[pod] xhtml2odt: take into account widths of td and th tags, from 'width' attributes or from the corresponding CSS property within 'style' attributes.

pod/styles_manager.py

@@ -63,7 +63,7 @@ HTML_TEXT_ODT_PARA = 'For XHTML element "%s", you must associate an ' \
                      'style.'
 # ------------------------------------------------------------------------------
 class Style:
-    '''Represents a paragraph style as found in styles.xml in a ODT file.'''
+    '''Represents a paragraph style as found in styles.xml in a ODT file'''
     numberRex = re.compile('(\d+)(.*)')
     def __init__(self, name, family):
         self.name = name
@@ -306,7 +306,7 @@ class StylesManager:
         return res
 
     def styleMatch(self, attrs, matchingAttrs):
-        '''p_match is a dict of attributes found on some HTML element.
+        '''p_attrs is a dict of attributes found on some HTML element.
            p_matchingAttrs is a dict of attributes corresponding to some style.
            This method returns True if p_attrs contains the winning (name,value)
            pairs that match those in p_matchingAttrs. Note that ALL attrs in

pod/xhtml2odt.py

@@ -12,6 +12,7 @@ from appy.pod.odf_parser import OdfEnvironment
 from appy.pod.styles_manager import Style
 from appy.shared.xml_parser import XmlEnvironment, XmlParser, escapeXml
 from appy.shared.utils import WhitespaceCruncher
+from appy.shared.css import CssStyles
 
 # To which ODT tags do HTML tags correspond ?
 HTML_2_ODT = {'h1':'h', 'h2':'h', 'h3':'h', 'h4':'h', 'h5':'h', 'h6':'h',
@@ -68,7 +69,7 @@ class HtmlElement:
         return self
 
     def getOdfTag(self, env):
-        '''Gets the raw ODF tag that corresponds to me.'''
+        '''Gets the raw ODF tag that corresponds to me'''
         res = ''
         if HTML_2_ODT.has_key(self.elem):
             res += '%s:%s' % (env.textNs, HTML_2_ODT[self.elem])
@@ -105,7 +106,7 @@ class HtmlElement:
         # Check elements that can't be found within a paragraph
         if (parentElem.elemType == 'para') and \
            (self.elem in NOT_INSIDE_P_OR_P):
-            # Oups, li->p wrongly considered as a conflict.
+            # Oups, li->p wrongly considered as a conflict
             if (parentElem.elem == 'li') and (self.elem in ('p', 'div')):
                 return ()
             return (parentElem.setConflictual(),)
@@ -464,16 +465,12 @@ class XhtmlEnvironment(XmlEnvironment):
             # If we are in the first row of a table, update columns count
             if not table.firstRowParsed:
                 table.nbOfColumns += colspan
-            if attrs.has_key('width') and (colspan == 1):
+            styles = CssStyles(elem, attrs)
-                # Get the width, keep figures only.
+            if hasattr(styles, 'width') and (colspan == 1):
-                width = ''
-                for c in attrs['width']:
-                    if c.isdigit(): width += c
-                width = int(width)
                 # Ensure self.columnWidths is long enough
                 while (len(table.columnWidths)-1) < table.cellIndex:
                     table.columnWidths.append(None)
-                table.columnWidths[table.cellIndex] = width
+                table.columnWidths[table.cellIndex] = styles.width.value
         return currentElem
 
     def onElementEnd(self, elem):

shared/css.py

@@ -1,3 +1,6 @@
+# ------------------------------------------------------------------------------
+import re
+
 # ------------------------------------------------------------------------------
 def parseStyleAttribute(value, asDict=False):
     '''Returns a list of CSS (name, value) pairs (or a dict if p_asDict is
@@ -11,4 +14,50 @@ def parseStyleAttribute(value, asDict=False):
         if asDict: res[name.strip()] = value.strip()
         else:      res.append( (name.strip(), value.strip()) )
     return res
+
+# ------------------------------------------------------------------------------
+class CssValue:
+    '''Represents a CSS value having unit "px" or "%": value and unit are
+       extracted in attributes of the same name. If no unit is specified, "px"
+       is assumed.'''
+    valueRex = re.compile('(\d+)(%|px)?')
+
+    def __init__(self, value):
+        value, unit = CssValue.valueRex.match(value)
+        if not unit: unit = 'px'
+        self.value = int(value)
+        self.unit = unit
+    def __str__(self): return '%d%s' % (self.value, self.unit)
+    def __repr__(self): return self.__str__()
+
+class CssStyles:
+    '''This class represents a set of styles collected from:
+       * an HTML "style" attribute;
+       * other attributes like "width".
+    '''
+    # The list of CSS properties having a unit (px or %)
+    withUnit = ('width', 'height')
+
+    def __init__(self, elem, attrs):
+        '''Analyses styles as found in p_attrs and sets, for every found style,
+           an attribute on self.'''
+        # First, parse the "style" attr if present
+        if attrs.has_key('style'):
+            styles = parseStyleAttribute(attrs['style'], asDict=True)
+            for name, value in styles.iteritems():
+                if name in CssStyles.withUnit:
+                    value = CssValue(value)
+                setattr(self, name.replace('-', ''), value)
+        # Parse attributes "width" and "height" if present. But they will not
+        # override corresponding attributes from the "styles" attributes if
+        # found.
+        for name in ('width', 'height'):
+            if not hasattr(self, name) and attrs.has_key(name):
+                setattr(self, name, CssValue(attrs[name]))
+
+    def __repr__(self):
+        res = '<CSS'
+        for name, value in self.__dict__.iteritems():
+            res += ' %s:%s' % (name, value)
+        return res + '>'
 # ------------------------------------------------------------------------------