mc/vfs/extfs/s3.in

#!/usr/bin/env python
# -*- coding: utf-8 -*-

#
#  Midnight Commander compatible EXTFS for accessing Amazon Web Services S3.
#  Written by Jakob Kemi <jakob.kemi@gmail.com> 2009
#
#  Copyright (c) 2009 Free Software Foundation, Inc.
#  This program is free software; you can redistribute it and/or modify
#  it under the terms of the GNU General Public License as published by
#  the Free Software Foundation; either version 2 of the License, or
#  (at your option) any later version.
#
#  This program is distributed in the hope that it will be useful,
#  but WITHOUT ANY WARRANTY; without even the implied warranty of
#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#  GNU General Public License for more details.
#
#  You should have received a copy of the GNU General Public License
#  along with this program; if not, write to the Free Software
#  Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
#
#
# Notes:
#  This EXTFS exposes buckets as directories and keys as files
#  Due to EXTFS limitations all buckets & keys have to be read initially which might
#  take quite some time.
#  Tested on Debian with Python 2.4-2.6 and boto 1.4c and 1.6b
#    (Python 2.6 might need -W ignore::DeprecationWarning due to boto using
#    deprecated module Popen2)
#
#
# Installation:
#  Make sure that boto <http://code.google.com/p/boto> (python-boto in Debian) is installed.
#  Preferably pytz (package python-tz in Debian) should be installed as well.
#
#  Save as executable file /share/mc/extfs/s3 (or wherever your mc expects to find extfs modules)
#  Add the the following to your extfs.ini (might exists as /usr/share/mc/extfs/extfs.ini):
# ----- begin extfs.ini -----
#  # Amazon S3
#  s3:
# ----- end extfs.ini   -----
#
#
# Settings: (should be set via environment)
#  Required:
#    AWS_ACCESS_KEY_ID         : Amazon AWS acces key (required)
#    AWS_SECRET_ACCESS_KEY     : Amazon AWS secret access key (required)
#  Optional:
#    MCVFS_EXTFS_S3_LOCATION   : where to create new buckets, "EU"(default) or "US"
#    MCVFS_EXTFS_S3_DEBUGFILE  : write debug info to this file (no info default)
#
#
# Usage:
#  Open dialog "Quick cd" (<alt-c>) and type: #s3 <enter> (or simply type ''cd #s3'' in shell line)
#
#
# History:
#  2009-02-07 Jakob Kemi <jakob.kemi@gmail.com>
#   - Updated instructions.
#   - Improved error reporting.
#
#  2009-02-06 Jakob Kemi <jakob.kemi@gmail.com>
#   - Threaded list command.
#   - Handle rm of empty "subdirectories" (as seen in mc).
#   - List most recent datetime and total size of keys as directory properties.
#   - List modification time in local time.
#
#  2009-02-05 Jakob Kemi <jakob.kemi@gmail.com>
#   - Initial version.
#

import sys
import os
import time
import re
import datetime


import boto
from boto.s3.connection import S3Connection
from boto.s3.key import Key
from boto.exception import BotoServerError


# Get settings from environment
USER=os.getenv('USER','0')
AWS_ACCESS_KEY_ID=os.getenv('AWS_ACCESS_KEY_ID')
AWS_SECRET_ACCESS_KEY=os.getenv('AWS_SECRET_ACCESS_KEY')
LOCATION = os.getenv('MCVFS_EXTFS_S3_LOCATION', 'EU').lower()
DEBUGFILE = os.getenv('MCVFS_EXTFS_S3_DEBUGFILE')

if not AWS_ACCESS_KEY_ID or not AWS_SECRET_ACCESS_KEY:
	sys.stderr.write('Missing AWS_ACCESS_KEY_ID or AWS_SECRET_ACCESS_KEY environment variables.\n')
	sys.exit(1)

