[sword-svn] r419 - trunk/modules/python
chrislit at crosswire.org
chrislit at crosswire.org
Sun Jun 9 14:53:34 MST 2013
Author: chrislit
Date: 2013-06-09 14:53:34 -0700 (Sun, 09 Jun 2013)
New Revision: 419
Modified:
trunk/modules/python/usfm2osis.py
Log:
simplification/cleanup of lookaheads for XML tags
Modified: trunk/modules/python/usfm2osis.py
===================================================================
--- trunk/modules/python/usfm2osis.py 2013-06-08 15:20:50 UTC (rev 418)
+++ trunk/modules/python/usfm2osis.py 2013-06-09 21:53:34 UTC (rev 419)
@@ -489,7 +489,7 @@
osis = re.sub('(\uFDE6<div type="subSubSubSubSection" subType="x-introduction">[^\uFDE2\uFDE3\uFDE4\uFDE5\uFDE6]+?)(?!\\c\b)', r'\1'+'</div>\uFDE6\n', osis, flags=re.DOTALL)
# \ip_text...
- osis = re.sub(r'\\ip\s+(.*?)(?=(\\(i?m|i?p|lit|cls|tr|io|iq|i?li|iex?|s|c)\b|</?div\b|<(p|closer)\b))', lambda m: '\uFDD3<p subType="x-introduction">\n' + m.group(1) + '\uFDD3</p>\n', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\ip\s+(.*?)(?=(\\(i?m|i?p|lit|cls|tr|io|iq|i?li|iex?|s|c)\b|<(/?div|p|closer)\b))', lambda m: '\uFDD3<p subType="x-introduction">\n' + m.group(1) + '\uFDD3</p>\n', osis, flags=re.DOTALL)
# \ipi_text...
# \im_text...
@@ -498,7 +498,7 @@
# \imq_text...
# \ipr_text...
pType = {'ipi':'x-indented', 'im':'x-noindent', 'imi':'x-noindent-indented', 'ipq':'x-quote', 'imq':'x-noindent-quote', 'ipr':'x-right'}
- osis = re.sub(r'\\(ipi|im|ipq|imq|ipr)\s+(.*?)(?=(\\(i?m|i?p|lit|cls|tr|io|iq|i?li|iex?|s|c)\b|</?div\b|<(p|closer)\b))', lambda m: '\uFDD3<p type="' + pType[m.group(1)] + '" subType="x-introduction">\n' + m.group(2) + '\uFDD3</p>\n', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\(ipi|im|ipq|imq|ipr)\s+(.*?)(?=(\\(i?m|i?p|lit|cls|tr|io|iq|i?li|iex?|s|c)\b|<(/?div|p|closer)\b))', lambda m: '\uFDD3<p type="' + pType[m.group(1)] + '" subType="x-introduction">\n' + m.group(2) + '\uFDD3</p>\n', osis, flags=re.DOTALL)
# \iq#_text...
osis = re.sub(r'\\iq\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\(iq[\d\s]|fig|q\d?|b)\b|<title\b))', r'<l level="1" subType="x-introduction">\1</l>', osis, flags=re.DOTALL)
@@ -511,16 +511,16 @@
#osis = re.sub('(<lg>.+?</lg>)', lambda m: m.group(1).replace('<lb type="x-p"/>', '</lg><lg>'), osis, flags=re.DOTALL) # re-handle \b that occurs within <lg>
# \ili#_text...
- osis = re.sub(r'\\ili\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\ili[\d\s]|<lb\b|<title\b|<item\b))', '<item type="x-indent-1" subType="x-introduction">\uFDE0'+r'\1'+'\uFDE0</item>', osis, flags=re.DOTALL)
- osis = re.sub(r'\\ili(\d)\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\ili[\d\s]|<lb\b|<title\b|<item\b))', r'<item type="x-indent-\1" subType="x-introduction">\uFDE0'+r'\2'+'\uFDE0</item>', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\ili\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\ili[\d\s]|<(lb|title|item)\b))', '<item type="x-indent-1" subType="x-introduction">\uFDE0'+r'\1'+'\uFDE0</item>', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\ili(\d)\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\ili[\d\s]|<(lb|title|item)\b))', r'<item type="x-indent-\1" subType="x-introduction">\uFDE0'+r'\2'+'\uFDE0</item>', osis, flags=re.DOTALL)
osis = osis.replace('\n</item>', '</item>\n')
osis = re.sub('(<item [^\uFDD0\uFDD1\uFDD3\uFDD4]+</item>)', '\uFDD3<list>'+r'\1'+'</list>\uFDD3', osis, flags=re.DOTALL)
