From 7c7f71224f1e1c6ae40c4cbfac54836907e561b4 Mon Sep 17 00:00:00 2001
From: Sunil Mohan Adapa <sunil@medhas.org>
Date: Thu, 27 Aug 2020 13:23:15 -0700
Subject: [PATCH] doc: wikiparser: Incorporate post processing fixes

Signed-off-by: Sunil Mohan Adapa <sunil@medhas.org>
Reviewed-by: James Valleroy <jvalleroy@mailbox.org>
---
 doc/Makefile                       |  8 +--
 doc/scripts/fixes.xslt             | 92 ------------------------------
 doc/scripts/manual-page-fixes.xslt | 84 ---------------------------
 doc/scripts/wikiparser.py          | 26 +++++++--
 4 files changed, 26 insertions(+), 184 deletions(-)
 delete mode 100644 doc/scripts/fixes.xslt
 delete mode 100644 doc/scripts/manual-page-fixes.xslt

diff --git a/doc/Makefile b/doc/Makefile
index 12781c817..f1ec5d05f 100644
--- a/doc/Makefile
+++ b/doc/Makefile
@@ -113,15 +113,15 @@ $(manual-pages-part-html): %.part.html: %.html
 $(manual-raw-xmls): %.raw.xml: %.raw.wiki $(SCRIPTS_DIR)/wikiparser.py
 	python3 $(SCRIPTS_DIR)/wikiparser.py $< > $@
 
-$(manual-xmls): %.xml: %.raw.xml $(SCRIPTS_DIR)/fixes.xslt
-	xsltproc --output $@ $(SCRIPTS_DIR)/fixes.xslt $<
+$(manual-xmls): %.xml: %.raw.xml
+	xmllint --format $< > $@
 
 $(manual-pages-raw-xml): %.raw.xml: %.raw.wiki $(SCRIPTS_DIR)/wikiparser.py
 	python3 $(SCRIPTS_DIR)/wikiparser.py --begin-marker="## BEGIN_INCLUDE" \
 	    --end-marker='## END_INCLUDE' $< > $@
 
-$(manual-pages-xml): %.xml: %.raw.xml $(SCRIPTS_DIR)/manual-page-fixes.xslt
-	xsltproc --output $@ $(SCRIPTS_DIR)/manual-page-fixes.xslt $<
+$(manual-pages-xml): %.xml: %.raw.xml
+	xmllint --format $< > $@
 
