首页 > 代码库 > 集成Nutch/Hbase/Solr构建搜索引擎之二：内容分析
集成Nutch/Hbase/Solr构建搜索引擎之二：内容分析
请先参见“集成Nutch/Hbase/Solr构建搜索引擎之一：安装及运行”，搭建测试环境
http://blog.csdn.net/jediael_lu/article/details/37329731
一、被索引的域 Schema.xml
在使用solr对Nutch抓取到的网页进行索引时，schema.xml被改变成以下内容。
文件中指定了哪些域被索引、存储等内容。
<?xml version="1.0" encoding="UTF-8" ?> <schema name="nutch" version="1.5"> <types> <fieldType name="string" class="solr.StrField" sortMissingLast="true" omitNorms="true"/> <fieldType name="long" class="solr.TrieLongField" precisionStep="0" omitNorms="true" positionIncrementGap="0"/> <fieldType name="float" class="solr.TrieFloatField" precisionStep="0" omitNorms="true" positionIncrementGap="0"/> <fieldType name="date" class="solr.TrieDateField" precisionStep="0" omitNorms="true" positionIncrementGap="0"/> <fieldType name="text" class="solr.TextField" positionIncrementGap="100"> <analyzer> <tokenizer class="solr.WhitespaceTokenizerFactory"/> <filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords.txt"/> <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1" catenateWords="1" catenateNumbers="1" catenateAll="0" splitOnCaseChange="1"/> <filter class="solr.LowerCaseFilterFactory"/> <filter class="solr.RemoveDuplicatesTokenFilterFactory"/> </analyzer> </fieldType> <fieldType name="url" class="solr.TextField" positionIncrementGap="100"> <analyzer> <tokenizer class="solr.StandardTokenizerFactory"/> <filter class="solr.LowerCaseFilterFactory"/> <filter class="solr.WordDelimiterFilterFactory" generateWordParts="1" generateNumberParts="1"/> </analyzer> </fieldType> </types> <fields> <field name="id" type="string" stored="true" indexed="true"/> <field name="_version_" type="long" indexed="true" stored="true"/> <!-- core fields --> <field name="batchId" type="string" stored="true" indexed="false"/> <field name="digest" type="string" stored="true" indexed="false"/> <field name="boost" type="float" stored="true" indexed="false"/> <!-- fields for index-basic plugin --> <field name="host" type="url" stored="false" indexed="true"/> <field name="url" type="url" stored="true" indexed="true" required="true"/> <field name="content" type="text" stored="false" indexed="true"/> <field name="title" type="text" stored="true" indexed="true"/> <field name="cache" type="string" stored="true" indexed="false"/> <field name="tstamp" type="date" stored="true" indexed="false"/> <!-- fields for index-anchor plugin --> <field name="anchor" type="string" stored="true" indexed="true" multiValued="true"/> <!-- fields for index-more plugin --> <field name="type" type="string" stored="true" indexed="true" multiValued="true"/> <field name="contentLength" type="long" stored="true" indexed="false"/> <field name="lastModified" type="date" stored="true" indexed="false"/> <field name="date" type="date" stored="true" indexed="true"/> <!-- fields for languageidentifier plugin --> <field name="lang" type="string" stored="true" indexed="true"/> <!-- fields for subcollection plugin --> <field name="subcollection" type="string" stored="true" indexed="true" multiValued="true"/> <!-- fields for feed plugin (tag is also used by microformats-reltag)--> <field name="author" type="string" stored="true" indexed="true"/> <field name="tag" type="string" stored="true" indexed="true" multiValued="true"/> <field name="feed" type="string" stored="true" indexed="true"/> <field name="publishedDate" type="date" stored="true" indexed="true"/> <field name="updatedDate" type="date" stored="true" indexed="true"/> <!-- fields for creativecommons plugin --> <field name="cc" type="string" stored="true" indexed="true" multiValued="true"/> <!-- fields for tld plugin --> <field name="tld" type="string" stored="false" indexed="false"/> </fields> <uniqueKey>id</uniqueKey> <defaultSearchField>content</defaultSearchField> <solrQueryParser defaultOperator="OR"/> </schema>分析上述文件，主要指定了以下内容：
