmldoc.IHTMLDocument2_write(txt);
42
string
title
=
htmldoc.title;
43
string
bodyText
=
htmldoc.body.innerText;
44
45
writer.DeleteDocuments(
new
Term(
"
url
"
, url));
//
删除旧的数据，以url键为主键，这样就避免重复
46
47
Document document
=
new
Document();
48
document.Add(
new
Field(
"
url
"
, url, Field.Store.YES, Field.Index.NOT_ANALYZED));
49
document.Add(
new
Field(
"
title
"
, title, Field.Store.YES, Field.Index.NOT_ANALYZED));
50
document.Add(
new
Field(
"
body
"
, bodyText, Field.Store.YES, Field.Index.ANALYZED, Lucene.Net.Documents.Field.TermVector.WITH_POSITIONS_OFFSETS));
51
52
writer.AddDocument(document);
53
log.Debug(
"
爬编号为
"
+
i.ToString()
+
"
的帖子结束
"
);
54
}
55
catch
(Exception ex)
56
{
57
log.Error(
"
爬编号为
"
+
i.ToString()
+
"
的帖子发生异常
"
, ex);
58
}
59
}
60
log.Debug(
"
结束索引，开始关闭Writer和Directory
"
);
61
writer.Close();
62
directory.Close();
63
log.Debug(
"
关闭Writer和Directiory完成
"
);
64
//
ClientScript.RegisterStartupScript(GetType(), "alert", "alert('索引完成')", true);
65
}
66
67
#endregion
68
69
70
private
static
int
GetMaxId(
string
siteURL)
//
获得最新的id帖子编号
71
{
72
WebClient wc
=
new
WebClient();
73
wc.Encoding
=
Encoding.UTF8;
74
string
html
=
wc.DownloadString(siteURL
+
"
tools/rss.aspx
"
);
75
XDocument doc
=
XDocument.Parse(html);
76
string
link
=
doc.Descendants(
"
item
"
).First().Element(
"
link
"
).Value;
77
System.Text.RegularExpressions.Regex regex
=
new
Regex(
@"
showtopic-(\d+)
"
);
78
Match match
=
regex.Match(link);
79
string
id
=
match.Groups[
1
].Value;
80
81
return
Convert.ToInt32(id);
82
}
83
84
85
86
}
Default.aspx页面前台html代码（表单form中）：
1
<
form
id
="form1"
action
="Default.aspx"
method
="get"
>
2
<
div
>
3
4
<
input
type
="text"
id
="txtKw"
value
="<%=Request["
kw"] %
>
" name ="kw" />
5
<
script
type
="text/javascript"
>
6
$(
"
#txtKw
"
).autocomplete({
7
source:
"
SearchSuggestion.ashx
"
, select:
function
(e, ui) {
8
$(
"
#txtKw
"
).val(ui.item.value);
9
$(
"
#sb
"
).click();
10
}
//
自动发出Ajax请求
11
});
12
</
script
>
13
14
<
input
type
="submit"
id
="sb"
value
="搜索"
/>
15
16
</
div
>
17
<
asp:Repeater
ID
="RepeaterResult"
runat
="server"
>
18
<
ItemTemplate
>
19
<
a
href
='<%#Eval("URL")%
>
' >
<%
#
Eval
(
"
TITLE
"
)
%>
</
a
>
20
<
br
/>
21
<
p
>
22
<%
#
Eval
(
"
BODY
"
)
%>
23
</
p
>
24
</
ItemTemplate
>
25
</
asp:Repeater
>
26
</
form
>
注意把viewstate禁用了，防止最后生成的客户端html代码中有一大堆viewstate的东东，显得我们不专业 呵呵~（EnableViewState ="false"）
我们知道一旦禁用viewstate所有跑在服务端的基本控件都不能用了，那些数控绑定和链接控件除外，所以我们回归原始的html。用get方法提交表单。
因此后台代码我们就在pageload方法中实现，通过用户输入的关键字Request["kw"]是否为空提交表单后判断是否加载运行。后台处理代码如下：（其中有注释我就不详细解释了）
1
public
partial
class
_Default : System.Web.UI.Page
2
{
3
protected
void
Page_Load(
object
sender, EventArgs e)
4
{
5
if
(
string
