def _get_image_size(self, el):
"""
If we can't find a height or width, return 0 for whichever is not
found, then rely on the `image` handler to strip those attributes. This
functionality can change once we integrate PIL.
"""
sizes = find_first(el, "ext")
if sizes is not None and sizes.get("cx"):
if sizes.get("cx"):
x = self._convert_image_size(int(sizes.get("cx")))
if sizes.get("cy"):
y = self._convert_image_size(int(sizes.get("cy")))
return ("%dpx" % x, "%dpx" % y)
shape = find_first(el, "shape")
if shape is not None and shape.get("style") is not None:
# If either of these are not set, rely on the method `image` to not
# use either of them.
x = 0
y = 0
styles = shape.get("style").split(";")
for s in styles:
if s.startswith("height:"):
y = s.split(":")[1]
if s.startswith("width:"):
x = s.split(":")[1]
return x, y
return 0, 0
def _get_image_size(self, el):
"""
If we can't find a height or width, return 0 for whichever is not
found, then rely on the `image` handler to strip those attributes. This
functionality can change once we integrate PIL.
"""
sizes = find_first(el, 'ext')
if sizes is not None and sizes.get('cx'):
if sizes.get('cx'):
x = self._convert_image_size(int(sizes.get('cx')))
if sizes.get('cy'):
y = self._convert_image_size(int(sizes.get('cy')))
return (
'%dpx' % x,
'%dpx' % y,
)
shape = find_first(el, 'shape')
if shape is not None and shape.get('style') is not None:
# If either of these are not set, rely on the method `image` to not
# use either of them.
x = 0
y = 0
styles = shape.get('style').split(';')
for s in styles:
if s.startswith('height:'):
y = s.split(':')[1]
if s.startswith('width:'):
x = s.split(':')[1]
return x, y
return 0, 0
def __init__(
self,
path,
convert_root_level_upper_roman=False,
*args,
**kwargs):
self._parsed = ''
self.block_text = ''
self.page_width = 0
self.convert_root_level_upper_roman = convert_root_level_upper_roman
self._image_data = {}
self._build_data(path, *args, **kwargs)
self.pre_processor = None
#divide by 20 to get to pt (Office works in 20th's of a point)
"""
see http://msdn.microsoft.com/en-us/library/documentformat
.openxml.wordprocessing.indentation.aspx
"""
if find_first(self.root, 'pgSz') is not None:
self.page_width = int(
find_first(self.root, 'pgSz').attrib['w']
) / 20
#all blank when we init
self.comment_store = None
self.visited = set()
self.list_depth = 0
self.rels_dict = self._parse_rels_root()
self.styles_dict = self._parse_styles()
self.parse_begin(self.root) # begin to parse
def _get_image_size(self, el):
"""
If we can't find a height or width, return 0 for whichever is not
found, then rely on the `image` handler to strip those attributes. This
functionality can change once we integrate PIL.
"""
sizes = find_first(el, 'ext')
if sizes is not None and sizes.get('cx'):
if sizes.get('cx'):
x = self._convert_image_size(int(sizes.get('cx')))
if sizes.get('cy'):
y = self._convert_image_size(int(sizes.get('cy')))
return (
'%dpx' % x,
'%dpx' % y,
)
shape = find_first(el, 'shape')
if shape is not None and shape.get('style') is not None:
# If either of these are not set, rely on the method `image` to not
# use either of them.
x = 0
y = 0
styles = shape.get('style').split(';')
for s in styles:
if s.startswith('height:'):
y = s.split(':')[1]
if s.startswith('width:'):
x = s.split(':')[1]
return x, y
return 0, 0
def __init__(self,
path,
convert_root_level_upper_roman=False,
*args,
**kwargs):
self._parsed = ''
self.block_text = ''
self.page_width = 0
self.convert_root_level_upper_roman = convert_root_level_upper_roman
self._image_data = {}
self._build_data(path, *args, **kwargs)
self.pre_processor = None
#divide by 20 to get to pt (Office works in 20th's of a point)
"""
see http://msdn.microsoft.com/en-us/library/documentformat
.openxml.wordprocessing.indentation.aspx
"""
if find_first(self.root, 'pgSz') is not None:
self.page_width = int(find_first(self.root,
'pgSz').attrib['w']) / 20
#all blank when we init
self.comment_store = None
self.visited = set()
self.list_depth = 0
self.rels_dict = self._parse_rels_root()
self.styles_dict = self._parse_styles()
self.parse_begin(self.root) # begin to parse
def _get_image_id(self, el):
# Drawings
blip = find_first(el, 'blip')
if blip is not None:
