Project

General

Profile

1 21 aaronmk
# XML DOM tree manipulation
2
3 73 aaronmk
import cgi
4
from HTMLParser import HTMLParser
5 3632 aaronmk
import itertools
6 1809 aaronmk
import re
7 21 aaronmk
from xml.dom import Node
8 299 aaronmk
import xml.dom.minidom as minidom
9 21 aaronmk
10 4331 aaronmk
import lists
11 73 aaronmk
import strings
12 331 aaronmk
import util
13 73 aaronmk
14 840 aaronmk
##### Escaping input
15
16 73 aaronmk
def escape(str_):
17
    return strings.to_unicode(cgi.escape(str_, True)).encode('ascii',
18
        'xmlcharrefreplace')
19
20
def unescape(str_): return HTMLParser().unescape(str_)
21
22 1814 aaronmk
##### Names
23
24
def strip_namespace(name):
25
    namespace, sep, base = name.partition(':')
26
    if sep != '': return base
27
    else: return name
28
29 2431 aaronmk
##### Nodes
30
31
def is_node(value): return isinstance(value, Node)
32
33 2008 aaronmk
##### Replacing a node
34
35
def remove(node): node.parentNode.removeChild(node)
36
37
def replace(old, new):
38
    '''@param new Node|None'''
39 3329 aaronmk
    assert old.parentNode != None # not removed from parent tree
40 4758 aaronmk
    if new == None: new = []
41 4331 aaronmk
    else: new = lists.mk_seq(new)
42 3329 aaronmk
43 4331 aaronmk
    olds_parent = old.parentNode
44
    if not new: olds_parent.removeChild(old)
45
    else: # at least one node
46
        last = new.pop(-1)
47
        olds_parent.replaceChild(last, old) # note arg order reversed
48
        for node in reversed(new):
49
            olds_parent.insertBefore(node, last) # there is no insertAfter()
50 2008 aaronmk
51 4234 aaronmk
def bool2str(val):
52
    '''For use with replace_with_text()'''
53
    if val: return '1'
54
    else: return None # remove node
55
56 2008 aaronmk
def replace_with_text(node, new):
57 3225 aaronmk
    '''
58
    @return The *new* node
59
    '''
60 4235 aaronmk
    if isinstance(new, bool): new = bool2str(new)
61 4757 aaronmk
    elif isinstance(new, int) or isinstance(new, float): new = str(new)
62 2008 aaronmk
    if util.is_str(new): new = node.ownerDocument.createTextNode(new)
63
    replace(node, new)
64 3225 aaronmk
    return new
65 2008 aaronmk
66 840 aaronmk
##### Element node contents
67 455 aaronmk
68 840 aaronmk
def is_elem(node): return node.nodeType == Node.ELEMENT_NODE
69
70 453 aaronmk
def is_completely_empty(node): return node.firstChild == None
71 135 aaronmk
72 301 aaronmk
def has_one_child(node):
73
    return node.firstChild != None and node.firstChild.nextSibling == None
74
75 4330 aaronmk
def only_child(node):
76
    if not has_one_child(node):
77 4491 aaronmk
        raise Exception('Must contain only one child:\n'+strings.ustr(node))
78 4330 aaronmk
    return node.firstChild
79
80 840 aaronmk
def is_simple(node):
81
    '''Whether every child recursively has no more than one child'''
82
    return (not is_elem(node) or is_completely_empty(node)
83
        or (has_one_child(node) and is_simple(node.firstChild)))
84
85 305 aaronmk
class NodeIter:
86
    def __init__(self, node): self.child = node.firstChild
87
88
    def __iter__(self): return self
89
90
    def curr(self):
91
        if self.child != None: return self.child
92
        raise StopIteration
93
94
    def next(self):
95
        child = self.curr()
96
        self.child = self.child.nextSibling
97
        return child
98
99 840 aaronmk
##### Comments
100 455 aaronmk
101 453 aaronmk
def is_comment(node): return node.nodeType == Node.COMMENT_NODE
102
103
def is_empty(node):
104
    for child in NodeIter(node):
105 4038 aaronmk
        if not (is_whitespace(child) or is_comment(child)): return False
106 453 aaronmk
    return True
107
108 1809 aaronmk
def clean_comment(str_):
109
    '''Sanitizes comment node contents. Strips invalid strings.'''
