Merge pull request ezsystems#155 from ezsystems/5.3-beta1fixes

5.3 beta1fixes

doc/INSTALL.md

@@ -9,9 +9,10 @@ Requirements:
 
 Installing:
 -----------
-1. Extract the ezfind extension and place it in the extensions folder.
+1. Extract the ezfind extension and place it in the ezpublish_legacy/extensions folder.
 2. Run it:
-   1. For small sites: use a regular java 1.7 JRE from Sun
+
+   1. For small sites: use a regular java 1.7 JRE from Oracle
       Start the Java based search engine (based on Solr).
       Enter `extension/ezfind/java` and start the Solr engine with the command:
 
@@ -20,16 +21,15 @@ Installing:
       ```
       (See configuration for more details)
       Make sure that the user running the java application has write access to
-      `extension/ezfind/java/solr/data` and `extension/ezfind/java/logs` .
-   2. For high performance, larger sites, use a 64 bit OS and 64 bit Java VM from Sun (1.5 recommended)
+      `extension/ezfind/java/solr/` and `extension/ezfind/java/logs` .
+   2. For high performance, larger sites, use a 64 bit OS and 64 bit Java VM from Oracle (1.7 recommended)
       you need to use some extra parameters for java too, the following is recommended for typical sites:
 
       ```bash
       $ java -server -d64 -Xmx1500m -Xms1500m -XX:+UseParallelGC -XX:+AggressiveOpts -XX:NewRatio=5 -jar start.jar
       ```
       This will: make sure the Java VM is started with 64 bit mode,
-      allocate a heap space of 1500MB which is used for Solr caches and objects,
-      the rest of the options are related to garbage collection
+
 3. Enable the extension in eZ Publish. Do this by opening settings/override/site.ini.append.php ,
    and add in the `[ExtensionSettings]` block:
 

doc/multicore-HOWTO.txt

@@ -2,62 +2,72 @@
 eZ find & Solr multicore how-to
 ===============================
 
-eZ find 2.2 introduces support for solr multicore feature. This finally allows
-for better language specific search & analysis.
+Since eZ Find 5.3 (which upgrades Solr to 4.7), running Solr in multicore mode
+is the default
 
 Starting solr multicore
 =======================
-A default configuration of solr multicore is shipped with ezfind.
 
-To start it, use the following command line instead of the default one::
-    $java/: java -Dsolr.solr.home=solr.multicore -jar start.jar
 
-Configuring solr for multicore
-==============================
-You will find in extension/ezfind/java a folder named solr.multicore. This is
-a distinct solr home that already provides multicore features.
+To start it, use the regular command line
+    $java/: java -jar start.jar
+
+Since eZ Find 5.3, the directory solr.multicore does not exist anymore!
+
+Configuring solr additional cores
+=================================
+
+You will find in extension/ezfind/java a folder named solr.languages. This
+directory contains language specific schema configurations.
+
 
 solr.xml
 --------
-Located in the solr.multicore folder mentionned above, this file is required to
-enable multicore.
+
 
 The <cores> element is used to configure the administration URL for the cores,
 as well as the list of available cores. Each 'core' sub-element configures a
 new core. The 'name' attribute determines what URL element is used to access
 this core, while the instanceDir attribute determines the subfolder of
-solr.multicore where this core configuration can be found.
+solr where this core configuration can be found.
 
 Creating a new core
 -------------------
 Note that even though only language specific cores are used by eZ find at the
 moment, cores can be used for much more than that.
 
-1) you need to copy one of the existing ones, for example
-extension/ezfind/java/solr.multicore/eng-GB to a new name, the easiest being
-the target locale name::
+1) you need to create a new directory under java/solr and then copy the default
+configuration, extension/ezfind/java/solr/ezp-default/conf to this new
+directory, typically the target locale name::
+
+    $ cd extension/ezfind/java/solr
+    $ mkdir <new-core>
+    $ cp -rp ezp-default/conf <new-core>
+
+Finally you should copy the appropriate files from
+extension/ezfind/java/solr.languages
+
+For example, for a  french language core:
+
+    $ cp ../solr.languages/fr/* <new-core>/conf
+
+2) Make sure Solr knows about the new core
+
+This can be done in two ways: edit the solr.xml file manually
 
-    $ezroot: cd extension/ezfind/java/solr.multicore
-    $solr.multicore: cp -rp eng-GB nor-NO
+2a) edit solr/solr.xml and add the newly created core here::
 
-2) edit solr.multicore/solr.xml and add the newly created core here::
+    <core name="<new-core>" instanceDir="<new-core>" />
 
-    <core name="eng-GB" instanceDir="eng-GB" />
-    <core name="nor-NO" instanceDir="nor-NO" />
+  Then restart Solr
 
-3) now, you must configure the core so that it actually knows what language
-it will handle, and will perform language specific analysis. To do so, edit
-the conf/schema.xml file located in this particular core::
+2b) go the the Solr admin interface
 
-    extension/ezfind/java/solr.multicore/nor-NO/conf/schema.xml
+Select the "Core Admin"
+Click "Add new core" and fill in the directory and name of the new core
 
-In this file, you will find multiple occurences of 'English' (if you've used
-eng-GB as the source). All of these will be attributes of a 'filter' XML element.
-Edit these, and replace 'English' with your language. The list of supported
-languages can be found on solr's wiki:
-http://wiki.apache.org/solr/AnalyzersTokenizersTokenFilters#solr.SnowballPorterFilterFactory
+You do not need to restart Solr, the core is added and ready to use
 
-For nor-NO, you would use 'Norwegian'.
 
 Tips
 ====

java/solr.languages/en/language-specific-fieldtypes.xml

@@ -24,7 +24,7 @@
         <filter class="solr.SynonymFilterFactory" synonyms="synonyms.txt" ignoreCase="true" expand="true"/>
         <filter class="solr.StopFilterFactory"
                 ignoreCase="true"
-                words="lang/stopwords_en.txt"
+                words="stopwords_en.txt"
         />
         <filter class="solr.LowerCaseFilterFactory"/>
         <filter class="solr.EnglishPossessiveFilterFactory"/>

java/solr.languages/fr/contractions_fr.txt

@@ -0,0 +1,15 @@
+# Set of French contractions for ElisionFilter
+# TODO: load this as a resource from the analyzer and sync it in build.xml
+l
+m
+t
+qu
+n
+s
+j
+d
+c
+jusqu
+quoiqu
+lorsqu
+puisqu

java/solr.languages/fr/language-specific-fieldtypes.xml

@@ -4,7 +4,7 @@
     <analyzer>
         <tokenizer class="solr.StandardTokenizerFactory"/>
         <!-- removes l', etc -->
-        <filter class="solr.ElisionFilterFactory" ignoreCase="true" articles="lang/contractions_fr.txt"/>
+        <filter class="solr.ElisionFilterFactory" ignoreCase="true" articles="contractions_fr.txt"/>
         <filter class="solr.LowerCaseFilterFactory"/>
         <filter class="solr.StopFilterFactory" ignoreCase="true" words="stopwords_fr.txt" format="snowball" />
         <filter class="solr.FrenchLightStemFilterFactory"/>

java/solr/ezp-default/conf/solrconfig.xml

@@ -863,7 +863,7 @@
       -->
   <!--  </requestHandler> -->
 
-    <requestHandler name="ezpublish" class="solr.SearchHandler" >
+    <requestHandler name="ezpublish" class="solr.SearchHandler" default="true" >
     <lst name="defaults">
       <str name="echoParams">explicit</str>
       <str name="defType">edismax</str>