# On drawing tags the id is actually whatever is returned from the
# embed attribute on the blip tag. Thanks a lot Microsoft.
return blip.get('embed')
# Picts
imagedata = find_first(el, 'imagedata')
if imagedata is not None:
return imagedata.get('id')
def _get_image_id(self, el):
# Drawings
blip = find_first(el, 'blip')
if blip is not None:
# On drawing tags the id is actually whatever is returned from the
# embed attribute on the blip tag. Thanks a lot Microsoft.
return blip.get('embed')
# Picts
imagedata = find_first(el, 'imagedata')
if imagedata is not None:
return imagedata.get('id')
def _parse_styles(self):
if self.styles_text is None:
return {}
tree = parse_xml_from_string(self.styles_text)
styles_dict = {}
for style in find_all(tree, "style"):
style_val = find_first(style, "name").attrib["val"]
run_properties = find_first(style, "rPr")
styles_dict[style.attrib["styleId"]] = {
"style_name": style_val,
"default_run_properties": self._parse_run_properties(run_properties),
}
return styles_dict
def test_find_first(self):
root = make_xml(b'<one><two><three v="1"/><three v="2"/></two></one>')
# Can't find the root element
result = find_first(root, 'one')
self.assertEqual(result, None)
result = find_first(root, 'three')
self.assertEqual(result.tag, 'three')
self.assertEqual(result.get('v'), '1')
result = find_first(root, 'two')
self.assertEqual(result.tag, 'two')
def test_find_first(self):
root = make_xml(b'<one><two><three v="1"/><three v="2"/></two></one>')
# Can't find the root element
result = find_first(root, "one")
self.assertEqual(result, None)
result = find_first(root, "three")
self.assertEqual(result.tag, "three")
self.assertEqual(result.get("v"), "1")
result = find_first(root, "two")
self.assertEqual(result.tag, "two")
def _parse_styles(self):
if self.styles_text is None:
return {}
tree = parse_xml_from_string(self.styles_text)
styles_dict = {}
for style in find_all(tree, 'style'):
style_val = find_first(style, 'name').attrib['val']
run_properties = find_first(style, 'rPr')
styles_dict[style.attrib['styleId']] = {
'style_name':
style_val,
'default_run_properties':
self._parse_run_properties(run_properties, ),
}
return styles_dict
def _parse_styles(self):
if self.styles_text is None:
return {}
tree = parse_xml_from_string(self.styles_text)
styles_dict = {}
for style in find_all(tree, 'style'):
style_val = find_first(style, 'name').attrib['val']
run_properties = find_first(style, 'rPr')
styles_dict[style.attrib['styleId']] = {
'style_name': style_val,
'default_run_properties': self._parse_run_properties(
run_properties,
),
}
return styles_dict
def _parse_styles(self):
styles_part = self.document.main_document_part.style_definitions_part
if not styles_part:
return {}
styles_root = styles_part.root_element
styles_dict = {}
for style in find_all(styles_root, 'style'):
style_val = find_first(style, 'name').attrib['val']
run_properties = find_first(style, 'rPr')
styles_dict[style.attrib['styleId']] = {
'style_name':
style_val,
'default_run_properties':
self._parse_run_properties(run_properties, ),
}
return styles_dict
def _parse_styles(self):
styles_part = self.document.main_document_part.style_definitions_part
if not styles_part:
return {}
styles_root = styles_part.root_element
styles_dict = {}
for style in find_all(styles_root, 'style'):
style_val = find_first(style, 'name').attrib['val']
run_properties = find_first(style, 'rPr')
styles_dict[style.attrib['styleId']] = {
'style_name': style_val,
'default_run_properties': self._parse_run_properties(
run_properties,
),
}
return styles_dict
def parse_table_cell(self, el, text):
v_merge = find_first(el, 'vMerge')
if v_merge is not None and ('restart' != v_merge.get('val', '')):
return ''
colspan = self.get_colspan(el)
rowspan = self._get_rowspan(el, v_merge)
if rowspan > 1:
rowspan = str(rowspan)
else:
rowspan = ''
return self.table_cell(text, colspan, rowspan)
def parse_table_cell(self, el, text):