# Setup logging
if DEBUGFILE:
	import logging
	logging.basicConfig(
		filename=DEBUGFILE,
		level=logging.DEBUG,
		format='%(asctime)s %(levelname)s %(message)s')
	logging.getLogger('boto').setLevel(logging.WARNING)
else:
	class Void(object):
		def __getattr__(self, attr):
			return self
		def __call__(self, *args, **kw):
			return self
	logging = Void()

logger=logging.getLogger('s3extfs')


def threadmap(fun, iterable, maxthreads=16):
	"""
	Quick and dirty threaded version of builtin method map.
	Propagates exception safely.
	"""
	from threading import Thread
	import Queue

	items = list(iterable)
	nitems = len(items)
	if nitems < 2:
		return map(fun, items)

	# Create and fill input queue
	input = Queue.Queue()
	output = Queue.Queue()

	for i,item in enumerate(items):
		input.put( (i,item) )

	class WorkThread(Thread):
		"""
		Takes one item from input queue (thread terminates when input queue is empty),
		performs fun, puts result in output queue
		"""
		def run(self):
			while True:
				try:
					(i,item) = input.get_nowait()
					try:
						result = fun(item)
						output.put( (i,result) )
					except:
						output.put( (None,sys.exc_info()) )
				except Queue.Empty:
					return

	# Start threads
	for i in range( min(len(items), maxthreads) ):
		t = WorkThread()
		t.setDaemon(True)
		t.start()

	# Wait for all threads to finish & collate results
	ret = []
	for i in range(nitems):
		try:
			i,res = output.get()
			if i == None:
				raise res[0],res[1],res[2]
		except Queue.Empty:
			break
		ret.append(res)

	return ret

logger.debug('started')

# Global S3 connection
s3 = S3Connection(AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY)
if LOCATION == 'eu':
	logger.debug('Using location EU for new buckets')
	S3LOCATION = boto.s3.connection.Location.EU
else:
	logger.debug('Using location US for new buckets')
	S3LOCATION = boto.s3.connection.Location.US

logger.debug('argv: ' + str(sys.argv))

try:
	cmd = sys.argv[1]
	args = sys.argv[2:]
except:
	sys.stderr.write('This program should be called from within MC\n')
	sys.exit(1)

def handleServerError(msg):
	e = sys.exc_info()
	msg += ', reason: ' + e[1].reason
	logger.error(msg, exc_info=e)
	sys.stderr.write(msg+'\n')
	sys.exit(1)

#
# Lists all S3 contents
#
if cmd == 'list':
	if len(args) > 0:
		path = args[0]
	else:
		path = ''

	logger.info('list')

	rs = s3.get_all_buckets()

	# Import python timezones (pytz)
	try:
		import pytz
	except:
		logger.warning('Missing pytz module, timestamps will be off')
		# A fallback UTC tz stub
		class pytzutc(datetime.tzinfo):
			def __init__(self):
				datetime.tzinfo.__init__(self)
				self.utc = self
				self.zone = 'UTC'
			def utcoffset(self, dt):
				return datetime.timedelta(0)
			def tzname(self, dt):
				return "UTC"
			def dst(self, dt):
				return datetime.timedelta(0)
		pytz = pytzutc()


	# Find timezone
	# (yes, timeZONE as in _geographic zone_ not EST/CEST or whatever crap we get from time.tzname)
	# http://regebro.wordpress.com/2008/05/10/python-and-time-zones-part-2-the-beast-returns/
	def getGuessedTimezone():
		# 1. check TZ env. var
		try:
			tz = os.getenv('TZ', '')
			return pytz.timezone(tz)
		except:
			pass
		# 2. check if /etc/timezone exists (Debian at least)
		try:
			if os.path.isfile('/etc/timezone'):
				tz = open('/etc/timezone', 'r').readline().strip()
				return pytz.timezone(tz)
		except:
			pass
		# 3. check if /etc/localtime is a _link_ to something useful
		try:
			if os.path.islink('/etc/localtime'):
				link = os.readlink('/etc/localtime')
				tz = '/'.join(p.split(os.path.sep)[-2:])
				return pytz.timezone(tz)
		except:
			pass
		# 4. use time.tzname which will probably be wrong by an hour 50% of the time.
		try:
			return pytz.timezone(time.tzname[0])
		except:
			pass
		# 5. use plain UTC ...
		return pytz.utc

	tz=getGuessedTimezone()
	logger.debug('Using timezone: ' + tz.zone)

