Add TopMargin & BottomMargin transformations

filak · Jan 24, 2020 · 61bb10e · 61bb10e
1 parent 6c43b7c
commit 61bb10e
Show file tree

Hide file tree

Showing 5 changed files with 168 additions and 18 deletions.
diff --git a/alto__hocr.xsl b/alto__hocr.xsl
@@ -71,7 +71,17 @@ License: MIT
  <xsl:template match="*:Page">
     <xsl:variable name="fname"><xsl:value-of select="//*:alto/*:Description/*:sourceImageInformation/*:fileName"/></xsl:variable>
     <div class="ocr_page" id="{mf:getId(@ID,'page',.)}" title="image {$fname}; bbox 0 0 {@WIDTH} {@HEIGHT}; ppageno 0">
+        <xsl:apply-templates select="*:TopMargin"/>
         <xsl:apply-templates select="*:PrintSpace"/>
+        <xsl:apply-templates select="*:BottomMargin"/>
+     </div>
+  </xsl:template>
+
+
+ <xsl:template match="*:TopMargin">
+    <div class="ocr_header" id="{mf:getId(@ID,'block',.)}" title="{mf:getBox(@HEIGHT,@WIDTH,@VPOS,@HPOS,@WC)}">
+        <xsl:apply-templates select="*:ComposedBlock"/>
+         <xsl:apply-templates select="*:TextBlock"/>
      </div>
   </xsl:template>
 
@@ -80,9 +90,17 @@ License: MIT
         <xsl:apply-templates select="*:ComposedBlock"/>
         <xsl:apply-templates select="*:TextBlock"/>
   </xsl:template>
-
-
- <xsl:template match="*:ComposedBlock">
+
+
+ <xsl:template match="*:BottomMargin">
+    <div class="ocr_footer" id="{mf:getId(@ID,'block',.)}" title="{mf:getBox(@HEIGHT,@WIDTH,@VPOS,@HPOS,@WC)}">
+        <xsl:apply-templates select="*:ComposedBlock"/>
+         <xsl:apply-templates select="*:TextBlock"/>
+     </div>
+  </xsl:template>
+
+
+  <xsl:template match="*:ComposedBlock">
     <div class="ocr_carea" id="{mf:getId(@ID,'block',.)}" title="{mf:getBox(@HEIGHT,@WIDTH,@VPOS,@HPOS,@WC)}">
          <xsl:apply-templates select="*:TextBlock"/>
      </div>

diff --git a/hocr__alto2.0.xsl b/hocr__alto2.0.xsl
@@ -30,6 +30,7 @@ License: MIT
         </alto>
   </xsl:template>
 
+
   <xsl:template match="*:head">
         <Description>
             <MeasurementUnit>pixel</MeasurementUnit>
@@ -53,16 +54,43 @@ License: MIT
             <xsl:apply-templates select="*:div[@class='ocr_page']"/>
         </Layout>
   </xsl:template>
+
+
   <xsl:template match="*:div[@class='ocr_page']">
         <!--  bbox 552 999 1724 1141 x1-L2-T3-R4-B5 -->
         <xsl:variable name="box" select="tokenize(mf:getBoxPage(@title), ' ')"/>
-            <Page ID="{@id}" PHYSICAL_IMG_NR="1" HEIGHT="{$box[5]}" WIDTH="{$box[4]}">
-                <PrintSpace HEIGHT="{$box[5]}" WIDTH="{$box[4]}" VPOS="0" HPOS="0">
+        <Page ID="{@id}" PHYSICAL_IMG_NR="1" HEIGHT="{$box[5]}" WIDTH="{$box[4]}">
+
+            <xsl:apply-templates select="*:div[@class='ocr_header']"/>
+
+            <PrintSpace HEIGHT="{$box[5]}" WIDTH="{$box[4]}" VPOS="0" HPOS="0">
+                <xsl:apply-templates select="*:div[@class='ocr_carea']"/>
+                <xsl:apply-templates select="*:p[@class='ocr_par']"/>
+            </PrintSpace>
+
+            <xsl:apply-templates select="*:div[@class='ocr_footer']"/>
+
+        </Page>
+  </xsl:template>
 
-                    <xsl:apply-templates/>
 
-                </PrintSpace>
-            </Page>
+  <xsl:template match="*:div[@class='ocr_header']">
+      <xsl:variable name="box" select="tokenize(mf:getBox(@title), ' ')"/>
+      <TopMargin ID="{@id}" HEIGHT="{number($box[5]) - number($box[3])}" WIDTH="{number($box[4]) - number($box[2])}" VPOS="{$box[3]}" HPOS="{$box[2]}">
+
+          <xsl:apply-templates/>
+
+      </TopMargin>
+  </xsl:template>
+
+
+  <xsl:template match="*:div[@class='ocr_footer']">
+      <xsl:variable name="box" select="tokenize(mf:getBox(@title), ' ')"/>
+      <BottomMargin ID="{@id}" HEIGHT="{number($box[5]) - number($box[3])}" WIDTH="{number($box[4]) - number($box[2])}" VPOS="{$box[3]}" HPOS="{$box[2]}">
+
+          <xsl:apply-templates/>
+
+      </BottomMargin>
   </xsl:template>
 