v_merge = find_first(el, "vMerge")
if v_merge is not None and ("restart" != v_merge.get("val", "")):
return ""
colspan = self.get_colspan(el)
rowspan = self._get_rowspan(el, v_merge)
if rowspan > 1:
rowspan = str(rowspan)
else:
rowspan = ""
return self.table_cell(text, colspan, rowspan)
def parse_table_cell(self, el, text):
v_merge = find_first(el, 'vMerge')
if v_merge is not None and (
'restart' != v_merge.get('val', '')):
return ''
colspan = self.get_colspan(el)
rowspan = self._get_rowspan(el, v_merge)
if rowspan > 1:
rowspan = str(rowspan)
else:
rowspan = ''
return self.table_cell(text, colspan, rowspan)
def _load(self):
self.document = WordprocessingDocument(path=self.path)
main_document_part = self.document.main_document_part
if main_document_part is None:
raise MalformedDocxException
self.root_element = main_document_part.root_element
self.numbering_root = None
numbering_part = main_document_part.numbering_definitions_part
if numbering_part:
self.numbering_root = numbering_part.root_element
pgSzEl = find_first(self.root_element, 'pgSz')
if pgSzEl is not None:
# pgSz is defined in twips, convert to points
pgSz = int(pgSzEl.attrib['w'])
self.page_width = pgSz / TWIPS_PER_POINT
self.styles_dict = self._parse_styles()
self.parse_begin(self.root_element)
def _load(self):
self.document = WordprocessingDocument(path=self.path)
main_document_part = self.document.main_document_part
if main_document_part is None:
raise MalformedDocxException
self.root_element = main_document_part.root_element
self.numbering_root = None
numbering_part = main_document_part.numbering_definitions_part
if numbering_part:
self.numbering_root = numbering_part.root_element
pgSzEl = find_first(self.root_element, 'pgSz')
if pgSzEl is not None:
# pgSz is defined in twips, convert to points
pgSz = int(pgSzEl.attrib['w'])
self.page_width = pgSz / TWIPS_PER_POINT
self.styles_dict = self._parse_styles()
self.parse_begin(self.root_element)
def parse_r(self, el, parsed):
"""
Parse the running text.
"""
text = parsed
if not text:
return ''
run_properties = {}
# Get the rPr for the current style, they are the defaults.
p = find_ancestor_with_tag(self.pre_processor, el, 'p')
paragraph_style = find_first(p, 'pStyle')
if paragraph_style is not None:
style = paragraph_style.get('val')
style_defaults = self.styles_dict.get(style, {})
run_properties.update(
style_defaults.get('default_run_properties', {}),
)
# Get the rPr for the current r tag, they are overrides.
run_properties_element = el.find('rPr')
if run_properties_element:
local_run_properties = self._parse_run_properties(
run_properties_element,
)
run_properties.update(local_run_properties)
inline_tag_handlers = {
'b': self.bold,
'i': self.italics,
'u': self.underline,
'caps': self.caps,
'smallCaps': self.small_caps,
'strike': self.strike,
'dstrike': self.strike,
'vanish': self.hide,
'webHidden': self.hide,
}
styles_needing_application = []
for property_name, property_value in run_properties.items():
# These tags are a little different, handle them separately
# from the rest.
# This could be a superscript or a subscript
if property_name == 'vertAlign':
if property_value == 'superscript':
styles_needing_application.append(self.superscript)
elif property_value == 'subscript':
styles_needing_application.append(self.subscript)
else:
if (
property_name in inline_tag_handlers and
self._is_style_on(property_value)
):
styles_needing_application.append(
inline_tag_handlers[property_name],
)
# Apply all the handlers.
for func in styles_needing_application:
text = func(text)
return text
def get_colspan(self, el):
grid_span = find_first(el, "gridSpan")
if grid_span is None:
return ""
return grid_span.attrib["val"]
def get_colspan(self, el):
grid_span = find_first(el, 'gridSpan')
if grid_span is None:
return ''
return grid_span.attrib['val']
def get_colspan(self, el):
grid_span = find_first(el, 'gridSpan')
if grid_span is None:
return ''
return grid_span.attrib['val']
