def PrintStats(filename):
if sys.platform == 'win32':
if not pythoncom.StgIsStorageFile(filename):
print "The file is not a storage file!"
return
# Open the file.
flags = storagecon.STGM_READ | storagecon.STGM_SHARE_EXCLUSIVE
stg_ = pythoncom.StgOpenStorage(filename, None, flags)
# Now see if the storage object supports Property Information.
try:
pss = stg_.QueryInterface(pythoncom.IID_IPropertySetStorage)
except pythoncom.com_error:
print "No summary information is available"
return
# Open the user defined properties.
ps = pss.Open(FMTID_UserDefinedProperties)
props = PIDSI_TITLE, PIDSI_SUBJECT, PIDSI_AUTHOR, PIDSI_CREATE_DTM
data = ps.ReadMultiple(props)
# Unpack the result into the items.
title, subject, author, created = data
print "Title:", title
print "Subject:", subject
print "Author:", author
print "Created:", created.Format()
def structured_storage(filename):
"""Pick out info from MS documents with embedded
structured storage(typically MS Word docs etc.)
Returns a dictionary of information found
"""
if not pythoncom.StgIsStorageFile(filename):
return {}
flags = storagecon.STGM_READ | storagecon.STGM_SHARE_EXCLUSIVE
storage = pythoncom.StgOpenStorage(filename, None, flags)
try:
properties_storage = storage.QueryInterface(pythoncom.IID_IPropertySetStorage)
except pythoncom.com_error:
return {}
property_sheet = properties_storage.Open(FMTID_USER_DEFINED_PROPERTIES)
try:
data = property_sheet.ReadMultiple(PROPERTIES)
finally:
property_sheet = None
title, subject, author, created_on, keywords, comments, template_used, \
updated_by, edited_on, printed_on, saved_on, \
n_pages, n_words, n_characters, \
application = data
result = {}
if title:
result['title'] = title
if subject:
result['subject'] = subject
if author:
result['author'] = author
if created_on:
result['created_on'] = created_on
if keywords:
result['keywords'] = keywords
if comments:
result['comments'] = comments
if template_used:
result['template_used'] = template_used
if updated_by:
result['updated_by'] = updated_by
if edited_on:
result['edited_on'] = edited_on
if printed_on:
result['printed_on'] = printed_on
if saved_on:
result['saved_on'] = saved_on
if n_pages:
result['n_pages'] = n_pages
if n_words:
result['n_words'] = n_words
if n_characters:
result['n_characters'] = n_characters
if application:
result['application'] = application
return result
def parse_msg_file(filepath):
mapi.MAPIInitialize((mapi.MAPI_INIT_VERSION, 0))
storage_flags = STGM_DIRECT | STGM_READ | STGM_SHARE_EXCLUSIVE
storage = pythoncom.StgOpenStorage(filepath, None, storage_flags, None, 0)
mapi_session = mapi.OpenIMsgSession()
message = mapi.OpenIMsgOnIStg(mapi_session, None, storage, None, 0,
mapi.MAPI_UNICODE)
return message
def _bind_to_filter(self, fileName):
"""
See if the file is a structured storage file or a normal file
and then return an ifilter interface by calling the appropriate bind/load function
"""
if pythoncom.StgIsStorageFile(fileName):
self.stg = pythoncom.StgOpenStorage(fileName, None, storagecon.STGM_READ | storagecon.STGM_SHARE_DENY_WRITE)
try:
self.f = ifilter.BindIFilterFromStorage(self.stg)
except pythoncom.com_error, e:
if e[0] == -2147467262: # 0x80004002: # no interface, try the load interface (this happens for some MSoft files)
self.f = ifilter.LoadIFilter(fileName)
else:
raise
def get_ifilter_for_file(filename, log=log):
"""
Deal with structured storage file if possible.