 
@@ -153,24 +181,28 @@ License: MIT
         </xsl:choose>
   </xsl:template>
 
-
+
+
 <xsl:function name="mf:getFname">
     <xsl:param name="titleString"/>
     <xsl:variable name="pPat">"</xsl:variable>
     <xsl:variable name="fpath" select="substring-after(tokenize(normalize-space($titleString),'; ')[1],'image &quot;')"/>
     <xsl:value-of select="reverse(tokenize(replace($fpath,$pPat,''),'\\'))[1]"/>
 </xsl:function>
-
+
+
 <xsl:function name="mf:getBoxPage">
     <xsl:param name="titleString"/>
     <xsl:value-of select="tokenize(normalize-space($titleString),'; ')[2]"/>
 </xsl:function>
 
+
 <xsl:function name="mf:getBox">
     <xsl:param name="titleString"/>
     <xsl:value-of select="tokenize(normalize-space($titleString),'; ')"/>
 </xsl:function>
 
+
 <xsl:function name="mf:getConfidence">
     <xsl:param name="titleString"/>
     <xsl:variable name="wconfString" select="tokenize(normalize-space($titleString),'; ')[2]" />
@@ -186,5 +218,6 @@ License: MIT
     </xsl:choose>
 
 </xsl:function>
-
+
+
   </xsl:stylesheet>
diff --git a/hocr__alto2.1.xsl b/hocr__alto2.1.xsl
@@ -29,6 +29,7 @@ License: MIT
             <xsl:apply-templates/>
         </alto>
   </xsl:template>
+
 
   <xsl:template match="*:head">
         <Description>
@@ -53,17 +54,44 @@ License: MIT
             <xsl:apply-templates select="*:div[@class='ocr_page']"/>
         </Layout>
   </xsl:template>
+
+
   <xsl:template match="*:div[@class='ocr_page']">
         <!--  bbox 552 999 1724 1141 x1-L2-T3-R4-B5 -->
         <xsl:variable name="box" select="tokenize(mf:getBoxPage(@title), ' ')"/>
         <Page ID="{@id}" PHYSICAL_IMG_NR="1" HEIGHT="{$box[5]}" WIDTH="{$box[4]}">
-            <PrintSpace HEIGHT="{$box[5]}" WIDTH="{$box[4]}" VPOS="0" HPOS="0">
 
-                <xsl:apply-templates/>
+            <xsl:apply-templates select="*:div[@class='ocr_header']"/>
 
+            <PrintSpace HEIGHT="{$box[5]}" WIDTH="{$box[4]}" VPOS="0" HPOS="0">
+                <xsl:apply-templates select="*:div[@class='ocr_carea']"/>
+                <xsl:apply-templates select="*:p[@class='ocr_par']"/>
             </PrintSpace>
+
+            <xsl:apply-templates select="*:div[@class='ocr_footer']"/>
+
         </Page>
   </xsl:template>
+
+
+  <xsl:template match="*:div[@class='ocr_header']">
+      <xsl:variable name="box" select="tokenize(mf:getBox(@title), ' ')"/>
+      <TopMargin ID="{@id}" HEIGHT="{number($box[5]) - number($box[3])}" WIDTH="{number($box[4]) - number($box[2])}" VPOS="{$box[3]}" HPOS="{$box[2]}">
+
+          <xsl:apply-templates/>
+
+      </TopMargin>
+  </xsl:template>
+
+
+  <xsl:template match="*:div[@class='ocr_footer']">
+      <xsl:variable name="box" select="tokenize(mf:getBox(@title), ' ')"/>
+      <BottomMargin ID="{@id}" HEIGHT="{number($box[5]) - number($box[3])}" WIDTH="{number($box[4]) - number($box[2])}" VPOS="{$box[3]}" HPOS="{$box[2]}">
+
+          <xsl:apply-templates/>
+
+      </BottomMargin>
+  </xsl:template>
 
 
   <xsl:template match="*:div[@class='ocr_carea']">
@@ -154,23 +182,27 @@ License: MIT
   </xsl:template>
 
 
+
 <xsl:function name="mf:getFname">
     <xsl:param name="titleString"/>
     <xsl:variable name="pPat">"</xsl:variable>
     <xsl:variable name="fpath" select="substring-after(tokenize(normalize-space($titleString),'; ')[1],'image &quot;')"/>
     <xsl:value-of select="reverse(tokenize(replace($fpath,$pPat,''),'\\'))[1]"/>
 </xsl:function>
 
+
 <xsl:function name="mf:getBoxPage">
     <xsl:param name="titleString"/>
     <xsl:value-of select="tokenize(normalize-space($titleString),'; ')[2]"/>
 </xsl:function>
 