（1）FiledType：域的类型
（2）Field：哪些域被索引、存储等，以及这个域是什么类型。
（3）uniqueKey：哪个域作为id，即文章的唯一标识。
（4）defaultSearchField：默认的搜索域
（5）solrQueryParser：OR，即使用OR来构建Query。
二、Nutch抓取的基本流程
./bin/crawl urls csdnitpub http://182.92.160.44:8983/solr/ 5
crawl <seedDir> <crawlID> <solrURL> <numberOfRounds>
<seedDir>：放置种子文件的目录
<crawlID> ：这个抓取任务的ID
<solrURL>：用于索引及搜索的solr地址
<numberOfRounds>：迭代次数，即抓取深度
一个nutch抓取流程如下：（1）InjectorJob开始第一个迭代（2）GeneratorJob（3）FetcherJob（4）ParserJob（5）DbUpdaterJob（6）SolrIndexerJob开始第二个迭代（2）GeneratorJob（3）FetcherJob（4）ParserJob（5）DbUpdaterJob（6）SolrIndexerJob
开始第三个迭代……
一个典型任务的日志如下：InjectorJob: starting at 2014-07-08 10:41:27
InjectorJob: Injecting urlDir: urls
InjectorJob: Using class org.apache.gora.hbase.store.HBaseStore as the Gora storage class.
InjectorJob: total number of urls rejected by filters: 0
InjectorJob: total number of urls injected after normalization and filtering: 2
Injector: finished at 2014-07-08 10:41:32, elapsed: 00:00:05
Tue Jul 8 10:41:33 CST 2014 : Iteration 1 of 5
Generating batchId
Generating a new fetchlist
GeneratorJob: starting at 2014-07-08 10:41:34
GeneratorJob: Selecting best-scoring urls due for fetch.
GeneratorJob: starting
GeneratorJob: filtering: false
GeneratorJob: normalizing: false
GeneratorJob: topN: 50000
GeneratorJob: finished at 2014-07-08 10:41:39, time elapsed: 00:00:05
GeneratorJob: generated batch id: 1404787293-26339
Fetching :
FetcherJob: starting
FetcherJob: batchId: 1404787293-26339
Fetcher: Your ‘http.agent.name‘ value should be listed first in ‘http.robots.agents‘ property.
FetcherJob: threads: 50
FetcherJob: parsing: false
FetcherJob: resuming: false
FetcherJob : timelimit set for : 1404798101129
Using queue mode : byHost
Fetcher: threads: 50
QueueFeeder finished: total 2 records. Hit by time limit :0
fetching http://www.csdn.net/ (queue crawl delay=5000ms)
Fetcher: throughput threshold: -1
Fetcher: throughput threshold sequence: 5
fetching http://www.itpub.net/ (queue crawl delay=5000ms)
-finishing thread FetcherThread47, activeThreads=48
-finishing thread FetcherThread46, activeThreads=47
-finishing thread FetcherThread45, activeThreads=46
-finishing thread FetcherThread44, activeThreads=45
-finishing thread FetcherThread43, activeThreads=44
-finishing thread FetcherThread42, activeThreads=43
-finishing thread FetcherThread41, activeThreads=42
-finishing thread FetcherThread40, activeThreads=41
-finishing thread FetcherThread39, activeThreads=40
-finishing thread FetcherThread38, activeThreads=39
-finishing thread FetcherThread37, activeThreads=38
-finishing thread FetcherThread36, activeThreads=37
-finishing thread FetcherThread35, activeThreads=36
-finishing thread FetcherThread34, activeThreads=35
-finishing thread FetcherThread33, activeThreads=34
-finishing thread FetcherThread32, activeThreads=33
-finishing thread FetcherThread31, activeThreads=32
