Major config changes for v2.0

.gitignore

@@ -53,3 +53,5 @@ js/ssSearch-debug.js
 js/ssSearch.js
 js/*.js.map
 
+# VSCode Configuration
+.idea

build.xml

@@ -104,7 +104,7 @@
   </pathconvert>
 
   <!--Load the configuration file as a property file-->
-  <xmlproperty file="${ssConfigFile}" keeproot="true"/>
+  <xmlproperty file="${ssConfigFile}" keeproot="true" collapseattributes="true"/>
 
   <!--Set the verbose flag to false by default; set to true to see detailed messages-->
   <property name="ssVerbose" value="false"/>
@@ -114,7 +114,7 @@
   <property name="ssVerboseReport" value="false"/>
 
   <!--The path to the search page, relative to the configuration directory-->
-  <property name="ssSearchFilePath" value="${configDir.converted}/${config.params.searchFile}"/>
+  <property name="ssSearchFilePath" value="${configDir.converted}/${config.params.searchPage.file}"/>
 
   <!--The name of the collection dir, derived from the search file-->
   <dirname property="ssCollectionDirName" file="${ssSearchFilePath}"/>
@@ -130,9 +130,12 @@
 
   <!--Output folder, which is forked, depending on whether or not an output folder
         is specified in the configuration file-->
-  <property name="ssOutputFolder" if:set="config.params.outputFolder"
-    value="${config.params.outputFolder}"/>
-  <property name="ssOutputFolder" unless:set="config.params.outputFolder" value="staticSearch"/>
+  <property name="ssOutputFolder"
+    if:set="config.params.output.dir"
+    value="${config.params.output.dir}"/>
+  <property name="ssOutputFolder"
+    unless:set="config.params.output.dir"
+    value="staticSearch"/>
 
   <!--Stemmer folder, which is forked, depending on whether or not a stemmer folder
         is specified in the configuration file-->
@@ -189,6 +192,7 @@
 
     <!-- If we're building the test data, we need to generate the VERSION file
          and also copy the .htm_ files to .html before starting. -->
+    <echo message="${ssConfigFileName}"/>
     <if>
       <equals arg1="configTest.xml" arg2="${ssConfigFileName}"/>
       <then>
@@ -235,11 +239,54 @@
       </then>
     </if>
   </target>
-
-  <target name="validate">
+  
+  <target name="checkConfigVersion">
     <description>
-      TARGET: validate
-      Task to validate that the source XHTML is well-formed XHTML. Note that this DOES
+      TARGET checkConfigVersion
+      Checks the supplied configuration file to see if it has 
+      @version="2" on its root element. If not, it invokes the
+      conversion process.
+    </description>
+    <echo message="Checking version of config file ${ssBaseDir.converted}/${ssConfig}..."/>
+    <fileset id="configOnly" dir="${ssBaseDir.converted}" includes="${ssConfig}">
+      <containsregexp expression="&lt;config[^&gt;]+version\s*=\s*.2."/>
+    </fileset>
+    <condition property="isV2" value="true" else="false">
+      <matches pattern="^2" string="${config.version}"/>
+    </condition>
+
+    <antcall target="convertConfigFile" unless:true="${isV2}"/>
+  </target>
+
+  <target name="convertConfigFile">
+    <description>
+      TARGET convertConfigFile
+      This runs an identity transform on the supplied config
+      file to create from it a version 2 file, and then reports
+      the outcome to the user.
+    </description>
+    <input addproperty="outputStrategy"
+           message="Your config file is for version 1 of staticSearch.&#x0a;
+It will be converted to version 2 format.&#x0a;
+Would you like to overwrite the existing file or &#x0a;
+create a new file? (o=overwrite, n=new)"
+           validargs="o,n"
+           defaultvalue="n"/>
+    <!--<echo message="${outputStrategy}"/>-->
+    <java classpath="${ssSaxon}" classname="net.sf.saxon.Transform" failonerror="true" fork="true">
+      <arg value="-xsl:${ssBaseDir.converted}/xsl/convert_v1_to_v2.xsl"/>
+      <arg value="-s:${ssBaseDir.converted}/${ssConfig}"/>
+      <arg value="--suppressXsltNamespaceCheck:on"/>
+      <arg value="ssBaseDir=${ssBaseDir.converted}"/>
+      <arg value="ssVerbose=${ssVerbose}"/>
+      <arg value="output=${outputStrategy}"/>
+    </java>
+  </target>
+
+  <target name="validate.html">
+    <description>
+      TARGET: validate.html
+      Target to check input files as well-formed XHTML. Note that this DOES
       NOT check whether or not it is valid XHTML: only that it is well-formed. This target also
       checks the config file and sees if it's invalid. We don't fail if it's invalid, though, since
       we must assume that some people just don't care: but raise the error anyway. 
@@ -255,6 +302,9 @@
         <exclude name="**/staticSearch_report.html"/>
       </fileset>
     </xmlvalidate>
+  </target>
+
+  <target name="validate.config" depends="checkConfigVersion">
     <echo message="Validating ${ssConfigFile} against the staticSearch schema..."/>
     <exec executable="java" failonerror="false">
       <arg line="-jar ${ssBaseDir.converted}/lib/jing.jar"/>
@@ -263,6 +313,13 @@
     </exec>
   </target>
 