See http://msdn2.microsoft.com/en-us/library/aa380369.aspx
"""
if pythoncom.StgIsStorageFile(filename):
storage_init_flags = STGM_READ | STGM_SHARE_DENY_WRITE
stg = pythoncom.StgOpenStorage(filename, None, storage_init_flags)
try:
filt = ifilter.BindIFilterFromStorage(stg)
except pythoncom.com_error, e:
if e[0] == -2147467262:
filt = load_ifilter(filename, log=log)
else:
raise
def readOfficeProperties(self, shellName):
if not pythoncom.StgIsStorageFile(filename):
self.abstract = 'not a Storage file'
flags = storagecon.STGM_READ | storagecon.STGM_SHARE_EXCLUSIVE
stg = pythoncom.StgOpenStorage(filename, None, flags)
try:
pss = stg.QueryInterface(pythoncom.IID_IPropertySetStorage)
except pythoncom.com_error:
self.abstract = ''
return
# open properties
ps = pss.Open(FMTID_UserDefinedProperties)
props = PIDSI_TITLE, PIDSI_SUBJECT, PIDSI_KEYWORDS
self.title, self.abstract, keywordText = ps.ReadMultiple(props)
self.keywords = string.split(keywordText, ',')
def testSimple(self):
with pytest.raises(
pythoncom.com_error):
pythoncom.StgOpenStorage("foo",
None,
0)
def _getException(self):
try:
pythoncom.StgOpenStorage("foo", None, 0)
except pythoncom.com_error as exc:
return exc
self.fail("Didn't get storage exception.")
def get_stats(path):
""" Function returns author,title,subject,keywords,comments,category using
the COM interface on windows"""
# this is code lifted from a message by [email protected]
# I changed some things
author = title = subject = keywords = comments = category = None
try:
#This is all MS stuff
pssread = pythoncom.StgOpenStorageEx(
path, storagecon.STGM_READ | storagecon.STGM_SHARE_EXCLUSIVE,
storagecon.STGFMT_FILE, 0, pythoncom.IID_IPropertySetStorage)
except:
try:
stg = pythoncom.StgOpenStorage(
path, None,
storagecon.STGM_READ | storagecon.STGM_SHARE_EXCLUSIVE)
pssread = stg.QueryInterface(pythoncom.IID_IPropertySetStorage)
except:
print("No extended storage")
else:
try:
ps = pssread.Open(
pythoncom.FMTID_SummaryInformation,
storagecon.STGM_READ | storagecon.STGM_SHARE_EXCLUSIVE)
except:
pass
else:
author,title,subject,keywords,comments = ps.ReadMultiple(\
(storagecon.PIDSI_AUTHOR, storagecon.PIDSI_TITLE,
storagecon.PIDSI_SUBJECT, storagecon.PIDSI_KEYWORDS,
storagecon.PIDSI_COMMENTS) )
try:
ps = pssread.Open(
pythoncom.FMTID_DocSummaryInformation,
storagecon.STGM_READ | storagecon.STGM_SHARE_EXCLUSIVE)
except:
pass
else:
category = ps.ReadMultiple((storagecon.PIDDSI_CATEGORY, ))[0]
stat_list = [author, title, subject, keywords, comments, category]
stat_dictionary = dict([(Field, stat_list[index])
for index, Field in enumerate(GET_STATS_FIELDS)
])
return stat_dictionary
else:
try:
ps = pssread.Open(
pythoncom.FMTID_SummaryInformation,
storagecon.STGM_READ | storagecon.STGM_SHARE_EXCLUSIVE)
except:
pass
else:
author, title, subject, keywords, comments = ps.ReadMultiple(
(storagecon.PIDSI_AUTHOR, storagecon.PIDSI_TITLE,
storagecon.PIDSI_SUBJECT, storagecon.PIDSI_KEYWORDS,
storagecon.PIDSI_COMMENTS))
try:
ps = pssread.Open(
pythoncom.FMTID_DocSummaryInformation,
storagecon.STGM_READ | storagecon.STGM_SHARE_EXCLUSIVE)
except:
pass
else:
category = ps.ReadMultiple((storagecon.PIDDSI_CATEGORY, ))[0]
try:
ps = pssread.Open(
pythoncom.FMTID_UserDefinedProperties,
storagecon.STGM_READ | storagecon.STGM_SHARE_EXCLUSIVE)
except:
pass
else:
pass
stat_list = [author, title, subject, keywords, comments, category]
stat_dictionary = dict([(Field, stat_list[index])
for index, Field in enumerate(GET_STATS_FIELDS)
])
return stat_dictionary
def _getException(self):
try:
pythoncom.StgOpenStorage("foo", None, 0)
except pythoncom.com_error, exc:
return exc