# \iot_text...
# \io#_text...(references range)
- osis = re.sub(r'\\io\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\io[t\d\s]|\\iex?|<lb\b|<title\b|<item\b))', '<item type="x-indent-1" subType="x-introduction">\uFDE1'+r'\1'+'\uFDE1</item>', osis, flags=re.DOTALL)
- osis = re.sub(r'\\io(\d)\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\io[t\d\s]|\\iex?|<lb\b|<title\b|<item\b))', r'<item type="x-indent-\1" subType="x-introduction">\uFDE1'+r'\2'+'\uFDE1</item>', osis, flags=re.DOTALL)
- osis = re.sub(r'\\iot\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\io[t\d\s]|\\iex?|<lb\b|<title\b|<item\b))', '<item type="head">\uFDE1'+r'\1'+'\uFDE1</item type="head">', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\io\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\io[t\d\s]|\\iex?|<(lb|title|item)\b))', '<item type="x-indent-1" subType="x-introduction">\uFDE1'+r'\1'+'\uFDE1</item>', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\io(\d)\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\io[t\d\s]|\\iex?|<(lb|title|item)\b))', r'<item type="x-indent-\1" subType="x-introduction">\uFDE1'+r'\2'+'\uFDE1</item>', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\iot\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\io[t\d\s]|\\iex?|<(lb|title|item)\b))', '<item type="head">\uFDE1'+r'\1'+'\uFDE1</item type="head">', osis, flags=re.DOTALL)
osis = osis.replace('\n</item>', '</item>\n')
osis = re.sub('(<item [^\uFDD0\uFDD1\uFDD3\uFDD4\uFDE0]+</item>)', '\uFDD3<div type="outline"><list>'+r'\1'+'</list></div>\uFDD3', osis, flags=re.DOTALL)
osis = re.sub('item type="head"', 'head', osis)
@@ -689,7 +689,7 @@
paragraphregex += '|phi|ps|psi|p1|p2|p3|p4|p5'
# \p(_text...)
- osis = re.sub(r'\\p\s+(.*?)(?=(\\(i?m|i?p|lit|cls|tr|p|'+paragraphregex+r')\b|<chapter eID|</?div\b|<(p|closer)\b))', lambda m: '\uFDD3<p>\n' + m.group(1) + '\uFDD3</p>\n', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\p\s+(.*?)(?=(\\(i?m|i?p|lit|cls|tr|p|'+paragraphregex+r')\b|<chapter eID|<(/?div|p|closer)\b))', lambda m: '\uFDD3<p>\n' + m.group(1) + '\uFDD3</p>\n', osis, flags=re.DOTALL)
# \pc(_text...)
# \pr(_text...)
@@ -706,17 +706,17 @@
# \psi # deprecated
# \p# # deprecated
pType = {'pc':'x-center', 'pr':'x-right', 'm':'x-noindent', 'pmo':'x-embedded-opening', 'pm':'x-embedded', 'pmc':'x-embedded-closing', 'pmr':'x-right', 'pi':'x-indented-1', 'pi1':'x-indented-1', 'pi2':'x-indented-2', 'pi3':'x-indented-3', 'pi4':'x-indented-4', 'pi5':'x-indented-5', 'mi':'x-noindent-indented', 'nb':'x-nobreak', 'phi':'x-indented-hanging', 'ps':'x-nobreakNext', 'psi':'x-nobreakNext-indented', 'p1':'x-level-1', 'p2':'x-level-2', 'p3':'x-level-3', 'p4':'x-level-4', 'p5':'x-level-5'}
- osis = re.sub(r'\\('+paragraphregex+r')\s+(.*?)(?=(\\(i?m|i?p|lit|cls|tr|'+paragraphregex+r')\b|<chapter eID|</?div\b|<(p|closer)\b))', lambda m: '\uFDD3<p type="' + pType[m.group(1)] + '">\n' + m.group(2) + '\uFDD3</p>\n', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\('+paragraphregex+r')\s+(.*?)(?=(\\(i?m|i?p|lit|cls|tr|'+paragraphregex+r')\b|<chapter eID|<(/?div|p|closer)\b))', lambda m: '\uFDD3<p type="' + pType[m.group(1)] + '">\n' + m.group(2) + '\uFDD3</p>\n', osis, flags=re.DOTALL)
