summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
-rwxr-xr-xversification/av11n.py38
1 files changed, 19 insertions, 19 deletions
diff --git a/versification/av11n.py b/versification/av11n.py
index 9ade84e..198eeb4 100755
--- a/versification/av11n.py
+++ b/versification/av11n.py
@@ -9,6 +9,7 @@
# special as for ordering.
#
# Invoke simply by calling the program and the file name.
+import io
import logging
# in normal state level should be debug.WARNING, debug.INFO and debug.DEBUG
# give additional information.
@@ -16,7 +17,12 @@ logging.basicConfig(format='%(levelname)s:%(funcName)s:%(message)s',
level=logging.WARNING)
import re
import sys
-verseid = re.compile(r'^.+\..+\..+$')
+try:
+ import lxml.etree as ET
+except ImportError:
+ import xml.etree.ElementTree as ET
+
+VERSEID_RE = re.compile(r'^.+\..+\..+$')
# Inform the user that we need the SWORD extension
try:
@@ -26,15 +32,6 @@ except ImportError:
"You do not have the SWORD library installed. Please install it.")
sys.exit(1)
-# Inform the user that we need pyquery, as it makes parsing XML files
-# that much easier
-try:
- from pyquery import PyQuery as pq # noqa
-except ImportError:
- logging.exception(
- "You do not appear to have PyQuery installed. Please install it.")
- sys.exit(2)
-
# Without the name of a file, we cannot proceed any further
if len(sys.argv) < 2 or sys.argv[1] == '--help':
print >>sys.stderr, "Usage: %s <OSISfile>" % sys.argv[0]
@@ -42,7 +39,8 @@ if len(sys.argv) < 2 or sys.argv[1] == '--help':
# Open the file
logging.debug('Opening %s' % (sys.argv[1],))
-d = pq(filename=sys.argv[1])
+
+tree = ET.parse(io.open(sys.argv[1], encoding='utf8')).getroot()
# Get the list of versifications
logging.debug('Fetching a list of v11ns')
vmgr = Sword.VersificationMgr.getSystemVersificationMgr()
@@ -50,7 +48,11 @@ av11ns = vmgr.getVersificationSystems()
# Get the list of all osisIDs
logging.debug('Fetching a list of OSIS IDs')
-ids = d("*[osisID]")
+ids = set()
+for item in tree.iter():
+ if 'osisID' in item.attrib:
+ ids.add(item.attrib['osisID'])
+
# Iterate each versification scheme
for v11n in av11ns:
print('Checking %s' % v11n.c_str())
@@ -82,20 +84,18 @@ for v11n in av11ns:
inNT = False
# Now iterate the ones we have in this file
- for e in ids:
- logging.debug('e = %s', e)
- osisid = e.attrib.get('osisID')
- #print 'Checking key %s' % (osisid,)
+ for osisid in ids:
+ logging.debug('Checking key %s', osisid)
if osisid in otkeyList:
otkeyList.remove(osisid)
elif osisid in ntkeyList:
ntkeyList.remove(osisid)
inNT = True
- elif verseid.match(osisid) and inNT:
+ elif VERSEID_RE.match(osisid) and inNT:
ntextraKeys.append(osisid)
- elif verseid.match(osisid) and not inNT:
+ elif VERSEID_RE.match(osisid) and not inNT:
otextraKeys.append(osisid)
- # Ignore it if not verseid.match()
+ # Ignore it if not VERSEID_RE.match()
# Now let's see what is left over
# Sets in Python cannot be ordered