【solr专题之四】关于VelocityResponseWriter

阅读量：5035 次

发布时间：2019-06-12

本文共 14119 字，大约阅读时间需要 47 分钟。

一、关于Velocity的基本配置

在Solr中，可以以多种方式返回搜索结果，如单纯的文本回复（XML、JSON、CSV等），也可以返回velocity，js等格式。而VelocityResponseWriter就是用于将返回velocity类型文本，以便直接用于结果呈现。

在Solr提供的example，其中的一个RequestHandler--/browse，使用了VelocityResponseWriter。其配置如下：


         
            
      
       explicit
             
             
      
       velocity
             
      
       browse
             
      
       layout
             
      
       Solritas_test
             
             
      
       edismax
             
      
                 text^0.5 features^1.0 name^1.2 sku^1.5 id^10.0 manu^1.1 cat^1.4          title^10.0 description^5.0 keywords^5.0 author^2.0 resourcename^1.0       
             
      
       text
             
      
       100%
             
      
       *:*
             
      
       10
             
      
       *,score
             
      
                text^0.5 features^1.0 name^1.2 sku^1.5 id^10.0 manu^1.1 cat^1.4         title^10.0 description^5.0 keywords^5.0 author^2.0 resourcename^1.0       
             
      
       text,features,name,sku,id,manu,cat,title,description,keywords,author,resourcename
             
      
       3
             
             
      
       on
             
      
       cat
             
      
       manu_exact
             
      
       content_type
             
      
       author_s
             
      
       ipod
             
      
       GB
             
      
       1
             
      
       cat,inStock
             
      
       after
             
      
       price
             
      
       0
             
      
       600
             
      
       50
             
      
       popularity
             
      
       0
             
      
       10
             
      
       3
             
      
       manufacturedate_dt
             
      
       NOW/YEAR-10YEARS
             
      
       NOW
             
      
       +1YEAR
             
      
       before
             
      
       after
             
             
      
       on
             
      
       content features title name
             
      
       html
             
      
       
      
             
              
      
       0
             
      
       title
             
      
       0
             
      
       name
             
      
       3
             
      
       200
             
      
       content
             
      
       750
             
             
      
       on
             
      
       false
             
      
       5
             
      
       2
             
      
       5
             
      
       true
             
      
       true
             
      
       5
             
      
       3
           
          
          
            
      
       spellcheck

关于velocity这个writer的定义如下：

处理一个流程的步骤如下：

1、根据请求url查找相关的配置

请求的url为：

则在solrConfig.xml中查找 /browse的配置，可以得出上述所示的结果。

2、其中有关velocity的内容如下：


          
             
       
        explicit
              
              
       
        velocity
              
       
        browse
              
       
        layout
              
       
        Solritas_test

从上述定义中开始分别查找显示层的内容(vm文件)与处理类的内容(wt的实现类。)

3、定位显示内容模板

根据v.template属性，定义到文件browse.vm，注意在配置中省略了后缀名vm。

由于存在v.layout属性，因此，此属性的值将作为模板，而

v.template中的内容将作为$content的内容。

v.layout: Template name that wraps main template (v.template). Main template renders to a $content that can be used in layout template.

layout.vm的内容如下：

#** *  Overall HTML page layout *#  #parse("head.vm")      
      
       Solr Admin
      
    
      
             #parse("header.vm")    
      
    
      
             #parse("tabs.vm")    
      
    
      
             $content    
      
    
      
             #parse("footer.vm")

4、定位ResponseWriter

从第2步的结果知道，使用的velocity，然后查找这个wt的定义，可以得到

即VelocityResponseWriter实现类，其定义如下：

public class VelocityResponseWriterextends Objectimplements QueryResponseWriter

与Solr返回Velocity相关的类只有4个：

以下是关于VelocityResponseWriter的官方说明：http://wiki.apache.org/solr/VelocityResponseWriter

Introduction

VelocityResponseWriter (aka ) enables Solr to respond with content generated from templates. Along with technologies like SolrJS, this makes Solr itself capable of driving sophisticated search interfaces without the need for an intermediate application server between the browser and Solr.

See for more information.

Contents

Instructions to use, Solr 1.4+

These steps will get you up and running for the examples below:

Download and install Solr 1.4.x

Fire up Solr: cd example; java -Dsolr.solr.home=../contrib/velocity/src/main/solr/ -jar start.jar

Index sample docs: cd example/exampledocs; java -jar post.jar *.xml

Hit the examples below...

Sample Usage

Renders browse.vm from conf/velocity. Faceted navigation included.

Renders browse.vm, but overrides the header.vm from conf/velocity with the specified value.

Renders browse.vm, adding in explanation views per hit, and a Velocity context dump at the end.