# \cls_text...
- osis = re.sub(r'\\m\s+(.+?)(?=(\\(i?m|i?p|lit|cls|tr)\b|<chapter eID|</?div\b|<(p|closer)\b))', lambda m: '\uFDD3<closer>' + m.group(1) + '\uFDD3</closer>\n', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\m\s+(.+?)(?=(\\(i?m|i?p|lit|cls|tr)\b|<chapter eID|<(/?div|p|closer)\b))', lambda m: '\uFDD3<closer>' + m.group(1) + '\uFDD3</closer>\n', osis, flags=re.DOTALL)
# \ph#(_text...)
# \li#(_text...)
osis = re.sub(r'\\ph\b\s*', r'\\li ', osis)
osis = re.sub(r'\\ph(\d)\b\s*', r'\\li\1 ', osis)
- osis = re.sub(r'\\li\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4\uFDE0\uFDE1\uFDD5\uFDD6\uFDD7\uFDD8\uFDD9\uFDDA\uFDDB\uFDDC\uFDDD\uFDDE'+r']|\\li[\d\s]|<lb\b|<title\b|<item\b|</?div\b|</?chapter\b))', r'<item type="x-indent-1">\1</item>', osis, flags=re.DOTALL)
- osis = re.sub(r'\\li(\d)\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4\uFDE0\uFDE1\uFDD5\uFDD6\uFDD7\uFDD8\uFDD9\uFDDA\uFDDB\uFDDC\uFDDD\uFDDE'+r']|\\li[\d\s]|<lb\b|<title\b|<item\b|</?div\b|</?chapter\b))', r'<item type="x-indent-\1">\2</item>', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\li\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4\uFDE0\uFDE1\uFDD5\uFDD6\uFDD7\uFDD8\uFDD9\uFDDA\uFDDB\uFDDC\uFDDD\uFDDE'+r']|\\li[\d\s]|<(lb|title|item|/?div|/?chapter)\b))', r'<item type="x-indent-1">\1</item>', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\li(\d)\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4\uFDE0\uFDE1\uFDD5\uFDD6\uFDD7\uFDD8\uFDD9\uFDDA\uFDDB\uFDDC\uFDDD\uFDDE'+r']|\\li[\d\s]|<(lb|title|item|/?div|/?chapter)\b))', r'<item type="x-indent-\1">\2</item>', osis, flags=re.DOTALL)
osis = osis.replace('\n</item>', '</item>\n')
osis = re.sub('(<item [^\uFDD0\uFDD1\uFDD3\uFDD4\uFDE0\uFDE1\uFDD5\uFDD6\uFDD7\uFDD8\uFDD9\uFDDA\uFDDB\uFDDC\uFDDD\uFDDE]+</item>)', '\uFDD3<list>'+r'\1'+'</list>\uFDD3', osis, flags=re.DOTALL)
@@ -747,14 +747,14 @@
osis = re.sub(r'\\qs\b\s(.+?)\\qs\*', r'<l type="selah">\1</l>', osis, flags=re.DOTALL)
# \q#(_text...)
- osis = re.sub(r'\\q\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4\uFDD5\uFDD6\uFDD7\uFDD8\uFDD9\uFDDA\uFDDB\uFDDC\uFDDD\uFDDE'+r']|\\q[\d\s]|\\fig|<l\b|<lb\b|<title\b|<list\b|</?div\b))', r'<l level="1">\1</l>', osis, flags=re.DOTALL)
- osis = re.sub(r'\\q(\d)\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4\uFDD5\uFDD6\uFDD7\uFDD8\uFDD9\uFDDA\uFDDB\uFDDC\uFDDD\uFDDE'+r']|\\q[\d\s]|\\fig|<l\b|<lb\b|<title\b|<list\b|</?div\b))', r'<l level="\1">\2</l>', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\q\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4\uFDD5\uFDD6\uFDD7\uFDD8\uFDD9\uFDDA\uFDDB\uFDDC\uFDDD\uFDDE'+r']|\\q[\d\s]|\\fig|<(l|lb|title|list|/?div)\b))', r'<l level="1">\1</l>', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\q(\d)\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4\uFDD5\uFDD6\uFDD7\uFDD8\uFDD9\uFDDA\uFDDB\uFDDC\uFDDD\uFDDE'+r']|\\q[\d\s]|\\fig|<(l|lb|title|list|/?div)\b))', r'<l level="\1">\2</l>', osis, flags=re.DOTALL)
