Update svg tooling.

Tweaked svg_cleaner a bit to remove display none, add option to
strip whitespace.  Stripping viewBox was too aggressive, two of our
emoji have non-default viewBox values, so emit it when x or y
are not zero.

Added script to invoke scour (now must be part of installation).

scour_svg.sh

@@ -0,0 +1,49 @@
+#!/bin/bash
+
+SRC_DIR=""
+DST_DIR=""
+
+SCOUR_ARGS="--strip-xml-prolog --enable-viewboxing --enable-id-stripping --enable-comment-stripping --shorten-ids --no-line-breaks --strip-xml-space"
+
+while [ $# != 0 ]; do
+  case "$1" in
+    -s) SRC_DIR=${2}
+        shift
+        shift
+        ;;
+    -d) DST_DIR=${2}
+        shift
+        shift
+        ;;
+    *) echo "unrecognized arg $1"
+       exit 1
+       ;;
+  esac
+done
+
+if [ -z "$SRC_DIR" ]; then
+  echo "missing source directory"
+  exit 1;
+fi
+
+if [ ! -d "$SRC_DIR" ]; then
+  echo "source dirctory '$SRC_DIR' does not exist"
+  exit 1;
+fi
+
+if [ -z "$DST_DIR" ]; then
+  echo "missing destination directory"
+  exit 1
+fi
+
+if [ ! -d "$DST_DIR" ]; then
+  echo "creating destination directory '$DST_DIR'"
+  mkdir -p "$DST_DIR"
+fi
+
+for file in "$SRC_DIR"/*.svg; do
+  dst="${file##*/}"
+  echo $dst
+  scour $SCOUR_ARGS -i "$file" -o "$DST_DIR/$dst"
+done
+  

svg_cleaner.py

@@ -72,18 +72,19 @@ class _Text_Node(object):
 class SvgCleaner(object):
   """Strip out unwanted parts of an svg file, primarily the xml declaration and
   doctype lines, comments, and some attributes of the outermost <svg> element.
-  The id will be replaced when it is inserted into the font.  viewBox causes
+  The id will be replaced when it is inserted into the font. (viewBox causes
   unwanted scaling when used in a font and its effect is difficult to
-  predict. version is unneeded, xml:space is ignored (we're processing spaces
+  predict, but for outside a font we need to keep it sometimes so we keep it).
+  version is unneeded, xml:space is ignored (we're processing spaces
   so a request to maintain them has no effect).  enable-background appears to
   have no effect.  x and y on the outermost svg element have no effect.  We
   keep width and height, and will elsewhere assume these are the dimensions
   used for the character box."""
 
-  def __init__(self):
+  def __init__(self, strip=False):
     self.reader = SvgCleaner._Reader()
     self.cleaner = SvgCleaner._Cleaner()
-    self.writer = SvgCleaner._Writer()
+    self.writer = SvgCleaner._Writer(strip)
 
   class _Reader(object):
     """Loosely based on fonttools's XMLReader.  This generates a tree of nodes,
@@ -130,7 +131,7 @@ class SvgCleaner(object):
 
   class _Cleaner(object):
     def _clean_elem(self, node):
-      viewBox, width, height = None, None, None
+      viewBox, x, y, width, height = None, None, None, None, None
       nattrs = {}
       for k, v in node.attrs.items():
         if node.name == 'svg' and k in [
@@ -153,14 +154,25 @@ class SvgCleaner(object):
         nattrs[k] = v
 
       if node.name == 'svg':
+        if viewBox:
+          x, y, width, height = viewBox.split()
         if not width or not height:
           if not viewBox:
             raise ValueError('no viewBox, width, or height')
-          width, height = viewBox.split()[2:]
         nattrs['width'] = width
         nattrs['height'] = height
+        # keep for svg use outside of font
+        if viewBox and (int(x) != 0 or int(y) != 0):
+          logging.warn('viewbox "%s" x: %s y: %s' % (viewBox, x, y));
+          nattrs['viewBox'] = viewBox
       node.attrs = nattrs
 
+      # if display:none, skip this and its children
+      style = node.attrs.get('style')
+      if (style and 'display:none' in style) or node.attrs.get('display') == 'none':
+          node.contents = []
+          return
+
       # scan contents. remove any empty text nodes, or empty 'g' element nodes.
       # if a 'g' element has no attrs and only one subnode, replace it with the
       # subnode.
@@ -212,6 +224,9 @@ class SvgCleaner(object):
     """For text nodes, replaces sequences of whitespace with a single space.
     For elements, replaces sequences of whitespace in attributes, and
     removes unwanted attributes from <svg> elements."""
+    def __init__(self, strip):
+      logging.warning('writer strip: %s' % strip);
+      self._strip = strip
 
     def _write_node(self, node, lines, indent):
       """Node is a node generated by _Reader, either a TextNode or an
@@ -222,7 +237,7 @@ class SvgCleaner(object):
         if node.text:
           lines.append(node.text)
       else:
-        margin = '  ' * indent
+        margin = '' if self._strip else '  ' * indent
         line = [margin]
         line.append('<%s' % node.name)
         # custom sort attributes of svg, yes this is a hack
@@ -258,7 +273,7 @@ class SvgCleaner(object):
       # the result.
       lines = []
       self._write_node(root, lines, 0)
-      return '\n'.join(lines)
+      return ''.join(lines) if self._strip else '\n'.join(lines)
 
   def tree_from_text(self, svg_text):
     return self.reader.from_text(svg_text)
@@ -276,7 +291,7 @@ class SvgCleaner(object):
     return self.tree_to_text(tree)
 
 
-def clean_svg_files(in_dir, out_dir, match_pat=None, clean=False):
+def clean_svg_files(in_dir, out_dir, match_pat=None, clean=False, strip=False):
   regex = re.compile(match_pat) if match_pat else None
   count = 0
 
@@ -286,7 +301,7 @@ def clean_svg_files(in_dir, out_dir, match_pat=None, clean=False):
 
   out_dir = tool_utils.ensure_dir_exists(out_dir, clean=clean)
 
-  cleaner = SvgCleaner()
+  cleaner = SvgCleaner(strip)
   for file_name in os.listdir(in_dir):
     if regex and not regex.match(file_name):
       continue
@@ -320,6 +335,9 @@ def main():
       metavar='regex', default=None)
   parser.add_argument(
       '-l', '--loglevel', help='log level name/value', default='warning')
+  parser.add_argument(
+      '-w', '--strip_whitespace', help='remove newlines and indentation',
+      action='store_true')
   args = parser.parse_args()
 
   tool_utils.setup_logging(args.loglevel)
@@ -331,7 +349,8 @@ def main():
     logging.info('Writing output to %s', args.out_dir)
 
   clean_svg_files(
-      args.in_dir, args.out_dir, match_pat=args.regex, clean=args.clean)
+      args.in_dir, args.out_dir, match_pat=args.regex, clean=args.clean,
+      strip=args.strip_whitespace)
 
 
 if __name__ == '__main__':