+  <target name="validate" depends="validate.config, validate.html">
+    <description>
+      TARGET: validate
+      Task to validate various input files, including the XHTML and the configuration file.
+    </description>
+  </target>
+
 
 
   <target name="config">

buildSchema.xml

@@ -111,8 +111,9 @@
      ****************************************************************-->
 
     <target name="patchOddSrc">
-        <description>Patches the staticSearch ODD to use RNG interleave
-        to work around bugs in TEI's regular processing.</description>
+        <description>Preprocesses the staticSearch ODD to expand any constructs
+           or modify features that are not handled by the TEI stylesheets
+           (e.g. interleave).</description>
         <echo message="${echo.separator}"/>
         <echo message="Patching the staticSearch ODD ..."/>
         <echo message="${echo.separator}"/>
@@ -240,6 +241,14 @@
             and documentation creation tasks.</description>
     </target>
 
+    <target name="justDocs" depends="patchOddSrc, odd2processedOdd, odd2Doc">
+        <description>
+            TARGET: justDocs
+
+            This target just creates the documentation; useful for quick proofing and rapid development.
+        </description>
+    </target>
+
 
 
 </project>

configTest.xml

@@ -1,36 +1,30 @@
 <?xml version="1.0" encoding="UTF-8"?>
+<!--Configuration file file:/Users/takeda/projects/staticSearch/configTest.xml converted to version 2 2025-02-08.-->
 <config xmlns="http://hcmc.uvic.ca/ns/staticSearch" version="2">
     <params>
-        <searchFile>test/search.html</searchFile>
-        <versionFile>test/VERSION</versionFile>
-        <recurse>true<!--Need to set recurse to true for the test--></recurse>
-        <phrasalSearch>true</phrasalSearch>
-        <wildcardSearch>true</wildcardSearch>
-        <createContexts>true</createContexts>
-        <resultsPerPage>5</resultsPerPage>
-        <minWordLength>2</minWordLength>
-        <!--NOTE: If phrasalSearch is set to TRUE, then
-        maxContexts prop will be ignored-->
-        <maxKwicsToHarvest>5</maxKwicsToHarvest>
-        <maxKwicsToShow>5</maxKwicsToShow>
-        <totalKwicLength>15</totalKwicLength>
-        <kwicTruncateString>...</kwicTruncateString>
-        <stopwordsFile>test/test_stopwords.txt</stopwordsFile>
-        <dictionaryFile>xsl/english_words.txt</dictionaryFile>
-        <outputFolder>ssTest</outputFolder>
+        <searchPage file="test/search.html"/>
+        <index recurse="true"/>
+        <stopwords file="test/test_stopwords.txt"/>
+        <dictionary file="xsl/english_words.txt"/>
+        <scoringAlgorithm name="raw"/>
+        <stemmer dir="stemmers/en/"/>
+        <tokenizer minWordLength="2"/>
+        <createContexts create="true"
+            phrasalSearch="true"
+            wildcardSearch="true"
+            maxKwicLength="15"
+            kwicTruncateString="..."/>
+        <results resultsPerPage="5" maxKwicsToShow="5" maxResults="1000"/>
+        <version file="test/VERSION"/>
+        <output dir="ssTest"/>
     </params>
-
     <rules>
-        <rule weight="2"
-            match="h1 | h2"/>
-        <rule weight="0"
-            match="span[@class='lineNum']"/>
-        <rule weight="0"
-            match="script | style"/>
+        <rule weight="2" match="h1 | h2"/>
+        <rule weight="0" match="span[@class='lineNum']"/>
+        <rule weight="0" match="script | style"/>
         <rule weight="10" match="span[@class='weighty']"/>
         <rule weight="5" match="td/span"/>
     </rules>
-
     <contexts>
         <context match="blockquote" label="Quotations"/>
         <context match="div[@class='l']"/>
@@ -39,12 +33,10 @@
         <context match="cite" label="Citations"/>
         <context match="p[contains-token(@class,'citation')]" label="Citations"/>
     </contexts>
-
     <excludes>
         <exclude type="index" match="html[@id='excluded']"/>
         <exclude match="meta[contains-token(@class,'excludedMeta')]" type="filter"/>
     </excludes>
-
     <filters>
         <filter filterName="Worth reading">
             <span xmlns="http://www.w3.org/1999/xhtml">
@@ -56,6 +48,5 @@
                 <span style="font-variant: small-caps;">Random Piffle</span>
             </span>
         </filter>
-
     </filters>
 </config>