+
 <xsl:function name="mf:getBox">
     <xsl:param name="titleString"/>
     <xsl:value-of select="tokenize(normalize-space($titleString),'; ')"/>
 </xsl:function>
 
+
 <xsl:function name="mf:getConfidence">
     <xsl:param name="titleString"/>
     <xsl:variable name="wconfString" select="tokenize(normalize-space($titleString),'; ')[2]" />
@@ -187,4 +219,5 @@ License: MIT
 
 </xsl:function>
 
+
 </xsl:stylesheet>
diff --git a/hocr__alto3.xsl b/hocr__alto3.xsl
@@ -25,11 +25,12 @@ License: MIT
   <xsl:variable name="langcodes" select="document('codes_lookup.xml')/*:codes/*:code" />
 
   <xsl:template match="/">
-        <alto xsi:schemaLocation="http://www.loc.gov/standards/alto/ns-v3# https://www.loc.gov/standards/alto/v3/alto.xsd">
+        <alto xsi:schemaLocation="http://www.loc.gov/standards/alto/ns-v4# https://www.loc.gov/standards/alto/v4/alto.xsd">
             <xsl:apply-templates/>
         </alto>
   </xsl:template>
 
+
   <xsl:template match="*:head">
         <Description>
             <MeasurementUnit>pixel</MeasurementUnit>
@@ -53,17 +54,44 @@ License: MIT
             <xsl:apply-templates select="*:div[@class='ocr_page']"/>
         </Layout>
   </xsl:template>
+
+
   <xsl:template match="*:div[@class='ocr_page']">
         <!--  bbox 552 999 1724 1141 x1-L2-T3-R4-B5 -->
         <xsl:variable name="box" select="tokenize(mf:getBoxPage(@title), ' ')"/>
         <Page ID="{@id}" PHYSICAL_IMG_NR="1" HEIGHT="{$box[5]}" WIDTH="{$box[4]}">
-            <PrintSpace HEIGHT="{$box[5]}" WIDTH="{$box[4]}" VPOS="0" HPOS="0">
 
-                <xsl:apply-templates/>
+            <xsl:apply-templates select="*:div[@class='ocr_header']"/>
 
+            <PrintSpace HEIGHT="{$box[5]}" WIDTH="{$box[4]}" VPOS="0" HPOS="0">
+                <xsl:apply-templates select="*:div[@class='ocr_carea']"/>
+                <xsl:apply-templates select="*:p[@class='ocr_par']"/>
             </PrintSpace>
+
+            <xsl:apply-templates select="*:div[@class='ocr_footer']"/>
+
         </Page>
   </xsl:template>
+
+
+  <xsl:template match="*:div[@class='ocr_header']">
+      <xsl:variable name="box" select="tokenize(mf:getBox(@title), ' ')"/>
+      <TopMargin ID="{@id}" HEIGHT="{number($box[5]) - number($box[3])}" WIDTH="{number($box[4]) - number($box[2])}" VPOS="{$box[3]}" HPOS="{$box[2]}">
+
+          <xsl:apply-templates/>
+
+      </TopMargin>
+  </xsl:template>
+
+
+  <xsl:template match="*:div[@class='ocr_footer']">
+      <xsl:variable name="box" select="tokenize(mf:getBox(@title), ' ')"/>
+      <BottomMargin ID="{@id}" HEIGHT="{number($box[5]) - number($box[3])}" WIDTH="{number($box[4]) - number($box[2])}" VPOS="{$box[3]}" HPOS="{$box[2]}">
+
+          <xsl:apply-templates/>
+
+      </BottomMargin>
+  </xsl:template>
 
 
   <xsl:template match="*:div[@class='ocr_carea']">
@@ -154,23 +182,27 @@ License: MIT
   </xsl:template>
 
 
+
 <xsl:function name="mf:getFname">
     <xsl:param name="titleString"/>
     <xsl:variable name="pPat">"</xsl:variable>
     <xsl:variable name="fpath" select="substring-after(tokenize(normalize-space($titleString),'; ')[1],'image &quot;')"/>
     <xsl:value-of select="reverse(tokenize(replace($fpath,$pPat,''),'\\'))[1]"/>
 </xsl:function>
 
+
 <xsl:function name="mf:getBoxPage">
     <xsl:param name="titleString"/>
     <xsl:value-of select="tokenize(normalize-space($titleString),'; ')[2]"/>
 </xsl:function>
 
+
 <xsl:function name="mf:getBox">
     <xsl:param name="titleString"/>
     <xsl:value-of select="tokenize(normalize-space($titleString),'; ')"/>
 </xsl:function>
 
+
 <xsl:function name="mf:getConfidence">
     <xsl:param name="titleString"/>
     <xsl:variable name="wconfString" select="tokenize(normalize-space($titleString),'; ')[2]" />
@@ -187,4 +219,5 @@ License: MIT
 
 </xsl:function>
 
+
 </xsl:stylesheet>