110
    return re.sub(r'-{2,}', r'-', str_) # comments can't contain '--'
111
112
def mk_comment(doc, str_): return doc.createComment(clean_comment(str_))
113
114 840 aaronmk
##### Child nodes that are elements
115 455 aaronmk
116 21 aaronmk
class NodeElemIter:
117
    def __init__(self, node): self.child = node.firstChild
118
119
    def __iter__(self): return self
120
121
    def curr(self):
122
        while self.child != None:
123 298 aaronmk
            if is_elem(self.child): return self.child
124 21 aaronmk
            self.child = self.child.nextSibling
125
        raise StopIteration
126
127
    def next(self):
128
        child = self.curr()
129
        self.child = self.child.nextSibling
130
        return child
131
132
def first_elem(node): return NodeElemIter(node).next()
133
134 450 aaronmk
def has_elems(node):
135
    try: first_elem(node); return True
136
    except StopIteration: return False
137
138 21 aaronmk
class NodeElemReverseIter:
139
    def __init__(self, node): self.child = node.lastChild
140
141
    def __iter__(self): return self
142
143
    def curr(self):
144
        while self.child != None:
145 298 aaronmk
            if is_elem(self.child): return self.child
146 21 aaronmk
            self.child = self.child.previousSibling
147
        raise StopIteration
148
149
    def next(self):
150
        child = self.curr()
151
        self.child = self.child.previousSibling
152
        return child
153
154
def last_elem(node): return NodeElemReverseIter(node).next()
155
156 4238 aaronmk
class NodeEntryIter:
157
    def __init__(self, node): self.iter_ = NodeElemIter(node)
158
159
    def __iter__(self): return self
160
161
    def next(self):
162
        entry = self.iter_.next()
163 4332 aaronmk
        values = list(NodeIter(entry))
164
        if not values: values = None
165
        elif len(values) == 1: values = values[0]
166
        return (entry.tagName, values)
167 4238 aaronmk
168 840 aaronmk
##### Parent nodes
169
170 1017 aaronmk
def parent(node):
171 999 aaronmk
    '''Does not treat the document object as the root node's parent, since it's
172
    not a true element node'''
173 1017 aaronmk
    parent_ = node.parentNode
174
    if parent_ != None and is_elem(parent_): return parent_
175
    else: return None
176
177
class NodeParentIter:
178
    '''See parent() for special treatment of root node.
179
    Note that the first element returned is the current node, not its parent.'''
180 21 aaronmk
    def __init__(self, node): self.node = node
181
182
    def __iter__(self): return self
183
184
    def curr(self):
185 1017 aaronmk
        if self.node != None: return self.node
186
        else: raise StopIteration
187 21 aaronmk
188
    def next(self):
189
        node = self.curr()
190 1017 aaronmk
        self.node = parent(self.node)
191 21 aaronmk
        return node
192
193 840 aaronmk
##### Element nodes containing text
194 661 aaronmk
195 298 aaronmk
def is_text_node(node): return node.nodeType == Node.TEXT_NODE
196
197 301 aaronmk
def is_text(node): return has_one_child(node) and is_text_node(node.firstChild)
198 21 aaronmk
199 2008 aaronmk
def value_node(node):
200 1832 aaronmk
    if is_elem(node):
201
        iter_ = NodeIter(node)
202
        util.skip(iter_, is_comment)
203 2008 aaronmk
        try: return iter_.next()
204 1832 aaronmk
        except StopIteration: return None
205 2008 aaronmk
    else: return node
206 21 aaronmk
207 2008 aaronmk
def value(node):
208 4029 aaronmk
    value_ = util.do_ignore_none(lambda n: n.nodeValue, value_node(node))
209 4034 aaronmk
    if value_ == strings.isspace_none_str: value_ = None # None equiv
210 4029 aaronmk
    return value_
