Cyclone-Team/gramps
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Strip out control characters, handle // names better

Browse Source 
svn: r903
This commit is contained in:
Don Allingham 2002-04-09 13:19:29 +00:00
parent c6f9757cca
commit 133765914e
1 changed files with 15 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
gramps/src/plugins/ReadGedcom.py
View File

@ -68,6 +68,7 @@ intRE = re.compile(r"\s*(\d+)\s*$")
lineRE = re.compile(r"\s*(\d+)\s+(\S+)\s*(.*)$") lineRE = re.compile(r"\s*(\d+)\s+(\S+)\s*(.*)$")
headRE = re.compile(r"\s*(\d+)\s+HEAD") headRE = re.compile(r"\s*(\d+)\s+HEAD")
nameRegexp= re.compile(r"/?([^/]*)(/([^/]*)(/([^/]*))?)?") nameRegexp= re.compile(r"/?([^/]*)(/([^/]*)(/([^/]*))?)?")
snameRegexp= re.compile(r"/([^/]*)/")
calRegexp = re.compile(r"\s*@#D([^@]+)@\s*(.*)$") calRegexp = re.compile(r"\s*@#D([^@]+)@\s*(.*)$")
fromtoRegexp = re.compile(r"\s*FROM\s+@#D([^@]+)@\s*(.*)\s+TO\s+@#D([^@]+)@\s*(.*)$") fromtoRegexp = re.compile(r"\s*FROM\s+@#D([^@]+)@\s*(.*)\s+TO\s+@#D([^@]+)@\s*(.*)$")
@ -155,6 +156,9 @@ class GedcomParser:
self.backoff = 0 self.backoff = 0
self.cnv = nocnv self.cnv = nocnv
self.trans = string.maketrans('','')
self.delc = self.trans[0:31]
self.file_obj = window.get_widget("file") self.file_obj = window.get_widget("file")
self.encoding_obj = window.get_widget("encoding") self.encoding_obj = window.get_widget("encoding")
self.created_obj = window.get_widget("created") self.created_obj = window.get_widget("created")
@ -219,6 +223,8 @@ class GedcomParser:
def get_next(self): def get_next(self):
if self.backoff == 0: if self.backoff == 0:
self.text = self.cnv(string.strip(self.f.readline())) self.text = self.cnv(string.strip(self.f.readline()))
self.text = string.translate(self.text,self.trans,self.delc)
self.index = self.index + 1 self.index = self.index + 1
l = string.split(self.text, None, 2) l = string.split(self.text, None, 2)
ln = len(l) ln = len(l)
@ -558,6 +564,11 @@ class GedcomParser:
return return
elif matches[1] == "NAME": elif matches[1] == "NAME":
name = Name() name = Name()
m = snameRegexp.match(matches[2])
if m:
n = m.groups()
names = ('','',n,'','')
else:
try: try:
names = nameRegexp.match(matches[2]).groups() names = nameRegexp.match(matches[2]).groups()
except: except: