1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
|
#! /usr/bin/env python
"""Promote the IDs from <label/> elements to the enclosing section / chapter /
whatever, then remove the <label/> elements. This allows *ML style internal
linking rather than the bogus LaTeX model.
Note that <label/>s in <title> elements are promoted two steps, since the
<title> elements are artificially created from the section parameter, and the
label really refers to the sectioning construct.
"""
__version__ = '$Revision$'
import errno
import string
import sys
import xml.dom.core
import xml.dom.esis_builder
# Workaround to deal with invalid documents (multiple root elements). This
# does not indicate a bug in the DOM implementation.
#
def get_documentElement(self):
docelem = None
for n in self._node.children:
if n.type == xml.dom.core.ELEMENT:
docelem = xml.dom.core.Element(n, self, self)
return docelem
xml.dom.core.Document.get_documentElement = get_documentElement
# Replace get_childNodes for the Document class; without this, children
# accessed from the Document object via .childNodes (no matter how many
# levels of access are used) will be given an ownerDocument of None.
#
def get_childNodes(self):
return xml.dom.core.NodeList(self._node.children, self, self)
xml.dom.core.Document.get_childNodes = get_childNodes
def get_first_element(doc, gi):
for n in doc.childNodes:
if n.nodeType == xml.dom.core.ELEMENT and n.tagName == gi:
return n
def extract_first_element(doc, gi):
node = get_first_element(doc, gi)
if node is not None:
doc.removeChild(node)
return node
def simplify(doc):
# Try to rationalize the document a bit, since these things are simply
# not valid SGML/XML documents as they stand, and need a little work.
documentclass = "document"
inputs = []
node = extract_first_element(doc, "documentclass")
if node is not None:
documentclass = node.getAttribute("classname")
node = extract_first_element(doc, "title")
if node is not None:
inputs.append(node)
# update the name of the root element
node = get_first_element(doc, "document")
if node is not None:
node._node.name = documentclass
while 1:
node = extract_first_element(doc, "input")
if node is None:
break
inputs.append(node)
if inputs:
docelem = doc.documentElement
inputs.reverse()
for node in inputs:
text = doc.createTextNode("\n")
docelem.insertBefore(text, docelem.firstChild)
docelem.insertBefore(node, text)
docelem.insertBefore(doc.createTextNode("\n"), docelem.firstChild)
while doc.firstChild.nodeType == xml.dom.core.TEXT:
doc.removeChild(doc.firstChild)
def cleanup_root_text(doc):
discards = []
skip = 0
for n in doc.childNodes:
prevskip = skip
skip = 0
if n.nodeType == xml.dom.core.TEXT and not prevskip:
discards.append(n)
elif n.nodeType == xml.dom.core.COMMENT:
skip = 1
for node in discards:
doc.removeChild(node)
def rewrite_desc_entries(doc, argname_gi):
argnodes = doc.getElementsByTagName(argname_gi)
for node in argnodes:
parent = node.parentNode
nodes = []
for n in parent.childNodes:
if n.nodeType != xml.dom.core.ELEMENT or n.tagName != argname_gi:
nodes.append(n)
desc = doc.createElement("description")
for n in nodes:
parent.removeChild(n)
desc.appendChild(n)
if node.childNodes:
# keep the <args>...</args>, newline & indent
parent.insertBefore(doc.createText("\n "), node)
else:
# no arguments, remove the <args/> node
parent.removeChild(node)
parent.appendChild(doc.createText("\n "))
parent.appendChild(desc)
parent.appendChild(doc.createText("\n"))
def handle_args(doc):
rewrite_desc_entries(doc, "args")
rewrite_desc_entries(doc, "constructor-args")
def handle_comments(doc, node=None):
if node is None:
node = doc
for n in node.childNodes:
if n.nodeType == xml.dom.core.ELEMENT:
if n.tagName == "COMMENT":
comment = doc.createComment(n.childNodes[0].data)
node.replaceChild(comment, n)
else:
handle_comments(doc, n)
def handle_labels(doc):
labels = doc.getElementsByTagName("label")
for label in labels:
id = label.getAttribute("id")
if not id:
continue
parent = label.parentNode
if parent.tagName == "title":
parent.parentNode.setAttribute("id", id)
else:
parent.setAttribute("id", id)
# now, remove <label id="..."/> from parent:
parent.removeChild(label)
def fixup_trailing_whitespace(doc, wsmap):
queue = [doc]
while queue:
node = queue[0]
del queue[0]
if node.nodeType == xml.dom.core.ELEMENT \
and wsmap.has_key(node.tagName):
ws = wsmap[node.tagName]
children = node.childNodes
children.reverse()
if children[0].nodeType == xml.dom.core.TEXT:
data = string.rstrip(children[0].data) + ws
children[0].data = data
children.reverse()
# hack to get the title in place:
if node.tagName == "title" \
and node.parentNode.firstChild.nodeType == xml.dom.core.ELEMENT:
node.parentNode.insertBefore(doc.createText("\n "),
node.parentNode.firstChild)
for child in node.childNodes:
if child.nodeType == xml.dom.core.ELEMENT:
queue.append(child)
def normalize(doc):
for node in doc.childNodes:
if node.nodeType == xml.dom.core.ELEMENT:
node.normalize()
def cleanup_trailing_parens(doc, element_names):
d = {}
for gi in element_names:
d[gi] = gi
rewrite_element = d.has_key
queue = []
for node in doc.childNodes:
if node.nodeType == xml.dom.core.ELEMENT:
queue.append(node)
while queue:
node = queue[0]
del queue[0]
if rewrite_element(node.tagName):
children = node.childNodes
if len(children) == 1 \
and children[0].nodeType == xml.dom.core.TEXT:
data = children[0].data
if data[-2:] == "()":
children[0].data = data[:-2]
else:
for child in node.childNodes:
if child.nodeType == xml.dom.core.ELEMENT:
queue.append(child)
def convert(ifp, ofp):
p = xml.dom.esis_builder.EsisBuilder()
p.feed(ifp.read())
doc = p.document
normalize(doc)
handle_args(doc)
handle_comments(doc)
simplify(doc)
handle_labels(doc)
fixup_trailing_whitespace(doc, {
"abstract": "\n",
"title": "",
"chapter": "\n\n",
"section": "\n\n",
"subsection": "\n\n",
"subsubsection": "\n\n",
"paragraph": "\n\n",
"subparagraph": "\n\n",
})
cleanup_root_text(doc)
cleanup_trailing_parens(doc, ["function", "method", "cfunction"])
try:
ofp.write(doc.toxml())
ofp.write("\n")
except IOError, (err, msg):
# Ignore EPIPE; it just means that whoever we're writing to stopped
# reading. The rest of the output would be ignored. All other errors
# should still be reported,
if err != errno.EPIPE:
raise
def main():
if len(sys.argv) == 1:
ifp = sys.stdin
ofp = sys.stdout
elif len(sys.argv) == 2:
ifp = open(sys.argv[1])
ofp = sys.stdout
elif len(sys.argv) == 3:
ifp = open(sys.argv[1])
ofp = open(sys.argv[2], "w")
else:
usage()
sys.exit(2)
convert(ifp, ofp)
if __name__ == "__main__":
main()
|