Using the VelocityResponseWriter in Solr Core

The VelocityResponseWriter is still a contrib component in Solr 1.4.x. In order to use it with the core distributions the following steps need to be followed:

The following jars need to be copied from contrib/velocity/src/main/solr/lib/ to $SOLR_HOME/lib:

apache-solr-velocity-1.4-dev.jar

velocity-1.6.1.jar

velocity-tools-2.0-beta3.jar

commons-beanutils-1.7.0.jar

commons-collections-3.2.1.jar

The VelocityResponseWriter uses a more recent version of the commons lang jar than the current version of Solr core, so the jar commons-lang-2.4.jar from .../contrib/velocity/src/main/solr/lib/ should replace $SOLR_HOME/lib/commons-lang-2.1.jar

Add some configuration for this to solrconfig.xml like this:

Set up a in solrconfig.xml:


         
            
      
       browse
             
      
       velocity.properties
             
      
       text/html;charset=UTF-8
             
      
       Solritas
             
      
       velocity
             
      
       dismax
             
      
       *:*
             
      
       10
             
      
       *,score
             
      
       on
             
      
       title
             
      
       1
             
      
                 text^0.5 title^1.5

Copy the .../contrib/velocity/src/main/solr/conf/velocity directory to $SOLR_HOME/conf/. This directory contains the Velocity templates that will be needed by the VelocityResponseWriter, and also a style sheet, main.css. The templates and style sheet can be edited to customize the display.

Instructions to use, Solr 4.0+

These steps will get you up and running for the examples below:

Check out Solr trunk: svn co

Build Solr: ant clean example

Fire up Solr: cd example; java -jar start.jar

Index sample docs: cd example/exampledocs; java -jar post.jar *.xml

Hit the examples below...

Sample Usage

Renders browse.vm from conf/velocity. Faceted navigation included.

Renders browse.vm, but overrides the header.vm from conf/velocity with the specified value.

Renders browse.vm, adding in explanation views per hit, and a Velocity context dump at the end.

Options (All Versions)

v.template: template name to use, without the .vm suffix. If not specified, "default"[.vm] will be used.

v.template.<name>: overrides a file system template

debugQuery: if true, default view displays explanations for each hit and additional debugging information in the footer.

v.json: Escapes and wraps Velocity generated response with v.json parameter as a function.

v.layout: Template name that wraps main template (v.template). Main template renders to a $content that can be used in layout template.

v.base_dir: overwrites default template load path (conf/velocity/).

v.properties: specifies a Velocity properties file to be applied, found using the Solr resource loader mechanism. If not specified, no .properties file is loaded. Example: v.properties=velocity.properties where velocity.properties can be found using Solr's resource loader mechanism, for example in the conf/ directory (not conf/velocity which is for templates only). The .properties file could also be located inside a JAR in the lib/ directory, or other locations.

v.contentType: sets the value of the HTTP response's Content-Type header (in case (x)html pages should be UTF-8 (instead of ISO-8859-1) encoded, make sure you set this option to text/xml;charset=UTF-8 (for XHTML) and text/html;charset=UTF-8 (for HTML), respectively)

Velocity Context

esc: a Velocity instance

date: a Velocity instance

list: a Velocity instance

math: a Velocity instance

number: a Velocity instance

page: a instance. page only is added to the context when response is a .

request: a

response: a most of the time, but in some cases where doesn't like the request handlers output (, for example, causes a parsing "response") the response will be a object.

sort: a Velocity instance

TODO

Ajax suggest

Integrate/adapt to SolrJS

Tie in SIMILE Timeline and SIMILE Exhibit

Add links in default footer to this wiki page, the Solr request as XML format, and SOLR-620

Fix multi-valued fields issue, and fl parameter usage.

Work on "dist" target so this works easily with a nightly build.

Make Velocity tools and engine configuration pluggable

二、Velocity文件定位过程

1、根据上述分析，首先定位layout.xml

  #parse("head.vm")      
    
     Solr Admin
    
    
    
           #parse("header.vm")    
    
    
    
           #parse("tabs.vm")    
    
    
    
           $content    
    
    
    
           #parse("footer.vm")

2、其中content的内容即为browse.vm


    
       #parse("pagination_top.vm")
    
## Show Error Message, if any
    
       #parse("error.vm")
    
## Render Results, actual matching docs
    
       #parse("results_list.vm")
    
    
       #parse("pagination_bottom.vm")

3、browse.vm中最主要的搜索结果为results_list.vm

#** *  Render the main Results List *### Usually displayed inside 
    
     #if($response.response.get('grouped'))  #foreach($grouping in $response.response.get('grouped'))    #parse("hit_grouped.vm")  #end#else  #foreach($doc in $response.results)    #parse("hit.vm")    ## Can get an extremely simple view of the doc    ## which might be nicer for debugging    ##parse("hit_plain.vm")  #end#end

一般情况下使用hit.vm作呈现，它对页面作了一些美工。

在某些情况下，如debug的时候，就使用hit_plain.vm进行呈现，此时将所有的属性呈现出来。

二者的对比效果如下：

4、先查看hit_plain.vm

#** *  An extremely plain / debug version of hit.vm *#
    
       ## For each field  #foreach( $fieldName in $doc.fieldNames )    ## For each value    #foreach( $value in $doc.getFieldValues($fieldName) )      
               ## Field Name        
        ## Field Value(s)        
    #end     ## end for each value  #end       ## end for each field
                            #if( $foreach.count == 1 )            $fieldName:          #end                         $esc.html($value)

