coresctostandoff.py
5.99 KB
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
#!/usr/bin/env python
import sys
import re
try:
import cElementTree as ET
except:
import xml.etree.cElementTree as ET
# tags of elements to exclude from standoff output
# (not used now; anything not explicitly converted is excluded)
EXCLUDED_TAGS = [
# "SP",
# "IT",
# "SB",
# "REF",
# "P",
# "B",
# "TITLE",
# "PAPER",
# "HEADER",
# "DIV",
# "BODY",
# "ABSTRACT",
# "THEAD",
# "TGROUP",
# "TBODY",
# "SUP",
# "EQN",
# "ENTRY",
# "XREF",
# "ROW",
# "EQ-S",
# "text",
# "datasection",
# "s",
# "mode2",
]
EXCLUDED_TAG = { t:True for t in EXCLUDED_TAGS }
# string to use to indicate elided text in output
ELIDED_TEXT_STRING = "[[[...]]]"
# maximum length of text strings printed without elision
MAXIMUM_TEXT_DISPLAY_LENGTH = 1000
# c-style string escaping for just newline, tab and backslash.
# (s.encode('string_escape') does too much for utf-8)
def c_escape(s):
return s.replace('\\', '\\\\').replace('\t','\\t').replace('\n','\\n')
def strip_ns(tag):
# remove namespace spec from tag, if any
return tag if tag[0] != '{' else re.sub(r'\{.*?\}', '', tag)
class Standoff:
def __init__(self, sid, element, start, end, text):
self.sid = sid
self.element = element
self.start = start
self.end = end
self.text = text
def compress_text(self, l):
if len(self.text) >= l:
el = len(ELIDED_TEXT_STRING)
sl = (l-el)/2
self.text = (self.text[:sl]+ELIDED_TEXT_STRING+self.text[-(l-sl-el):])
def tag(self):
return strip_ns(self.element.tag)
def attrib(self):
# remove namespace specs from attribute names, if any
attrib = {}
for a in self.element.attrib:
if a[0] == "{":
an = re.sub(r'\{.*?\}', '', a)
else:
an = a
attrib[an] = self.element.attrib[a]
return attrib
def __str__(self):
return "X%d\t%s %d %d\t%s\t%s" % \
(self.sid, self.tag(), self.start, self.end,
c_escape(self.text.encode("utf-8")),
" ".join(['%s="%s"' % (k.encode("utf-8"), v.encode("utf-8"))
for k,v in self.attrib().items()]))
def txt(s):
return s if s is not None else ""
next_free_so_id = 1
def text_and_standoffs(e, curroff=0, standoffs=None):
global next_free_so_id
if standoffs == None:
standoffs = []
startoff = curroff
# to keep standoffs in element occurrence order, append
# a placeholder before recursing
so = Standoff(next_free_so_id, e, 0, 0, "")
next_free_so_id += 1
standoffs.append(so)
setext, dummy = subelem_text_and_standoffs(e, curroff+len(txt(e.text)), standoffs)
text = txt(e.text) + setext
curroff += len(text)
so.start = startoff
so.end = curroff
so.text = text
return (text, standoffs)
def subelem_text_and_standoffs(e, curroff, standoffs):
startoff = curroff
text = ""
for s in e:
stext, dummy = text_and_standoffs(s, curroff, standoffs)
text += stext
text += txt(s.tail)
curroff = startoff + len(text)
return (text, standoffs)
def empty_elements(e, tags=None):
if tags is None or strip_ns(e.tag) in tags:
e.clear()
for c in e:
empty_elements(c, tags)
def add_space(e):
if strip_ns(e.tag) in ('title', ):
e.tail = (e.tail if e.tail is not None else '') + '\n'
for c in e:
add_space(c)
def convert_coresc1(s):
sostrings = []
# create a textbound of the type specified by the "type"
# attribute.
tid = "T%d" % convert_coresc1._idseq
sostrings.append('%s\t%s %d %d\t%s' % \
(tid, s.attrib()['type'], s.start, s.end,
s.text.encode('utf-8')))
# TODO: consider converting "advantage" and "novelty" attributes
convert_coresc1._idseq += 1
return sostrings
convert_coresc1._idseq = 1
convert_function = {
'CoreSc1' : convert_coresc1,
'annotationART' : convert_coresc1,
}
def main(argv=[]):
if len(argv) != 4:
print >> sys.stderr, "Usage:", argv[0], "IN-XML OUT-TEXT OUT-SO"
return -1
in_fn, out_txt_fn, out_so_fn = argv[1:]
# "-" for STDIN / STDOUT
if in_fn == "-":
in_fn = "/dev/stdin"
if out_txt_fn == "-":
out_txt_fn = "/dev/stdout"
if out_so_fn == "-":
out_so_fn = "/dev/stdout"
tree = ET.parse(in_fn)
root = tree.getroot()
# remove unannotated, (primarily) non-content elements
empty_elements(root, set(['article-categories',
'copyright-statement', 'license',
'copyright-holder', 'copyright-year',
'journal-meta', 'article-id',
'back',
'fig', 'table-wrap',
'contrib-group',
'aff', 'author-notes',
'pub-date',
'volume', 'issue',
'fpage', 'lpage',
'history'
]))
add_space(root)
text, standoffs = text_and_standoffs(root)
# filter
standoffs = [s for s in standoffs if not s.tag() in EXCLUDED_TAG]
# convert selected elements
converted = []
for s in standoffs:
if s.tag() in convert_function:
converted.extend(convert_function[s.tag()](s))
# else:
# converted.append(s)
standoffs = converted
for so in standoffs:
try:
so.compress_text(MAXIMUM_TEXT_DISPLAY_LENGTH)
except AttributeError:
pass
# open output files
out_txt = open(out_txt_fn, "wt")
out_so = open(out_so_fn, "wt")
out_txt.write(text.encode("utf-8"))
for so in standoffs:
print >> out_so, so
out_txt.close()
out_so.close()
if __name__ == "__main__":
sys.exit(main(sys.argv))