 $(manual-pages-html): %.html: %.xml
 	xsltproc --output $@ /usr/share/xml/docbook/stylesheet/docbook-xsl/xhtml5/docbook.xsl $<
diff --git a/doc/scripts/fixes.xslt b/doc/scripts/fixes.xslt
deleted file mode 100644
index 752a3a637..000000000
--- a/doc/scripts/fixes.xslt
+++ /dev/null
@@ -1,92 +0,0 @@
-<?xml version="1.0" encoding="UTF-8"?>
-<!--
-# SPDX-License-Identifier: AGPL-3.0-or-later
--->
-<xsl:stylesheet version="1.0"
-                xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
-
-  <!-- Set output format to Docbook -->
-  <xsl:output
-      method="xml"
-      encoding="utf-8"
-      indent="yes"
-      doctype-public="-//OASIS//DTD DocBook XML V4.4//EN"
-      doctype-system="http://www.docbook.org/xml/4.4/docbookx.dtd"/>
-
-  <!-- Copy all nodes and attributes by default -->
-  <xsl:template match="@*|node()">
-    <xsl:copy>
-      <xsl:apply-templates select="@*|node()"/>
-    </xsl:copy>
-  </xsl:template>
-
-  <!-- Replace / in title with a space -->
-  <xsl:template match="articleinfo/title/text()">
-    <xsl:call-template name="clean-title">
-      <xsl:with-param name="title" select="."/>
-    </xsl:call-template>
-  </xsl:template>
-
-  <xsl:template name="clean-title">
-    <xsl:param name="title"/>
-    <xsl:choose>
-      <xsl:when test="contains($title, '/')">
-        <!-- Skip copying the language code -->
-        <xsl:choose>
-          <xsl:when test="not(string-length(substring-before($title, '/'))=2)
-                          and
-                          not(string-length(substring-before($title, '/'))=5 and contains(substring-before($title, '/'), '-'))">
-            <xsl:value-of select="substring-before($title, '/')"/>
-            <xsl:text> </xsl:text>
-          </xsl:when>
-        </xsl:choose>
-        <xsl:call-template name="clean-title">
-          <xsl:with-param name="title" select="substring-after($title, '/')"/>
-        </xsl:call-template>
-      </xsl:when>
-      <xsl:otherwise>
-        <xsl:value-of select="$title"/>
-      </xsl:otherwise>
-    </xsl:choose>
-  </xsl:template>
-
-  <!-- Remove revision history -->
-  <xsl:template match="revhistory"/>
-
-  <!-- Convert all image dimensions from pixels to points -->
-  <xsl:template match="@width[parent::imagedata]|@depth[parent::imagedata]">
-    <xsl:attribute name="{name()}">
-      <xsl:value-of select=". div 2"/>
-      <xsl:text>pt</xsl:text>
-    </xsl:attribute>
-  </xsl:template>
-
-  <!-- Convert all image source URLs to relative paths -->
-  <xsl:template match="@fileref[parent::imagedata]">
-    <xsl:attribute name="fileref">
-      <xsl:text>images/</xsl:text>
-      <xsl:call-template name="filename">
-        <xsl:with-param name="path" select="."/>
-      </xsl:call-template>
-    </xsl:attribute>
-  </xsl:template>
-
-  <!-- Output just the filename from a URL -->
-  <xsl:template name="filename">
-    <xsl:param name="path"/>
-    <xsl:choose>
-      <xsl:when test="contains($path, 'target=')">
-        <xsl:value-of select="substring-after($path, 'target=')"/>
-      </xsl:when>
-      <xsl:when test="not(contains($path, '/'))">
-        <xsl:value-of select="$path"/>
-      </xsl:when>
-      <xsl:otherwise>
-        <xsl:call-template name="filename">
-          <xsl:with-param name="path" select="substring-after($path, '/')"/>
-        </xsl:call-template>
-      </xsl:otherwise>
-    </xsl:choose>
-  </xsl:template>
-
-</xsl:stylesheet>
diff --git a/doc/scripts/manual-page-fixes.xslt b/doc/scripts/manual-page-fixes.xslt
deleted file mode 100644
index a9c713bb4..000000000
--- a/doc/scripts/manual-page-fixes.xslt
+++ /dev/null
@@ -1,84 +0,0 @@
-<?xml version="1.0" encoding="UTF-8"?>
-<!--
-# SPDX-License-Identifier: AGPL-3.0-or-later
--->
-<xsl:stylesheet version="1.0"
-                xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
-
-  <!-- Set output format to Docbook -->
-  <xsl:output
-      method="xml"
-      encoding="utf-8"
-      indent="yes"
-      doctype-public="-//OASIS//DTD DocBook XML V4.4//EN"
-      doctype-system="http://www.docbook.org/xml/4.4/docbookx.dtd"/>
-
-  <!-- Copy all nodes and attributes by default -->
-  <xsl:template match="@*|node()">
-    <xsl:copy>
-      <xsl:apply-templates select="@*|node()"/>
-    </xsl:copy>
-  </xsl:template>
-
-  <!-- Split and get the last word in the title -->
-  <xsl:template match="articleinfo/title/text()">
-    <xsl:call-template name="clean-title">
-      <xsl:with-param name="title" select="."/>
-    </xsl:call-template>
-  </xsl:template>
-
-  <xsl:template name="clean-title">
-    <xsl:param name="title"/>
-    <xsl:choose>
-      <xsl:when test="contains($title, '/')">
-        <xsl:text></xsl:text>
-        <xsl:call-template name="clean-title">
-          <xsl:with-param name="title" select="substring-after($title, '/')"/>
-        </xsl:call-template>
-      </xsl:when>
-      <xsl:otherwise>
-        <xsl:value-of select="$title"/>
-      </xsl:otherwise>
-    </xsl:choose>
-  </xsl:template>
-
-  <!-- Remove revision history -->
-  <xsl:template match="revhistory"/>
-
-  <!-- Convert all image dimensions from pixels to points -->
-  <xsl:template match="@width[parent::imagedata]|@depth[parent::imagedata]">
-    <xsl:attribute name="{name()}">
-      <xsl:value-of select=". div 2"/>
-      <xsl:text>pt</xsl:text>
-    </xsl:attribute>
-  </xsl:template>
-
-  <!-- Convert all image source URLs to relative paths -->
-  <xsl:template match="@fileref[parent::imagedata]">
-    <xsl:attribute name="fileref">
-      <xsl:text>images/</xsl:text>
-      <xsl:call-template name="filename">
-        <xsl:with-param name="path" select="."/>
-      </xsl:call-template>
-    </xsl:attribute>
-  </xsl:template>
-
-  <!-- Output just the filename from a URL -->
-  <xsl:template name="filename">
-    <xsl:param name="path"/>
-    <xsl:choose>
-      <xsl:when test="contains($path, 'target=')">
-        <xsl:value-of select="substring-after($path, 'target=')"/>
-      </xsl:when>
-      <xsl:when test="not(contains($path, '/'))">
-        <xsl:value-of select="$path"/>
-      </xsl:when>
-      <xsl:otherwise>
-        <xsl:call-template name="filename">
-          <xsl:with-param name="path" select="substring-after($path, '/')"/>
-        </xsl:call-template>
-      </xsl:otherwise>
-    </xsl:choose>
-  </xsl:template>
-
-</xsl:stylesheet>
diff --git a/doc/scripts/wikiparser.py b/doc/scripts/wikiparser.py
index eae4831e3..fa630efb0 100644
--- a/doc/scripts/wikiparser.py
+++ b/doc/scripts/wikiparser.py
@@ -227,7 +227,7 @@ class EmbeddedAttachment(EmbeddedLink):
 