# \qr_text...
# \qc_text...
# \qm#(_text...)
qType = {'qr':'x-right', 'qc':'x-center', 'qm':'x-embedded" level="1', 'qm1':'x-embedded" level="1', 'qm2':'x-embedded" level="2', 'qm3':'x-embedded" level="3', 'qm4':'x-embedded" level="4', 'qm5':'x-embedded" level="5'}
- osis = re.sub(r'\\(qr|qc|qm\d)\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4\uFDD5\uFDD6\uFDD7\uFDD8\uFDD9\uFDDA\uFDDB\uFDDC\uFDDD\uFDDE'+r']|\\q[\d\s]|\\fig|<l\b|<lb\b|<title\b|<list\b|</?div\b))', lambda m: '<l type="' + qType[m.group(1)] + '">' + m.group(2) + '</l>', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\(qr|qc|qm\d)\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4\uFDD5\uFDD6\uFDD7\uFDD8\uFDD9\uFDDA\uFDDB\uFDDC\uFDDD\uFDDE'+r']|\\q[\d\s]|\\fig|<(l|lb|title|list|/?div)\b))', lambda m: '<l type="' + qType[m.group(1)] + '">' + m.group(2) + '</l>', osis, flags=re.DOTALL)
osis = osis.replace('\n</l>', '</l>\n')
osis = re.sub('(<l [^\uFDD0\uFDD1\uFDD3\uFDD4\uFDD5\uFDD6\uFDD7\uFDD8\uFDD9\uFDDA\uFDDB\uFDDC\uFDDD\uFDDE]+</l>)', r'<lg>\1</lg>', osis, flags=re.DOTALL)
@@ -777,7 +777,7 @@
"""
# \tr_
- osis = re.sub(r'\\tr\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\tr\s|<lb\b|<title\b))', r'<row>\1</row>', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\tr\b\s*(.*?)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\tr\s|<(lb|title)\b))', r'<row>\1</row>', osis, flags=re.DOTALL)
# \th#_text...
# \thr#_text...
@@ -786,7 +786,7 @@
tType = {'th':' role="label"', 'thr':' role="label" type="x-right"', 'tc':'', 'tcr':' type="x-right'}
osis = re.sub(r'\\(thr?|tcr?)\d*\b\s*(.*?)(?=(\\t[hc]|</row))', lambda m: '<cell' + tType[m.group(1)] + '>' + m.group(2) + '</cell>', osis, flags=re.DOTALL)
- osis = re.sub(r'(<row>.*?</row>)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\tr\s|<lb\b|<title\b))', r'<table>\1</table>', osis, flags=re.DOTALL)
+ osis = re.sub(r'(<row>.*?</row>)(?=(['+'\uFDD0\uFDD1\uFDD3\uFDD4'+r']|\\tr\s|<(lb|title)\b))', r'<table>\1</table>', osis, flags=re.DOTALL)
return osis
@@ -968,7 +968,7 @@
osis = re.sub(r'\\k\s+(.+?)\\k\*', r'<seg type="keyword">\1</seg>', osis, flags=re.DOTALL)
# \lit
- osis = re.sub(r'\\lit\s+(.*?)(?=(\\(i?m|i?p|nb|lit|cls|tr)\b|<chapter eID|</?div\b|<(p|closer)\b))', lambda m: '\uFDD3<p type="x-liturgical">\n' + m.group(1) + '\uFDD3</p>\n', osis, flags=re.DOTALL)
+ osis = re.sub(r'\\lit\s+(.*?)(?=(\\(i?m|i?p|nb|lit|cls|tr)\b|<(chapter eID|/?div|p|closer)\b))', lambda m: '\uFDD3<p type="x-liturgical">\n' + m.group(1) + '\uFDD3</p>\n', osis, flags=re.DOTALL)
# \dc_...\dc* # TODO: Find an example---should this really be transChange?
osis = re.sub(r'\\dc\b\s*(.+?)\\dc\*', r'<transChange type="added" editions="dc">\1</transChange>', osis, flags=re.DOTALL)
More information about the sword-cvs
mailing list