Saskwoch patch 1

mbox-extract-attachments.py

@@ -24,6 +24,7 @@
 
 # Related RFCs: 2047, 2044, 1522
 
+# 20150907 JMW Python Version 3.4.3 and other changes
 
 __author__ = "Pablo Castellano <[email protected]>"
 __license__ = "GNU GPLv3+"
@@ -36,21 +37,35 @@
 import os
 import sys
 import email
-
+import codecs	# 20150907 JMW	For codec error handling
 
 BLACKLIST = ('signature.asc', 'message-footer.txt', 'smime.p7s')
 VERBOSE = 1
 
 attachments = 0 #Count extracted attachment
 skipped = 0
 
+# 20150907 JMW	For codec error handling
+# 		Taken from http://www.gossamer-threads.com/lists/python/python/780611#780611
+def replace_spc_error_handler(error): 
+# error is an UnicodeEncodeError/UnicodeDecodeError instance 
+# with these attributes: 
+# object = unicode object being encoded 
+# start:end = slice of object with error 
+# reason = error message 
+# Must return a tuple (replacement unicode object, 
+# index into object to continue encoding) 
+# or raise the same or another exception 
+	return (u' ' * (error.end-error.start), error.end) 
+
+
 # Search for filename or find recursively if it's multipart
 def extract_attachment(payload):
 	global attachments, skipped
 	filename = payload.get_filename()
 
 	if filename is not None:
-		print "\nAttachment found!"
+		print("\nAttachment found!")
 		if filename.find('=?') != -1:
 			ll = email.header.decode_header(filename)
 			filename = ""
@@ -60,7 +75,7 @@ def extract_attachment(payload):
 		if filename in BLACKLIST:
 			skipped = skipped + 1
 			if (VERBOSE >= 1):
-				print "Skipping %s (blacklist)\n" %filename
+				print("Skipping %s (blacklist)\n" %filename)
 			return
 
 		# Puede no venir especificado el nombre del archivo??		
@@ -70,15 +85,15 @@ def extract_attachment(payload):
 		content = payload.as_string()
 		# Skip headers, go to the content
 		fh = content.find('\n\n')
-		content = content[fh:]
+		content = content[fh:].encode('utf-8')	# 20150709 JMW	Address error I was getting
 
 		# if it's base64....
 		if payload.get('Content-Transfer-Encoding') == 'base64':
 			content = base64.decodestring(content)
 		# quoted-printable
 		# what else? ...
 
-		print "Extracting %s (%d bytes)\n" %(filename, len(content))
+		print("Extracting %s (%d bytes)\n" %(filename, len(content)))
 
 		n = 1
 		orig_filename = filename
@@ -87,74 +102,81 @@ def extract_attachment(payload):
 			n = n+1
 
 		try:
-			fp = open(filename, "w")
+			fp = open(filename, "wb")	# 201500907 JMW	Needed to make binary
 #			fp = open(str(i) + "_" + filename, "w")
 			fp.write(content)
 		except IOError:
-			print "Aborted, IOError!!!"
+			print("Aborted, IOError!!!")
 			sys.exit(2)
 		finally:
 			fp.close()	
 
 		attachments = attachments + 1
 	else:
 		if payload.is_multipart():
-			for payl in payload.get_payload():
+			for payl in payload.get_payload(decode=False):	# 20150907 JMW Needed to make decode=False
 				extract_attachment(payl)
 
 
 ###
-print "Extract attachments from mbox files"
-print "Copyright (C) 2012 Pablo Castellano"
-print "This program comes with ABSOLUTELY NO WARRANTY."
-print "This is free software, and you are welcome to redistribute it under certain conditions."
-print
+print("Extract attachments from mbox files")
+print("Copyright (C) 2012 Pablo Castellano")
+print("This program comes with ABSOLUTELY NO WARRANTY.")
+print("This is free software, and you are welcome to redistribute it under certain conditions.")
+print()
+
+codecs.register_error("replace_spc", replace_spc_error_handler)	# 20150907 JMW	Register error handler
 
 if len(sys.argv) < 2 or len(sys.argv) > 3:
-	print "Usage: %s <mbox_file> [directory]" %sys.argv[0]
+	print("Usage: %s <mbox_file> [directory]" %sys.argv[0])
 	sys.exit(0)
 
 filename = sys.argv[1]
 directory = os.path.curdir
 
 if not os.path.exists(filename):
-	print "File doesn't exist:", filename
+	print("File doesn't exist:", filename)
 	sys.exit(1)
 
 if len(sys.argv) == 3:
 	directory = sys.argv[2]
 	if not os.path.exists(directory) or not os.path.isdir(directory):
-		print "Directory doesn't exist:", directory
+		print("Directory doesn't exist:", directory)
 		sys.exit(1)
 
 mb = mailbox.mbox(filename)
-nmes = len(mb)
+# nmes = len(mb)	# 20150907 JMW Commented out as not used and a performance hit
 
 os.chdir(directory)
 
-for i in range(len(mb)):
+for i in range(len(mb)):	# 20150907 JMW With >140k msgs I used a numeric literal here for performance reasons
 	if (VERBOSE >= 2):
-		print "Analyzing message number", i
+		print("Analyzing message number", i)
 
 	mes = mb.get_message(i)
 	em = email.message_from_string(mes.as_string())
 
 	subject = em.get('Subject')
-	if subject.find('=?') != -1:
-		ll = email.header.decode_header(subject)
-		subject = ""
-		for l in ll:
-			subject = subject + l[0]
+	# 20150907 JMW Ended up adding exception handling to skip two messages that were throwing NonType
+	try:
+		if subject.find('=?') != -1:
+			ll = email.header.decode_header(subject)
+			subject = ""
+			for l in ll:
+				subject = subject + l[0].decode('utf-8', "replace_spc")	# 20150907 JMW Explicit decode
+	except AttributeError:
+		print("NonType encountered")
+		continue
 
 	em_from = em.get('From')
 	if em_from.find('=?') != -1:
 		ll = email.header.decode_header(em_from)
 		em_from = ""
 		for l in ll:
-			em_from = em_from + l[0]
+			em_from = em_from + l[0].decode('utf-8', "replace_spc")	# 20150907 JMW Explicit decode
 
 	if (VERBOSE >= 2):
-		print "%s - From: %s" %(subject, em_from)
+		print("%s - From: %s" %(subject, em_from))
 
 	filename = mes.get_filename()
 
@@ -165,6 +187,6 @@ def extract_attachment(payload):
 	else:
 		extract_attachment(em)
 
-print "\n--------------"
-print "Total attachments extracted:", attachments
-print "Total attachments skipped:", skipped
+print("\n--------------")
+print("Total attachments extracted:", attachments)
+print("Total attachments skipped:", skipped)