#if( $foreach.count == 1 ) $fieldName: #end	$esc.html($value)

就是将属性名与属性值呈现出来。

5、再看看hit.vm

#** *  Called for each matching document but then *  calls one of product_doc, join_doc or richtext_doc *  depending on which fields the doc has *##set($docId = $doc.getFieldValue('id'))
    
       ## Has a "name" field ?  #if($doc.getFieldValue('name'))    #parse("product_doc.vm")  ## Has a "compName_s" field ?  #elseif($doc.getFieldValue('compName_s'))    #parse("join_doc.vm")  ## Fallback to richtext_doc  #else    #parse("richtext_doc.vm")  #end

6、可以直接看richtest_doc.vm

#** *  Render a complex document in the results list *### Load Mime-Type List and Mapping#parse('mime_type_lists.vm')## Sets:## * supportedMimeTypes, AKA supportedtypes## * mimeExtensionsMap, AKA extMap## Title#if($doc.getFieldValue('title'))  #set($title = $esc.html($doc.getFirstValue('title')))#else  #set($title = "["+$doc.getFieldValue('id')+"]")#end## URL#if($doc.getFieldValue('url'))  #set($url = $doc.getFieldValue('url'))#elseif($doc.getFieldValue('resourcename'))  #set($url = "file:///$doc.getFieldValue('resourcename')")#else  #set($url = "$doc.getFieldValue('id')")#end## Sort out Mime-Type#set($ct = $list.get($doc.getFirstValue('content_type').split(";"),0))#set($filename = $doc.getFieldValue('resourcename'))#set($filetype = false)#set($filetype = $mimeExtensionsMap.get($ct))## TODO: falling back to file extension is convenient,## except when you don't have an icon for that extension## example "application/vnd.openxmlformats-officedocument.wordprocessingml.document"## document with a .docx extension.## It'd be nice to fall back to an "unknown" or the existing "file" type## We sort of do this below, but only if the filename has no extension## (anything after the last dot).#if(!$filetype)  #set($filetype = $filename.substring($filename.lastIndexOf(".")).substring(1))#end## #if(!$filetype)##   #set($filetype = "file")## #end## #if(!$supportedMimeTypes.contains($filetype))##   #set($filetype = "file")## #end## Row 1: Icon and Title and mlt link
    
       ## Icon  ## Small file type icons from http://www.splitbrain.org/projects/file_icons (public domain)  
       ## Title, hyperlinked  
         $title  ## Link for MLT / More Like This / Find Similar  
         #if($params.getBool('mlt', false) == false)              More Like This    #end  
    
## Row 2?: ID / URL
    
       #Id: #field('id')  ##自己修改  ##Time: #field('tstamp')
    
## Resource Name
    
       #if($doc.getFieldValue('resourcename'))    Resource name: $filename   #elseif($url)    URL: $url  #end  #if($ct)    ($ct)  #end
    
## Author#if($doc.getFieldValue('author'))  
    
         Author: #field('author')  
    
#end## Last_Modified Date#if($doc.getFieldValue('last_modified'))  
    
         last-modified:    #field('last_modified')  
    
#end## Main content of doc
    
       #field('content')
    
## Display Similar Documents / MLT = More Like This
    
       #set($mlt = $mltResults.get($docId))  #set($mltOn = $params.getBool('mlt'))  #if($mltOn == true)    
     
            Similar Items    
     
  #end  ## If has MLT enabled An Entries to show  #if ($mltOn && $mlt && $mlt.size() > 0)    
     
            #foreach($mltHit in $mlt)        #set($mltId = $mltHit.getFieldValue('id'))        
                
                   
                      $mltId          
       
          
                   
                      Title:                        $mltHit.getFieldValue('title')          
       
          
                   
                      Author:                        $mltHit.getFieldValue('author')            
                      Description:                        $mltHit.getFieldValue('description')          
       
        
      #end    ## end for each mltHit in $mlt    
     
  ## Else MLT Enabled but no mlt results for this query  #elseif($mltOn && $mlt.size() == 0)    
     
      No Similar Items Found
     
  #end
    
  ## div class=mlt#parse('debug.vm')

由于本文件是example自带的呈现文件，其属性也按照自带的schemal.xml定义，并不适用于nutch的schema。

因此，若要改变呈现的内容，可以直接修改此文件。如将

## Row 2?: ID / URL
    
       #Id: #field('id')

改为：

## Row 2?: ID / URL
    
       ##自己修改  #Time: #field('tstamp')

则在页面不再显示id，而是显示时间。

转载于:https://www.cnblogs.com/eaglegeek/p/4557879.html

你可能感兴趣的文章

Mongodb 基本命令

查看>>

Qt中QTableView中加入Check列实现

[leetcode] 1. Two Sum

Windbg调试Sql Server 进程

解决 No Entity Framework provider found for the ADO.NET provider

查看>>

SVN服务器搭建和使用（三）（转载）

查看>>

Android 自定义View (三) 圆环交替等待效果

查看>>

设置虚拟机虚拟机中fedora上网配置-bridge连接方式（图解）

查看>>