.IsNullOrEmpty(Request[
"
kw
"
]))
6
{
7
return
;
8
}
9
else
10
{
11
12
string
kw
=
Request[
"
kw
"
];
13
new
SearchLogTableAdapter().Insert(Guid.NewGuid(), DateTime.Now, Request.UserHostAddress, kw);
//
把关键词插入数据库
14
15
string
indexPath
=
Server.MapPath(
"
~/Index
"
);
//
获得要搜索的文本路径（已经被lucene建立好的文本的索引文件）
16
FSDirectory directory
=
FSDirectory.Open(
new
DirectoryInfo(indexPath),
new
NativeFSLockFactory());
17
IndexReader indexReader
=
IndexReader.Open(directory,
true
);
18
IndexSearcher searcher
=
new
IndexSearcher(indexReader);
//
加入搜索者
19
PhraseQuery query
=
new
PhraseQuery();
//
查询条件
20
foreach
(
string
word
in
segString(kw))
//
分词
21
{
22
query.Add(
new
Term(
"
body
"
, word));
//
加入分词的查询条件
23
24
}
25
query.SetSlop(
1000
);
//
相邻1000个字有效
26
27
TopScoreDocCollector collector
=
TopScoreDocCollector.create(
1000
,
true
);
//
创建收集器最多收集1000个文本
28
searcher.Search(query,
null
, collector);
//
开始查询，使用query条件，结果放入collector中
29
TopDocs topDocs
=
collector.TopDocs();
//
获得结果
30
int
sum
=
collector.GetTotalHits();
//
得到结果条数
31
List
<
SearchResult
>
list
=
new
List
<
SearchResult
>
();
32
foreach
(ScoreDoc scoreDoc
in
topDocs.scoreDocs)
33
{
34
35
int
docId
=
scoreDoc.doc;
//
拿到搜到的文档ID
36
Document document
=
searcher.Doc(docId);
//
根据文档ID创建DOCUMENT
37
string
url
=
document.Get(
"
url
"
);
38
string
title
=
document.Get(
"
title
"
);
39
string
body
=
document.Get(
"
body
"
);
40
41
SimpleHTMLFormatter simpleHtmlFormatter
=
new
SimpleHTMLFormatter(
"
<font color='red'>
"
,
"
</font>
"
);
//
设置高亮,还可以实现其他功能。提示之类的~~
42
Highlighter highlighter
=
new
Highlighter(simpleHtmlFormatter,
new
Segment());
43
highlighter.FragmentSize
=
200
;
44
body
=
highlighter.GetBestFragment(kw, body);
//
把关键字设为高亮
45
46
SimpleHTMLFormatter simpleHtmlFormatter1
=
new
SimpleHTMLFormatter(
"
<font style='background-color:red'>
"
,
"
</font>
"
);
//
把标题背景设为红
47
Highlighter highlighter1
=
new
Highlighter(simpleHtmlFormatter1,
new
Segment());
48
highlighter1.FragmentSize
=
200
;
49
title
=
highlighter1.GetBestFragment(title, title);
50
51
SearchResult result
=
new
SearchResult()
52
{
53
URL
=
url,
54
TITLE
=
title,
55
BODY
=
body
56
};
57
list.Add(result);
58
59
}
60
searcher.Close();
61
indexReader.Close();
62
directory.Close();
63
RepeaterResult.DataSource
=
list;
64
RepeaterResult.DataBind();
65
}
66
67
}
68
private
static
string
[] segString(
string
s)
69
{
70
Segment segment
=
new
Segment();
71
return
(from wordInfo
in
segment.DoSegment(s) select wordInfo.Word).ToArray();
72
}
73
74
}
75
public
class
SearchResult
76
{
77
public
string
URL {
get
;
set
; }
78
public
string
TITLE {
get
;
set
; }
79
public
string
BODY {
get
;
set
; }
80
}
接下去我们实现输入关键词自动补全。用JQueryUI和AJAX请求后台关键词效果。
未完，待续。。。