Merge sibling <g> nodes with identical attributes

In some cases, gnuplot generates a very suboptimal SVG content of the
following pattern:

        <g color="black" fill="none" stroke="currentColor">
        <path d="m82.5 323.3v-4.1" stroke="#000"/>
        </g>
        <g color="black" fill="none" stroke="currentColor">
        <path d="m116.4 323.3v-4.1" stroke="#000"/>
        </g>
        ... repeated 10+ more times here ...
        <g color="black" fill="none" stroke="currentColor">
        <path d="m65.4 72.8v250.5h420v-250.5h-420z" stroke="#000"/>
        </g>

A more optimal pattern would be:

        <g color="black" fill="none" stroke="#000">
        <path d="m82.5 323.3v-4.1"/>
        <path d="m116.4 323.3v-4.1"/>
        ... 10+ more paths here ...
        <path d="m65.4 72.8v250.5h420v-250.5h-420z"/>
        </g>

This patch enables that optimization by handling the merging of two
sibling <g> entries that have identical attributes.  In the above
example that does not solve the rewrite from "currentColor" to "#000"
for the stroke attribute.  However, the existing code already handles
that automatically after the <g> elements have been merged.

This change provides comparable results to --create-groups as shown by
the following diagram while being a distinct optimization:

       +----------------------------+-------+--------+
       |           Test             | Size  |  in %  |
       +----------------------------+-------+--------+
       | baseline                   | 17961 |  100%  |
       | baseline + --create-groups | 17418 |  97.0% |
       | patched                    | 16939 |  94.3% |
       | patched + --create-groups  | 16855 |  93.8% |
       +----------------------------+-------+--------+

The image used in the size table above was generated based on the
instructions from https://bugs.debian.org/858039#10 with gnuplot 5.2
patchlevel 2.  Beyond the test-based "--create-groups", the following
scour command-line parameters were used:
      --enable-id-stripping --enable-comment-stripping \
      --shorten-ids --indent=none

Note that the baseline was scour'ed repeatedly to stablize the image
size.

Signed-off-by: Niels Thykier <niels@thykier.net>

scour/scour.py

@@ -1134,6 +1134,75 @@ def moveCommonAttributesToParentGroup(elem, referencedElements):
     return num
 
 
+def mergeSiblingGroupsWithCommonAttributes(elem):
+    """
+    Merge two or more sibling <g> elements with the identical attributes.
+
+    This function acts recursively on the given element.
+    """
+
+    num = 0
+    i = elem.childNodes.length - 1
+    while i >= 0:
+        currentNode = elem.childNodes.item(i)
+        if currentNode.nodeType != Node.ELEMENT_NODE or currentNode.nodeName != 'g' or \
+           currentNode.namespaceURI != NS['SVG']:
+            i -= 1
+            continue
+        attributes = {a.nodeName: a.nodeValue for a in currentNode.attributes.values()}
+        runStart, runEnd = i, i
+        runElements = 1
+        while runStart > 0:
+            nextNode = elem.childNodes.item(runStart - 1)
+            if nextNode.nodeType == Node.ELEMENT_NODE:
+                if nextNode.nodeName != 'g' or nextNode.namespaceURI != NS['SVG']:
+                    break
+                nextAttributes = {a.nodeName: a.nodeValue for a in nextNode.attributes.values()}
+                hasNoMergeTags = (True for n in nextNode.childNodes
+                                  if n.nodeType == Node.ELEMENT_NODE
+                                  and n.nodeName in ('title', 'desc')
+                                  and n.namespaceURI == NS['SVG'])
+                if attributes != nextAttributes or any(hasNoMergeTags):
+                    break
+                else:
+                    runElements += 1
+                    runStart -= 1
+            else:
+                runStart -= 1
+
+        # Next loop will start from here
+        i = runStart - 1
+
+        if runElements < 2:
+            continue
+
+        # Find the <g> entry that starts the run (we might have run
+        # past it into a text node or a comment node.
+        while True:
+            node = elem.childNodes.item(runStart)
+            if node.nodeType == Node.ELEMENT_NODE and node.nodeName == 'g' and node.namespaceURI == NS['SVG']:
+                break
+            runStart += 1
+        primaryGroup = elem.childNodes.item(runStart)
+        runStart += 1
+        nodes = elem.childNodes[runStart:runEnd+1]
+        for node in nodes:
+            if node.nodeType == Node.ELEMENT_NODE and node.nodeName == 'g' and node.namespaceURI == NS['SVG']:
+                # Merge
+                primaryGroup.childNodes.extend(node.childNodes)
+                node.childNodes = []
+            else:
+                primaryGroup.childNodes.append(node)
+            elem.childNodes.remove(node)
+
+    # each child gets the same treatment, recursively
+    for childNode in elem.childNodes:
+        if childNode.nodeType == Node.ELEMENT_NODE:
+            num += mergeSiblingGroupsWithCommonAttributes(childNode)
+
+    return num
+
+
 def createGroupsForCommonAttributes(elem):
     """
     Creates <g> elements to contain runs of 3 or more
@@ -3658,6 +3727,8 @@ def scourString(in_string, options=None):
     while removeDuplicateGradients(doc) > 0:
         pass
 