         xml = '<inlinemediaobject><imageobject>'
         xml += '<imagedata '
-        props = {'fileref': target}
+        props = {'fileref': map_local_files(target)}
         if self.params:
             params = self.params.split(',')
             for param in params:
@@ -235,7 +235,7 @@ class EmbeddedAttachment(EmbeddedLink):
                 if prop == 'height':
                     prop = 'depth'
 
-                props[prop] = value
+                props[prop] = convert_image_units(value)
 
         props = [f'{prop}="{value}"' for prop, value in sorted(props.items())]
         xml += ' '.join(props)
@@ -538,6 +538,24 @@ def get_url_text(url):
     return url
 
 
+def convert_image_units(value):
+    """Covert wiki image units to docbook image units."""
+    value = int(value)
+    value = value / 2.0 if value % 2 else int(value / 2)
+    return str(value) + 'pt'
+
+
+def map_local_files(path):
+    """Map files to locally existing paths."""
+    if 'target=' in path:
+        path = path.partition('target=')[2]
+
+    if '/' in path:
+        path = path.rsplit('/', maxsplit=1)[1]
+
+    return f'images/{path}'
+
+
 def resolve_url(url, context):
     """Expand a URL into a full path."""
     if re.match(r'https?://', url) or url.startswith('mailto:') or \
@@ -1673,13 +1691,13 @@ Let\\'s Encrypt</ulink>'
 
     >>> generate_inner_docbook([EmbeddedAttachment('cockpit-enable.png')])
     '<inlinemediaobject><imageobject>\
-<imagedata fileref="cockpit-enable.png"/></imageobject>\
+<imagedata fileref="images/cockpit-enable.png"/></imageobject>\
 <textobject><phrase>cockpit-enable.png</phrase></textobject>\
 </inlinemediaobject>'
     >>> generate_inner_docbook([EmbeddedAttachment('Backups_Step1_v49.png', \
 [PlainText('Backups: Step 1')], 'width=800')])
     '<inlinemediaobject><imageobject>\
-<imagedata fileref="Backups_Step1_v49.png" width="800"/></imageobject>\
+<imagedata fileref="images/Backups_Step1_v49.png" width="400pt"/></imageobject>\
 <textobject><phrase>Backups: Step 1</phrase></textobject>\
 </inlinemediaobject>'