	# AWS time is on format: 2009-01-07T16:43:39.000Z
	# we "want" MM-DD-YYYY hh:mm (in localtime)
	expr = re.compile(r'^(\d{4})-(\d{2})-(\d{2})T(\d{2}):(\d{2}):(\d{2})\.\d{3}Z$')
	def convDate(awsdatetime):
		m = expr.match(awsdatetime)
		ye,mo,da,ho,mi,se = map(int,m.groups())

		dt = datetime.datetime(ye,mo,da,ho,mi,se, tzinfo=pytz.utc)
		return dt.astimezone(tz).strftime('%m-%d-%Y %H:%M')


	def bucketList(b):
		totsz = 0
		mostrecent = '1970-01-01T00:00:00.000Z'
		ret = []
		for k in b.list():
			mostrecent = max(mostrecent, k.last_modified)
			datetime = convDate(k.last_modified)
			ret.append('%10s %3d %-8s %-8s %d %s %s\n' % (
				'-rw-r--r--', 1, USER, USER, k.size, datetime, b.name+'/'+k.name)
			)
			totsz += k.size

		datetime=convDate(mostrecent)
		sys.stdout.write('%10s %3d %-8s %-8s %d %s %s\n' % (
			'drwxr-xr-x', 1, USER, USER, totsz, datetime, b.name)
		)
		for line in ret:
			sys.stdout.write(line)

	threadmap(bucketList, rs)

#
# Fetch file from S3
#
elif cmd == 'copyout':
	archivename = args[0]
	storedfilename = args[1]
	extractto = args[2]

	bucket,key = storedfilename.split('/', 1)
	logger.info('copyout bucket: %s, key: %s'%(bucket, key))

	try:
		b = s3.get_bucket(bucket)
		k = b.get_key(key)

		out = open(extractto, 'w')

		k.open(mode='r')
		for buf in k:
			out.write(buf)
		k.close()
		out.close()
	except BotoServerError:
		handleServerError('Unable to fetch key "%s"'%(key))

#
# Upload file to S3
#
elif cmd == 'copyin':
	archivename = args[0]
	storedfilename = args[1]
	sourcefile = args[2]

	bucket,key = storedfilename.split('/', 1)
	logger.info('copyin bucket: %s, key: %s'%(bucket, key))

	try:
		b = s3.get_bucket(bucket)
		k = b.new_key(key)
		k.set_contents_from_file(fp=open(sourcefile,'r'))
	except BotoServerError:
		handleServerError('Unable to upload key "%s"' % (key))

#
# Remove file from S3
#
elif cmd == 'rm':
	archivename = args[0]
	storedfilename = args[1]

	bucket,key = storedfilename.split('/', 1)
	logger.info('rm bucket: %s, key: %s'%(bucket, key))

	try:
		b = s3.get_bucket(bucket)
		b.delete_key(key)
	except BotoServerError:
		handleServerError('Unable to remove key "%s"' % (key))

#
# Create directory
#
elif cmd == 'mkdir':
	archivename = args[0]
	dirname = args[1]

	logger.info('mkdir dir: %s' %(dirname))
	if '/' in dirname:
		logger.warning('skipping mkdir')
		pass
	else:
		bucket = dirname
		try:
			s3.create_bucket(bucket, location=boto.s3.connection.Location.EU)
		except BotoServerError:
			handleServerError('Unable to create bucket "%s"' % (bucket))

#
# Remove directory
#
elif cmd == 'rmdir':
	archivename = args[0]
	dirname = args[1]

	logger.info('rmdir dir: %s' %(dirname))
	if '/' in dirname:
		logger.warning('skipping rmdir')
		pass
	else:
		bucket = dirname
		try:
			b = s3.get_bucket(bucket)
			s3.delete_bucket(b)
		except BotoServerError:
			handleServerError('Unable to delete bucket "%s"' % (bucket))

#
# Run from S3
#
elif cmd == 'run':
	archivename = args[0]
	storedfilename = args[1]
	arguments = args[2:]

	bucket,key = storedfilename.split('/', 1)
	logger.info('run bucket: %s, key: %s'%(bucket, key))

	os.execv(storedfilename, arguments)
else:
	logger.error('unhandled, bye')
	sys.exit(1)

logger.debug('command handled')
sys.exit(0)