211 2008 aaronmk
212 4031 aaronmk
def is_whitespace(node):
213
    return is_text_node(node) and (node.nodeValue == ''
214
        or node.nodeValue.isspace())
215 1720 aaronmk
216 143 aaronmk
def set_value(node, value):
217 2008 aaronmk
    value_node_ = value_node(node)
218
    if value != None:
219
        if value_node_ != None:
220
            value_node_.nodeValue = value
221
        else:
222
            assert is_elem(node)
223
            node.appendChild(node.ownerDocument.createTextNode(value))
224 2009 aaronmk
    elif value_node_ != None:
225 4029 aaronmk
        if is_elem(node): remove(value_node_)
226
        else:
227 4033 aaronmk
            if is_text_node(node): value = strings.isspace_none_str # None equiv
228 4029 aaronmk
            node.nodeValue = value
229 22 aaronmk
230 86 aaronmk
class NodeTextEntryIter:
231
    def __init__(self, node): self.iter_ = NodeElemIter(node)
232
233
    def __iter__(self): return self
234
235 1176 aaronmk
    def next(self):
236
        entry = self.iter_.next()
237 3632 aaronmk
        if is_empty(entry): value_ = None
238
        elif is_text(entry): value_ = value(entry)
239 4328 aaronmk
        else: value_ = only_child(entry)
240 839 aaronmk
        return (entry.tagName, value_)
241 86 aaronmk
242 792 aaronmk
def is_text_node_entry(val): return util.is_str(val[1])
243
244 3632 aaronmk
def non_empty(iterable):
245
    return itertools.ifilter(lambda i: i[1] != None, iterable)
246
247 792 aaronmk
class TextEntryOnlyIter(util.CheckedIter):
248
    def __init__(self, iterable):
249 3632 aaronmk
        util.CheckedIter.__init__(self, is_text_node_entry, non_empty(iterable))
250 792 aaronmk
251 1836 aaronmk
##### IDs
252
253
def get_id(node):
254
    '''If the node doesn't have an ID, assumes the node itself is the ID.
255
    @return None if the node doesn't have an ID or a value
256
    '''
257
    id_ = node.getAttribute('id')
258
    if id_ != '': return id_
259
    else: return value(node) # assume the node itself is the ID
260
261
def set_id(node, id_): node.setAttribute('id', id_)
262
263 2008 aaronmk
##### Child nodes
264 455 aaronmk
265 4318 aaronmk
def prune_empty(node):
266
    '''Removes node if it's empty'''
267
    if is_empty(node): remove(node)
268
269
def prune_children(node):
270
    '''Removes empty children'''
271
    for child in NodeElemIter(node): prune_empty(child)
272
273 4321 aaronmk
def prune(node):
274
    '''Removes empty children and then node if it's empty'''
275
    prune_children(node)
276
    prune_empty(node)
277
278 135 aaronmk
def set_child(node, name, value):
279
    '''Note: does not remove any existing child of the same name'''
280
    child = node.ownerDocument.createElement(name)
281 143 aaronmk
    set_value(child, value)
282 135 aaronmk
    node.appendChild(child)
283
284 1814 aaronmk
def by_tag_name(node, name, last_only=False, ignore_namespace=False):
285
    '''last_only optimization returns only the last matching node'''
286
    if ignore_namespace: filter_name = strip_namespace
287
    else: filter_name = lambda name: name
288
    name = filter_name(name)
289
290 22 aaronmk
    children = []
291 888 aaronmk
    if last_only: iter_ = NodeElemReverseIter(node)
292
    else: iter_ = NodeElemIter(node)
293
    for child in iter_:
294 1814 aaronmk
        if filter_name(child.tagName) == name:
295 22 aaronmk
            children.append(child)
296
            if last_only: break
297
    return children
298 28 aaronmk
299 3226 aaronmk
def merge(from_, into):
300 3331 aaronmk
    '''Merges two nodes of the same tag name and their newly-adjacent children.
301
    @post The into node is saved; the from_ node is deleted.