+    if options.group_collapse:
+        _num_elements_removed += mergeSiblingGroupsWithCommonAttributes(doc.documentElement)
     # create <g> elements if there are runs of elements with the same attributes.
     # this MUST be before moveCommonAttributesToParentGroup.
     if options.group_create:

testscour.py

@@ -2075,6 +2075,21 @@ class MustKeepGInSwitch2(unittest.TestCase):
                          'Erroneously removed a <g> in a <switch>')
 
 
+class GroupSiblingMerge(unittest.TestCase):
+
+    def test_sibling_merge(self):
+        doc = scourXmlFile('unittests/group-sibling-merge.svg',
+                           parse_args([]))
+        self.assertEqual(doc.getElementsByTagName('g').length, 5,
+                         'Merged sibling <g> tags with similar values')
+
+    def test_sibling_merge_disabled(self):
+        doc = scourXmlFile('unittests/group-sibling-merge.svg',
+                           parse_args(['--disable-group-collapsing']))
+        self.assertEqual(doc.getElementsByTagName('g').length, 8,
+                         'Sibling merging is disabled by --disable-group-collapsing')
+
+
 class GroupCreation(unittest.TestCase):
 
     def runTest(self):

unittests/group-sibling-merge.svg

@@ -0,0 +1,29 @@
+<?xml version="1.0" encoding="UTF-8" standalone="no"?>
+<svg xmlns="http://www.w3.org/2000/svg">
+<desc>Produced by GNUPLOT 5.2 patchlevel 8</desc>
+<rect width="900" height="600" fill="none"/>
+<g color="black" fill="none">
+<path d="m88.5 564h9m777.5 0h-9" stroke="#000"/>
+<g transform="translate(80.2,567.9)" fill="#000" font-family="Arial" font-size="12" text-anchor="end">
+<text><tspan font-family="Arial">0</tspan></text>
+</g>
+</g>
+<g color="black" fill="none">
+<path d="m88.5 473h9m777.5 0h-9" stroke="#000"/>
+<g transform="translate(80.2,476.9)" fill="#000" font-family="Arial" font-size="12" text-anchor="end">
+<text><tspan font-family="Arial">5000</tspan></text>
+</g>
+</g>
+<g color="black" fill="none">
+<path d="m88.5 382h9m777.5 0h-9" stroke="#000"/>
+<g transform="translate(80.2,385.9)" fill="#000" font-family="Arial" font-size="12" text-anchor="end">
+<text><tspan font-family="Arial">10000</tspan></text>
+</g>
+</g>
+<g color="black" fill="none">
+<path d="m88.5 291h9m777.5 0h-9" stroke="#000"/>
+<g transform="translate(80.2,294.9)" fill="#000" font-family="Arial" font-size="12" text-anchor="end">
+<text><tspan font-family="Arial">15000</tspan></text>
+</g>
+</g>
+</svg>