Merge branch 'develop' into release

kingsdigitallab · Sep 2, 2024 · ff93b8b · ff93b8b
2 parents c5c8171 + 21b488b
commit ff93b8b
Show file tree

Hide file tree

Showing 11 changed files with 925 additions and 0 deletions.
diff --git a/.gitignore b/.gitignore
@@ -29,3 +29,5 @@ tei_all.rng
 /apps/checktexts/edition
 .idea/
 /apps/entities/xslt/tei-to-json.sef.json
+/apps/textsearch/clone
+/apps/textsearch/to-be-indexed
diff --git a/_build/autopull.sh b/_build/autopull.sh
@@ -49,6 +49,8 @@ if [[ $h1 != $h2 ]]; then
   # fi
 fi
 
+su - $GITUSER -c "export SITE_ENV=$SITE_ENV; cd $PROJECT_DIR/apps/textsearch/ && npm run index"
+
 # update tweets
 # su - $GITUSER -c "cd $PROJECT_DIR && npm run tweets -w frontend"
 

diff --git a/apps/textsearch/html-to-html-modern.xslt b/apps/textsearch/html-to-html-modern.xslt
@@ -0,0 +1,76 @@
+<xsl:stylesheet version="2.0" 
+  xmlns="http://www.w3.org/1999/xhtml" 
+  xmlns:html="http://www.w3.org/1999/xhtml" 
+  xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+  xmlns:tei="http://www.tei-c.org/ns/1.0" 
+>
+
+  <xsl:template match="@*|node()">
+    <xsl:call-template name="copy-element" />
+  </xsl:template>
+
+  <xsl:template name="copy-element">
+    <xsl:copy>
+      <xsl:apply-templates select="@*|node()"/>
+    </xsl:copy>
+  </xsl:template>
+
+  <xsl:template match="/">
+    <html>
+      <head>
+      </head>
+      <body data-pagefind-body="" 
+        data-pagefind-meta="url[data-url],title[data-title],version[data-version]" 
+        data-pagefind-filter="book[data-book],version[data-version]"
+        data-pagefind-sort="book-page[data-book-page]"
+        data-title="#title#" 
+        data-url="#url#" 
+        data-book="#book#" 
+        data-version="#version#"
+        data-book-page="#bookPage#"
+        >
+        <xsl:call-template name="copy-element" />
+      </body>
+    </html>
+  </xsl:template>
+
+  <!-- <span class="tei-note tei-type-term info-box" data-tei="note" data-tei-type="term"> -->
+  <xsl:template match="span[contains(@class, 'info-box') and not (contains(@class, 'has-info-box'))]">
+  </xsl:template>
+
+  <xsl:template match="span[@data-tei-resp='ednote']">
+  </xsl:template>
+
+  <xsl:template match="*[contains(@class, 'hidden')]">
+  </xsl:template>
+
+  <xsl:template match="span[@data-tei-type='event-mark']">
+  </xsl:template>
+
+  <!-- VERSION-SPECIFIC -->
+
+  <!-- <span class="tei-w" data-tei="w" data-tei-norm="sins"><span class="norm">sins</span><span class="orig">sinns</span></span> -->
+  <xsl:template match="span[@class='orig']">
+  </xsl:template>
+
+  <!-- <span class="tei-abbr" data-tei="abbr">hon.< -->
+  <xsl:template match="span[@data-tei='abbr']">
+  </xsl:template>
+
+  <xsl:template match="span[@data-tei='surplus']">
+  </xsl:template>
+
+  <xsl:template match="span[@data-tei='sic']">
+  </xsl:template>
+
+  <!-- <span class="tei-measure tei-type-currency not-a-word" data-tei="measure" data-tei-type="currency" data-tei-unit="£">6<span class="tei-am" data-tei="am" data-tei-rend="superscript italic">l</span></span> -->
+  <xsl:template match="span[@data-tei='am']">
+  </xsl:template>
+
+  <xsl:template match="del[@data-tei='del']">
+  </xsl:template>
+
+  <xsl:template match="span[@data-tei='fw' and contains(@data-tei-place, 'margin') and not(contains(@class, 'tei-type-head'))]">
+  </xsl:template>
+
+</xsl:stylesheet>
diff --git a/apps/textsearch/html-to-html-semidip.xslt b/apps/textsearch/html-to-html-semidip.xslt