302
    '''
303
    if from_ == None or into == None: return # base case
304
    if from_.tagName != into.tagName: return # not mergeable
305
306
    from_first = from_.firstChild # save before merge
307 3226 aaronmk
    for child in NodeIter(from_): into.appendChild(child)
308
    remove(from_)
309
310 3331 aaronmk
    # Recurse
311
    merge(from_first, from_first.previousSibling) # = into.lastChild
312 3226 aaronmk
313 3331 aaronmk
def merge_by_name(root, name):
314
    '''Merges siblings in root with the given name'''
315
    children = by_tag_name(root, name)
316
    child0 = children.pop(0)
317
    for child in children: merge(child, child0)
318
319 840 aaronmk
##### XML documents
320 455 aaronmk
321 133 aaronmk
def create_doc(root='_'):
322 303 aaronmk
    return minidom.getDOMImplementation().createDocument(None, root, None)
323 133 aaronmk
324 840 aaronmk
##### Printing XML
325 455 aaronmk
326 304 aaronmk
prettyxml_config = dict(addindent='    ', newl='\n')
327 331 aaronmk
toprettyxml_config = prettyxml_config.copy()
328
util.rename_key(toprettyxml_config, 'addindent', 'indent')
329 304 aaronmk
330 840 aaronmk
##### minidom modifications
331 455 aaronmk
332 1720 aaronmk
#### Module
333
334 299 aaronmk
minidom._write_data = lambda writer, data: writer.write(escape(data))
335 73 aaronmk
336 305 aaronmk
minidom.Node.__iter__ = lambda self: NodeIter(self)
337
338 857 aaronmk
def __Node_str(self):
339 860 aaronmk
    return strings.remove_extra_newl(self.toprettyxml(**toprettyxml_config))
340 796 aaronmk
minidom.Node.__str__ = __Node_str
341
minidom.Node.__repr__ = __Node_str
342
minidom.Element.__repr__ = __Node_str
343 301 aaronmk
344 1720 aaronmk
#### Node
345 888 aaronmk
346 315 aaronmk
minidom.Node.pop = lambda self: self.removeChild(self.lastChild)
347
348
def __Node_clear(self):
349
    while not is_empty(self): self.pop()
350
minidom.Node.clear = __Node_clear
351
352 1720 aaronmk
#### Text
353
354
__Text_writexml_orig = minidom.Text.writexml
355
def __Text_writexml(self, *args, **kw_args):
356
    if is_whitespace(self): pass # we add our own whitespace
357
    else: __Text_writexml_orig(self, *args, **kw_args)
358
minidom.Text.writexml = __Text_writexml
359
360
#### Attr
361
362
def __Attr_str(self): return escape(self.name)+'="'+escape(self.value)+'"'
363
minidom.Attr.__str__ = __Attr_str
364
minidom.Attr.__repr__ = __Attr_str
365
366
#### Element
367
368 301 aaronmk
def __Element_write_opening(self, writer, indent='', addindent='', newl=''):
369 1563 aaronmk
    writer.write(indent+'<'+escape(self.tagName))
370 298 aaronmk
    for attr_idx in xrange(self.attributes.length):
371 4491 aaronmk
        writer.write(' '+strings.ustr(self.attributes.item(attr_idx)))
372 298 aaronmk
    writer.write('>'+newl)
373 301 aaronmk
minidom.Element.write_opening = __Element_write_opening
374 73 aaronmk
375 301 aaronmk
def __Element_write_closing(self, writer, indent='', addindent='', newl=''):
376 1563 aaronmk
    writer.write('</'+escape(self.tagName)+'>'+newl)
377 301 aaronmk
minidom.Element.write_closing = __Element_write_closing
378 298 aaronmk
379 1720 aaronmk
__Element_writexml_orig = minidom.Element.writexml
380 301 aaronmk
def __Element_writexml(self, writer, indent='', addindent='', newl=''):
381 306 aaronmk
    if isinstance(indent, int): indent = addindent*indent
382 774 aaronmk
    if is_simple(self):
383
        writer.write(indent)
384 1720 aaronmk
        __Element_writexml_orig(self, writer)
385 774 aaronmk
        writer.write(newl)
386 1720 aaronmk
    else: __Element_writexml_orig(self, writer, indent, addindent, newl)
387 301 aaronmk
minidom.Element.writexml = __Element_writexml
388 28 aaronmk
389 1720 aaronmk
#### Document
390
391 301 aaronmk
def __Document_write_opening(self, writer, indent='', addindent='', newl='',
392
    encoding=None):
393
    xmlDecl = '<?xml version="1.0" '
394
    if encoding != None: xmlDecl += 'encoding="'+escape(encoding)+'"'
395
    xmlDecl += '?>'+newl
396
    writer.write(xmlDecl)
397
    assert has_one_child(self)
398
    assert is_elem(self.firstChild)
399
    self.firstChild.write_opening(writer, indent, addindent, newl)
400
minidom.Document.write_opening = __Document_write_opening
401
402
def __Document_write_closing(self, writer, indent='', addindent='', newl=''):
403
    self.firstChild.write_closing(writer, indent, addindent, newl)
404
minidom.Document.write_closing = __Document_write_closing