-finishing thread FetcherThread30, activeThreads=31
-finishing thread FetcherThread29, activeThreads=30
-finishing thread FetcherThread48, activeThreads=29
-finishing thread FetcherThread27, activeThreads=29
-finishing thread FetcherThread26, activeThreads=28
-finishing thread FetcherThread25, activeThreads=27
-finishing thread FetcherThread24, activeThreads=26
-finishing thread FetcherThread23, activeThreads=25
-finishing thread FetcherThread22, activeThreads=24
-finishing thread FetcherThread21, activeThreads=23
-finishing thread FetcherThread20, activeThreads=22
-finishing thread FetcherThread19, activeThreads=21
-finishing thread FetcherThread18, activeThreads=20
-finishing thread FetcherThread17, activeThreads=19
-finishing thread FetcherThread16, activeThreads=18
-finishing thread FetcherThread15, activeThreads=17
-finishing thread FetcherThread14, activeThreads=16
-finishing thread FetcherThread13, activeThreads=15
-finishing thread FetcherThread12, activeThreads=14
-finishing thread FetcherThread11, activeThreads=13
-finishing thread FetcherThread10, activeThreads=12
-finishing thread FetcherThread9, activeThreads=11
-finishing thread FetcherThread8, activeThreads=10
-finishing thread FetcherThread7, activeThreads=9
-finishing thread FetcherThread5, activeThreads=8
-finishing thread FetcherThread4, activeThreads=7
-finishing thread FetcherThread3, activeThreads=6
-finishing thread FetcherThread2, activeThreads=5
-finishing thread FetcherThread49, activeThreads=4
-finishing thread FetcherThread6, activeThreads=3
-finishing thread FetcherThread28, activeThreads=2
-finishing thread FetcherThread0, activeThreads=1
fetch of http://www.itpub.net/ failed with: java.io.IOException: unzipBestEffort returned null
-finishing thread FetcherThread1, activeThreads=0
0/0 spinwaiting/active, 2 pages, 1 errors, 0.4 0 pages/s, 93 93 kb/s, 0 URLs in 0 queues
-activeThreads=0
FetcherJob: done
Parsing :
ParserJob: starting
ParserJob: resuming: false
ParserJob: forced reparse: false
ParserJob: batchId: 1404787293-26339
Parsing http://www.csdn.net/
http://www.csdn.net/ skipped. Content of size 92777 was truncated to 59561
Parsing http://www.itpub.net/
ParserJob: success
CrawlDB update for csdnitpub
DbUpdaterJob: starting
DbUpdaterJob: done
Indexing csdnitpub on SOLR index -> http://182.92.160.44:8983/solr/
SolrIndexerJob: starting
SolrIndexerJob: done.
SOLR dedup -> http://182.92.160.44:8983/solr/
Tue Jul 8 10:42:18 CST 2014 : Iteration 2 of 5
Generating batchId
Generating a new fetchlist
GeneratorJob: starting at 2014-07-08 10:42:19
GeneratorJob: Selecting best-scoring urls due for fetch.
GeneratorJob: starting
GeneratorJob: filtering: false
GeneratorJob: normalizing: false
GeneratorJob: topN: 50000
GeneratorJob: finished at 2014-07-08 10:42:25, time elapsed: 00:00:05
GeneratorJob: generated batch id: 1404787338-30453
Fetching :
FetcherJob: starting
FetcherJob: batchId: 1404787338-30453
Fetcher: Your ‘http.agent.name‘ value should be listed first in ‘http.robots.agents‘ property.
FetcherJob: threads: 50
FetcherJob: parsing: false
FetcherJob: resuming: false
FetcherJob : timelimit set for : 1404798146676
Using queue mode : byHost
Fetcher: threads: 50
QueueFeeder finished: total 0 records. Hit by time limit :0
附一些说明：
Crawling the Web is already explained above. You can add more URLs in the seed.txt file and crawl the same.