@@ -0,0 +1,35 @@
+<xsl:stylesheet version="2.0" 
+  xmlns="http://www.w3.org/1999/xhtml" 
+  xmlns:html="http://www.w3.org/1999/xhtml" 
+  xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+  xmlns:tei="http://www.tei-c.org/ns/1.0" 
+>
+
+  <xsl:template match="@*|node()">
+    <xsl:call-template name="copy-element" />
+  </xsl:template>
+
+  <xsl:template name="copy-element">
+    <xsl:copy>
+      <xsl:apply-templates select="@*|node()"/>
+    </xsl:copy>
+  </xsl:template>
+
+  <xsl:template match="/">
+    <html>
+      <head>
+      </head>
+      <body data-pagefind-body="" 
+        data-pagefind-meta="url[data-url],title[data-title]" 
+        data-title="#title#" 
+        data-url="#url#" 
+        data-pagefind-filter="book[data-book],version[data-version]"
+        data-book="#book#" 
+        data-version="#version#" 
+        >
+        <xsl:call-template name="copy-element" />
+      </body>
+    </html>
+  </xsl:template>
+
+</xsl:stylesheet>
diff --git a/apps/textsearch/html-to-html.sef.jsont b/apps/textsearch/html-to-html.sef.jsont
@@ -0,0 +1 @@
+{"N":"package","version":"20","packageVersion":"1","saxonVersion":"SaxonJS 2.6","target":"JS","targetVersion":"2","name":"TOP-LEVEL","relocatable":"false","buildDateTime":"2024-08-23T01:24:27.625+01:00","ns":"xml=~ =http://www.w3.org/1999/xhtml html=http://www.w3.org/1999/xhtml xsl=~ tei=http://www.tei-c.org/ns/1.0","C":[{"N":"co","id":"0","uniform":"true","binds":"1","C":[{"N":"template","flags":"os","module":"html-to-html.xslt","slots":"200","baseUri":"file:///home/jeff/src/prj/alice-thornton/apps/textsearch/html-to-html.xslt","name":"Q{}copy-element","line":"12","expand-text":"false","sType":"1 ","C":[{"N":"copy","sType":"1 ","flags":"cin","role":"body","line":"13","C":[{"N":"applyT","sType":"* ","line":"14","mode":"#unnamed","bSlot":"0","C":[{"N":"docOrder","sType":"*N u[NT,NP,NC,NE]","role":"select","line":"14","C":[{"N":"union","op":"|","sType":"*N u[NT,NP,NC,NE]","ns":"= xml=~ fn=~ html=http://www.w3.org/1999/xhtml xsl=~ tei=http://www.tei-c.org/ns/1.0 ","C":[{"N":"axis","name":"attribute","nodeTest":"*NA"},{"N":"axis","name":"child","nodeTest":"*N u[NT,NP,NC,NE]"}]}]}]}]}]}]},{"N":"co","id":"1","binds":"0","C":[{"N":"mode","onNo":"TC","flags":"","patternSlots":"0","prec":"","C":[{"N":"templateRule","rank":"0","prec":"0","seq":"1","ns":"xml=~ =http://www.w3.org/1999/xhtml html=http://www.w3.org/1999/xhtml xsl=~ tei=http://www.tei-c.org/ns/1.0","minImp":"0","flags":"s","slots":"200","baseUri":"file:///home/jeff/src/prj/alice-thornton/apps/textsearch/html-to-html.xslt","line":"18","module":"html-to-html.xslt","expand-text":"false","match":"/","prio":"-0.5","matches":"ND","C":[{"N":"p.nodeTest","role":"match","test":"ND","sType":"1ND","ns":"= xml=~ fn=~ html=http://www.w3.org/1999/xhtml xsl=~ tei=http://www.tei-c.org/ns/1.0 "},{"N":"elem","name":"html","sType":"1NE nQ{http://www.w3.org/1999/xhtml}html ","nsuri":"http://www.w3.org/1999/xhtml","namespaces":"=http://www.w3.org/1999/xhtml html=http://www.w3.org/1999/xhtml tei=http://www.tei-c.org/ns/1.0","role":"action","line":"19","C":[{"N":"sequence","sType":"*NE ","C":[{"N":"elem","name":"head","sType":"1NE nQ{http://www.w3.org/1999/xhtml}head ","nsuri":"http://www.w3.org/1999/xhtml","namespaces":"=http://www.w3.org/1999/xhtml html=http://www.w3.org/1999/xhtml tei=http://www.tei-c.org/ns/1.0","line":"20","C":[{"N":"empty","sType":"0 "}]},{"N":"elem","name":"body","sType":"1NE nQ{http://www.w3.org/1999/xhtml}body ","nsuri":"http://www.w3.org/1999/xhtml","namespaces":"=http://www.w3.org/1999/xhtml html=http://www.w3.org/1999/xhtml tei=http://www.tei-c.org/ns/1.0","line":"29","C":[{"N":"sequence","sType":"* ","C":[{"N":"att","name":"data-pagefind-body","nsuri":"","sType":"1NA ","C":[{"N":"str","sType":"1AS ","val":""}]},{"N":"att","name":"data-pagefind-meta","nsuri":"","sType":"1NA ","C":[{"N":"str","sType":"1AS ","val":"url[data-url],title[data-title]"}]},{"N":"att","name":"data-title","nsuri":"","sType":"1NA ","C":[{"N":"str","sType":"1AS ","val":"#title#"}]},{"N":"att","name":"data-url","nsuri":"","sType":"1NA ","C":[{"N":"str","sType":"1AS ","val":"#url#"}]},{"N":"att","name":"data-pagefind-filter","nsuri":"","sType":"1NA ","C":[{"N":"str","sType":"1AS ","val":"book[data-book],version[data-version]"}]},{"N":"att","name":"data-book","nsuri":"","sType":"1NA ","C":[{"N":"str","sType":"1AS ","val":"#book#"}]},{"N":"att","name":"data-version","nsuri":"","sType":"1NA ","C":[{"N":"str","sType":"1AS ","val":"#version#"}]},{"N":"callT","bSlot":"0","sType":"* ","name":"Q{}copy-element","line":"30"}]}]}]}]}]},{"N":"templateRule","rank":"1","prec":"0","seq":"0","ns":"xml=~ =http://www.w3.org/1999/xhtml html=http://www.w3.org/1999/xhtml xsl=~ tei=http://www.tei-c.org/ns/1.0","minImp":"0","flags":"s","slots":"200","baseUri":"file:///home/jeff/src/prj/alice-thornton/apps/textsearch/html-to-html.xslt","line":"8","module":"html-to-html.xslt","expand-text":"false","match":"@*|node()","prio":"-0.5","matches":"NA","C":[{"N":"p.nodeTest","role":"match","test":"NA","sType":"1NA"},{"N":"callT","bSlot":"0","sType":"* ","name":"Q{}copy-element","role":"action","line":"9"}]},{"N":"templateRule","rank":"2","prec":"0","seq":"0","ns":"xml=~ =http://www.w3.org/1999/xhtml html=http://www.w3.org/1999/xhtml xsl=~ tei=http://www.tei-c.org/ns/1.0","minImp":"0","flags":"s","slots":"200","baseUri":"file:///home/jeff/src/prj/alice-thornton/apps/textsearch/html-to-html.xslt","line":"8","module":"html-to-html.xslt","expand-text":"false","match":"@*|node()","prio":"-0.5","matches":"N u[NT,NP,NC,NE]","C":[{"N":"p.nodeTest","role":"match","test":"N u[NT,NP,NC,NE]","sType":"1N u[NT,NP,NC,NE]"},{"N":"callT","bSlot":"0","sType":"* ","name":"Q{}copy-element","role":"action","line":"9"}]}]}]},{"N":"overridden"},{"N":"output","C":[{"N":"property","name":"Q{http://saxon.sf.net/}stylesheet-version","value":"20"}]},{"N":"decimalFormat"}],"Σ":"6f3b471f"}
diff --git a/apps/textsearch/package-lock.json b/apps/textsearch/package-lock.json
diff --git a/apps/textsearch/package.json b/apps/textsearch/package.json
@@ -0,0 +1,21 @@
+{
+  "name": "textsearch",
+  "version": "0.1.0",
+  "main": "textsearch.js",
+  "scripts": {
+    "test": "echo \"Error: no test specified\" && exit 1",
+    "extract": "rm -rf clone && git clone https://github.com/kingsdigitallab/alice-thornton.git -b dts --single-branch clone",
+    "transform": "node textsearch.js",
+    "load": "npx pagefind --site to-be-indexed --output-path ../../frontend/_site/pagefind",
+    "index": "npm run extract && npm run transform && npm run load",
+    "tl": "npm run transform && npm run load"
+  },
+  "author": "",
+  "license": "ISC",
+  "description": "",
+  "dependencies": {
+    "pagefind": "^1.1.0",
+    "saxon-js": "^2.6.0",
+    "xmldom": "^0.6.0"
+  }
+}