configTest_v1.xml

@@ -0,0 +1,61 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<config xmlns="http://hcmc.uvic.ca/ns/staticSearch" version="1">
+   <params>
+      <searchFile>test/search.html</searchFile>
+      <versionFile>test/VERSION</versionFile>
+      <recurse>true<!--Need to set recurse to true for the test--></recurse>
+      <phrasalSearch>true</phrasalSearch>
+      <wildcardSearch>true</wildcardSearch>
+      <createContexts>true</createContexts>
+      <resultsPerPage>5</resultsPerPage>
+      <minWordLength>2</minWordLength>
+      <!--NOTE: If phrasalSearch is set to TRUE, then
+        maxContexts prop will be ignored-->
+      <maxKwicsToHarvest>5</maxKwicsToHarvest>
+      <maxKwicsToShow>5</maxKwicsToShow>
+      <totalKwicLength>15</totalKwicLength>
+      <kwicTruncateString>...</kwicTruncateString>
+      <stopwordsFile>test/test_stopwords.txt</stopwordsFile>
+      <dictionaryFile>xsl/english_words.txt</dictionaryFile>
+      <outputFolder>ssTest</outputFolder>
+   </params>
+
+   <rules>
+      <rule weight="2"
+         match="h1 | h2"/>
+      <rule weight="0"
+         match="span[@class='lineNum']"/>
+      <rule weight="0"
+         match="script | style"/>
+      <rule weight="10" match="span[@class='weighty']"/>
+      <rule weight="5" match="td/span"/>
+   </rules>
+
+   <contexts>
+      <context match="blockquote" label="Quotations"/>
+      <context match="div[@class='l']"/>
+      <context match="span[@class='note'] | *[contains-token(@class,'sidenotes')]"
+         label="Notes"/>
+      <context match="cite" label="Citations"/>
+      <context match="p[contains-token(@class,'citation')]" label="Citations"/>
+   </contexts>
+
+   <excludes>
+      <exclude type="index" match="html[@id='excluded']"/>
+      <exclude match="meta[contains-token(@class,'excludedMeta')]" type="filter"/>
+   </excludes>
+
+   <filters>
+      <filter filterName="Worth reading">
+         <span xmlns="http://www.w3.org/1999/xhtml">
+            <i>Really</i> worth reading: 
+         </span>
+      </filter>
+      <filter filterName="Random nonsense">
+         <span xmlns="http://www.w3.org/1999/xhtml" lang="en">
+            <span style="font-variant: small-caps;">Random Piffle</span>
+         </span>
+      </filter>
+
+   </filters>
+</config>

config_documentation.xml

@@ -1,39 +1,33 @@
 <?xml version="1.0" encoding="UTF-8"?>
+<!--Configuration file file:/Users/takeda/projects/staticSearch/config_documentation.xml converted to version 2 2025-02-19.-->
 <config xmlns="http://hcmc.uvic.ca/ns/staticSearch" version="2">
-    <params>
-        <searchFile>docs/index.html</searchFile>
-        <versionFile>test/VERSION</versionFile>
-        <stemmerFolder>en</stemmerFolder>
-        <recurse>false</recurse>
-        <phrasalSearch>true</phrasalSearch>
-        <wildcardSearch>true</wildcardSearch>
-        <createContexts>true</createContexts>
-        <!--NOTE: If phrasalSearch is set to TRUE, then
-        maxContexts prop will be ignored-->
-        <maxKwicsToHarvest>3</maxKwicsToHarvest>
-        <maxKwicsToShow>10</maxKwicsToShow>
-        <totalKwicLength>15</totalKwicLength>
-        <kwicTruncateString>...</kwicTruncateString>
-        <stopwordsFile>xsl/english_stopwords.txt</stopwordsFile>
-        <dictionaryFile>xsl/english_words.txt</dictionaryFile>
-        <outputFolder>search</outputFolder>
-    </params>
-
-    <rules>
-        <rule weight="2"
-            match="h1 | h2"/>
-        <!--ignore the TOC-->
-        <rule weight="0"
-            match="div[contains-token(@class,'tei_front')]"/>
-
-        <!--Don't index codeblocks in table cells-->
-        <rule weight="0" match="td/pre"/>
-
-    </rules>
-
-    <excludes>
+   <params>
+      <searchPage file="docs/index.html"/>
+      <index recurse="false"/>
+      <stopwords file="xsl/english_stopwords.txt"/>
+      <dictionary file="xsl/english_words.txt"/>
+      <scoringAlgorithm name="raw"/>
+      <stemmer dir="stemmers/en"/>
+      <tokenizer minWordLength="2"/>
+      <createContexts create="true"
+                      phrasalSearch="true"
+                      wildcardSearch="true"
+                      maxKwicLength="15"
+                      kwicTruncateString="..."/>
+      <results resultsPerPage="0" maxKwicsToShow="10" maxResults="2000"/>
+      <version file="test/VERSION"/>
+      <output dir="search"/>
+   </params>
+   <rules>
+      <rule weight="2" match="h1 | h2"/>
+      <!--ignore the TOC-->
+      <rule weight="0" match="div[contains-token(@class,'tei_front')]"/>
+      <!--Don't index codeblocks in table cells-->
+      <rule weight="0" match="td/pre"/>
+   </rules>
+   <excludes>
         <!--Ignore the master SS doc from the search-->
-        <exclude type="index" match="html[matches(document-uri(/),'staticSearch.html$')]"/>
-    </excludes>
-
+      <exclude type="index"
+               match="html[matches(document-uri(/),'staticSearch.html$')]"/>
+   </excludes>
 </config>