def useRoot(self, root):
stream = root.getSubIStream()
ministream = guessParser(stream)
if not ministream:
warning("Unable to create the OLE2 mini stream parser!")
return
self._extract(ministream, main_document=False)
def getFieldAttribute(fieldset, key, attrname):
try:
field = fieldset[key]
if field.hasValue():
return getattr(field, attrname)
except HACHOIR_ERRORS, err:
warning("Unable to get %s of field %s/%s: %s" % (
attrname, fieldset.path, key, err))
def getFieldAttribute(fieldset, key, attrname):
try:
field = fieldset[key]
if field.hasValue():
return getattr(field, attrname)
except HACHOIR_ERRORS, err:
warning("Unable to get %s of field %s/%s: %s" %
(attrname, fieldset.path, key, err))
def getField(self, fieldset, main_document, name):
if name not in fieldset:
return None
# _feedAll() is needed to make sure that we get all fragments
# eg. summary[0], summary[1], ..., summary[n]
fieldset._feedAll()
field = fieldset[name]
if main_document:
stream = field.getSubIStream()
field = guessParser(stream)
if not field:
warning("Unable to create the OLE2 parser for %s!" % name)
return None
return field
def useProperties(self, properties):
last = properties["total_str_length"]
offset0 = last.address + last.size
for index in properties.array("property"):
# Search name and value
value = properties.getFieldByAddress(offset0+index["value_offset"].value*8)
if not value:
continue
value = value.value
if not value:
continue
name = properties.getFieldByAddress(offset0+index["name_offset"].value*8)
if not name:
continue
name = name.value
if name not in self.PROP_TO_KEY:
warning("Skip %s=%r" % (name, value))
continue
key = self.PROP_TO_KEY[name]
setattr(self, key, value)
for key, value in self.parser_args.iteritems():
setattr(parser_obj, key, value)
return parser_obj
except ValidateError, err:
res = unicode(err)
if fallback and self.fallback:
fb = parser
except HACHOIR_ERRORS, err:
res = unicode(err)
if warn:
if parser == self.other:
warn = info
warn(_("Skip parser '%s': %s") % (parser.__name__, res))
fallback = False
if self.use_fallback and fb:
warning(_("Force use of parser '%s'") % fb.__name__)
return fb(stream)
def guessParser(stream):
return QueryParser(stream.tags).parse(stream)
def createParser(filename, real_filename=None, tags=None):
"""
Create a parser from a file or returns None on error.
Options:
- filename (unicode): Input file name ;
- real_filename (str|unicode): Real file name.
"""
for key, value in self.parser_args.iteritems():
setattr(parser_obj, key, value)
return parser_obj
except ValidateError, err:
res = unicode(err)
if fallback and self.fallback:
fb = parser
except HACHOIR_ERRORS, err:
res = unicode(err)
if warn:
if parser == self.other:
warn = info
warn(_("Skip parser '%s': %s") % (parser.__name__, res))
fallback = False
if self.use_fallback and fb:
warning(_("Force use of parser '%s'") % fb.__name__)
return fb(stream)
def guessParser(stream):
return QueryParser(stream.tags).parse(stream)
def createParser(filename, real_filename=None, tags=None):
"""
Create a parser from a file or returns None on error.
Options:
- filename (unicode): Input file name ;
- real_filename (str|unicode): Real file name.
"""
def safe_func(*args, **kw):
try:
func(*args, **kw)
except HACHOIR_ERRORS, err:
warning("Error when calling function %s(): %s" % (
func.__name__, err))
def safe_func(*args, **kw):
try:
func(*args, **kw)
except HACHOIR_ERRORS, err:
warning("Error when calling function %s(): %s" %
(func.__name__, err))