When a user invokes a crawling command in Apache Nutch 1.x, CrawlDB is generated by Apache Nutch which is nothing but a directory and which contains details about crawling. In Apache 2.x, CrawlDB is not present. Instead, Apache Nutch keeps all the crawling data directly in the database. In our case, we have used Apache HBase, so all crawling data would go inside Apache HBase. The following are details of how each function of crawling works.
A crawling cycle has four steps, in which each is implemented as a Hadoop MapReduce job:
? GeneratorJob
? FetcherJob
? ParserJob (optionally done while fetching using ‘fetch.parse‘)
? DbUpdaterJob
Additionally, the following processes need to be understood:
? InjectorJob
? Invertlinks
? Indexing with Apache Solr
First of all, the job of an Injector is to populate initial rows for the web table. The InjectorJob will initialize crawldb with the URLs that we have provided. We need to run the InjectorJob by providing certain URLs, which will then be inserted into crawlDB.
Then the GeneratorJob will use these injected URLs and perform the operation. The table which is used for input and output for these jobs is called webpage, in which
every row is a URL (web page). The row key is stored as a URL with reversed host components so that URLs from the same TLD and domain can be kept together and
form a group. In most NoSQL stores, row keys are sorted and give an advantage.
Using specific rowkey filtering, scanning will be faster over a subset, rather than scanning over the entire table. Following are the examples of rowkey listing:
? org.apache..www:http/
? org.apache.gora:http/
Let‘s define each step in depth so that we can understand crawling step-by-step.
Apache Nutch contains three main directories, crawlDB, linkdb, and a set of segments. crawlDB is the directory which contains information about every URL that is known to Apache Nutch. If it is fetched, crawlDB contains the details when it was fetched. The linkdatabase or linkdb contains all the links to each URL which will include source URL and also the anchor text of the link. A set of segments is a URL set, which is fetched as a unit. This directory will contain the following subdirectories:
? A crawl_generate job will be used for a set of URLs to be fetched
? A crawl_fetch job will contain the status of fetching each URL
? A content will contain the content of rows retrieved from every URL
Now let‘s understand each job of crawling in detail.
三、与抓取相关的一些hbase操作
（1）进入bhase shell
[root@jediael44 hbase-0.90.4]# ./bin/hbase shell
HBase Shell; enter ‘help<RETURN>‘ for list of supported commands.
Type "exit<RETURN>" to leave the HBase Shell
Version 0.90.4, r1150278, Sun Jul 24 15:53:29 PDT 2011
（2）查看所有表
hbase(main):001:0> list
TABLE
20140710_webpage
TestCrawl_webpage
csdnitpub_webpage
stackoverflow_webpage
4 row(s) in 0.7620 seconds
（3）查看某个表的结构
hbase(main):002:0> describe ‘20140710_webpage‘
DESCRIPTION ENABLED
{NAME => ‘20140710_webpage‘, FAMILIES => [{NAME => ‘f‘, BLOOMFILTER => ‘NONE‘, true
REPLICATION_SCOPE => ‘0‘, VERSIONS => ‘1‘, COMPRESSION => ‘NONE‘, TTL => ‘21474
83647‘, BLOCKSIZE => ‘65536‘, IN_MEMORY => ‘false‘, BLOCKCACHE => ‘true‘}, {NAM
E => ‘h‘, BLOOMFILTER => ‘NONE‘, REPLICATION_SCOPE => ‘0‘, VERSIONS => ‘1‘, COM
PRESSION => ‘NONE‘, TTL => ‘2147483647‘, BLOCKSIZE => ‘65536‘, IN_MEMORY => ‘fa
lse‘, BLOCKCACHE => ‘true‘}, {NAME => ‘il‘, BLOOMFILTER => ‘NONE‘, REPLICATION_
SCOPE => ‘0‘, VERSIONS => ‘1‘, COMPRESSION => ‘NONE‘, TTL => ‘2147483647‘, BLOC
KSIZE => ‘65536‘, IN_MEMORY => ‘false‘, BLOCKCACHE => ‘true‘}, {NAME => ‘mk‘, B
LOOMFILTER => ‘NONE‘, REPLICATION_SCOPE => ‘0‘, VERSIONS => ‘1‘, COMPRESSION =>
‘NONE‘, TTL => ‘2147483647‘, BLOCKSIZE => ‘65536‘, IN_MEMORY => ‘false‘, BLOCK
CACHE => ‘true‘}, {NAME => ‘mtdt‘, BLOOMFILTER => ‘NONE‘, REPLICATION_SCOPE =>
‘0‘, VERSIONS => ‘1‘, COMPRESSION => ‘NONE‘, TTL => ‘2147483647‘, BLOCKSIZE =>
‘65536‘, IN_MEMORY => ‘false‘, BLOCKCACHE => ‘true‘}, {NAME => ‘ol‘, BLOOMFILTE
R => ‘NONE‘, REPLICATION_SCOPE => ‘0‘, VERSIONS => ‘1‘, COMPRESSION => ‘NONE‘,
TTL => ‘2147483647‘, BLOCKSIZE => ‘65536‘, IN_MEMORY => ‘false‘, BLOCKCACHE =>
‘true‘}, {NAME => ‘p‘, BLOOMFILTER => ‘NONE‘, REPLICATION_SCOPE => ‘0‘, VERSION
S => ‘1‘, COMPRESSION => ‘NONE‘, TTL => ‘2147483647‘, BLOCKSIZE => ‘65536‘, IN_
MEMORY => ‘false‘, BLOCKCACHE => ‘true‘}, {NAME => ‘s‘, BLOOMFILTER => ‘NONE‘,
REPLICATION_SCOPE => ‘0‘, VERSIONS => ‘1‘, COMPRESSION => ‘NONE‘, TTL => ‘21474
83647‘, BLOCKSIZE => ‘65536‘, IN_MEMORY => ‘false‘, BLOCKCACHE => ‘true‘}]}
1 row(s) in 0.0880 seconds
（4）查看某列的HTML文件
hbase(main):010:0* get ‘20140710_webpage‘, ‘ar.com.admival.www:http/‘, ‘f:cnt‘
COLUMN CELL
f:cnt timestamp=1404983104070, value=http://www.mamicode.com/x0D/x0A o="urn:schemas-microsoft-com:office:office" xmlns="http://www.w3.org/TR/REC-html40">\x0D\x0
A\x0D\x0A<head>\x0D\x0A<meta http-equiv="Content-Type" content="text/html; charset=windows-
1252">\x0D\x0A<meta http-equiv="Content-Language" content="es">\x0D\x0A\x0D\x0A<title>CAMBI
O DE CHEQUES | cambio de cheque | cheque | cheques | compra de cheques de terceros | cambio
cheques | cheques no a la orden </title>\x0D\x0A<link rel="shortcut icon" ref="favivon.ico
" />\x0D\x0A<meta name="GENERATOR" content="Microsoft FrontPage 5.0">\x0D\x0A<meta name="Pr
ogId" content="FrontPage.Editor.Document">\x0D\x0A<meta name=keywords content="cambio de ch
eques, compro cheques, canje de cheques, Compro cheques de terceros, cambio cheques">\x0D\x
0A<meta name="description" http-equiv="description" content="Admival.com.ar cambia por efec
tivo sus cheques de terceros, posdatados, al d\xECa, de Sociedades, personales, cheques no
a la orden, cheques de indemnizaci\xF3n por despido, etc. " /> \x0D\x0A<meta name="robots"
content="index,follow">\
未完……
其中行的名称可以通过scan ‘20140710_webpage‘得到。