#!/usr/bin/env python3.3
import argparse, os
parser = argparse.ArgumentParser()
parser.add_argument('-c', '--config', dest = 'config', help = 'config file location', metavar = 'FILE', default = 'samesite.conf')
args = parser.parse_args()
assert os.access(args.config, os.R_OK), "Fatal error: can't read {}".format(args.config)
import configparser
config = configparser.ConfigParser({
'port': '8008',
'verbose': 'no',
'noetag': 'no',
'noparts': 'no',
'strip': '',
'sub': '',
'proto': 'http',
})
config.read(args.config)
cache_dir = os.path.realpath(os.path.dirname(args.config))
import re
for section in config.sections():
if section != 'DEFAULT':
if 'dir' in config[section]:
if not re.compile('^/.*').match(config[section]['dir']):
config[section]['dir'] = cache_dir + os.sep + section
thisDir = re.compile('^(.*)/$').match(config[section]['dir'])
if thisDir:
config[section]['dir'] = thisDir.group(1)
if not re.compile('^/(.*)$').match(config[section]['dir']):
config[section]['dir'] = cache_dir + os.sep + config[section]['dir']
else:
config[section]['dir'] = cache_dir + os.sep + section
if not 'root' in config[section]:
config[section]['root'] = section
#assert options.port or os.access(options.log, os.R_OK), 'Log file unreadable'
const_desc_fields = set(['Content-Length', 'Last-Modified', 'Pragma'])
const_ignore_fields = set([
'Accept-Ranges', 'Age',
'Cache-Control', 'Connection', 'Content-Type',
'Date',
'Expires',
'Referer',
'Server',
'Via',
'X-CCC', 'X-CID', 'X-Cache', 'X-Cache-Lookup', 'X-Livetool', 'X-Powered-By',
])
block_size = 8192
import bsddb3.dbshelve, copy, datetime, http.server, spacemap, urllib.request, urllib.error
class MyRequestHandler(http.server.BaseHTTPRequestHandler):
def __process(self):
# reload means file needs to be reloaded to serve request
reload = False
# recheck means file needs to be checked, this also means that if file hav been modified we can serve older copy
recheck = False
# file_stat means file definitely exists
file_stat = None
# requested_ranges holds data about any range requested
requested_ranges = None
# records holds data from index locally, should be written back upon successfull completion
record = None
myPath = re.compile('^(.*?)(\?.*)$').match(self.path)
if myPath:
my_path = myPath.group(1)
else:
my_path = self.path
if not config.has_section(self.headers['Host']):
config.add_section(self.headers['Host'])
config[self.headers['Host']]['root'] = self.headers['Host']
config[self.headers['Host']]['dir'] = cache_dir + os.sep + self.headers['Host']
config_host = config[self.headers['Host']]
if config_host['sub'] != None and config_host['strip'] != None and len(config_host['strip']) > 0:
string = re.compile(config_host['strip']).sub(config_host['sub'], my_path)
my_path = string
my_path_b = my_path.encode('utf-8')
info = 'Checking file: ' + my_path
if not os.access(config_host['dir'], os.X_OK):
os.mkdir(config_host['dir'])
# this is file index - everything is stored in this file
# _parts - list of stored parts of file
# _time - last time the file was checked
# everything else is just the headers
index = bsddb3.dbshelve.open(config_host['dir'] + os.sep + '.index')
desc_fields = const_desc_fields.copy()
ignore_fields = const_ignore_fields.copy()
if config_host['noetag'] == 'no':
desc_fields.add('ETag')
else:
ignore_fields.add('ETag')
proxy_ignored = set([
'Accept', 'Accept-Charset', 'Accept-Encoding', 'Accept-Language',
'Cache-Control', 'Connection', 'Content-Length', 'Cookie',
'Host',
'If-Modified-Since', 'If-None-Match', 'If-Unmodified-Since',
'Referer',
'UA-CPU', 'User-Agent',
'Via',
'X-Forwarded-For', 'X-Last-HR', 'X-Last-HTTP-Status-Code', 'X-Old-UID', 'X-Removed', 'X-Real-IP', 'X-Retry-Count',
])
print('===============[ {} request ]==='.format(self.command))
for header in self.headers:
if header in proxy_ignored:
pass
elif header in ('Range'):
isRange = re.compile('bytes=(\d+)-(\d+)').match(self.headers[header])
if isRange:
requested_ranges = spacemap.SpaceMap({int(isRange.group(1)): int(isRange.group(2)) + 1})
else:
return()
elif header in ('Pragma'):
if my_path_b in index:
index[my_path_b][header] = self.headers[header]
else:
print('Unknown header - ', header, ': ', self.headers[header], sep='')
return()
print(header, self.headers[header])
# creating file name from my_path
file_name = config_host['dir'] + os.sep + re.compile('%20').sub(' ', my_path)
# partial file or unfinished download
temp_name = config_host['dir'] + os.sep + '.parts' + re.compile('%20').sub(' ', my_path)
# creating empty placeholder in index
# if there's no space map and there's no file in real directory - we have no file
# if there's an empty space map - file is full
# space map generally covers every bit of file we don't posess currently
if not my_path_b in index:
info += '\nThis one is new.'
reload = True
record = {}
else:
# forcibly checking file if no file present
record = index[my_path_b]
if os.access(file_name, os.R_OK):
info += '\nFull file found.'
file_stat = os.stat(file_name)
elif '_parts' in index[my_path_b] and os.access(temp_name, os.R_OK):
info += '\nPartial file found.'
file_stat = os.stat(temp_name)
recheck = True
else:
info += '\nFile not found or inaccessible.'
record['_parts'] = None
reload = True
if not '_parts' in record:
record['_parts'] = None
if record['_parts'] == None:
recheck = True
# forcibly checking file if file size doesn't match with index data
if not reload:
if '_parts' in record and record['_parts'] == spacemap.SpaceMap():
if 'content-length' in record and file_stat and file_stat.st_size != int(record['content-length']):
info += '\nFile size is {} and stored file size is {}.'.format(file_stat.st_size, record['content-length'])
record['_parts'] = None
reload = True
# forcibly checking file if index holds Pragma header
if not reload and 'pragma' in record and record['pragma'] == 'no-cache':
info +='\nPragma on: recheck imminent.'
recheck = True
# skipping file processing if there's no need to recheck it and we have checked it at least 4 hours ago
if not recheck and not reload and '_time' in record and (record['_time'] - datetime.datetime.now() + datetime.timedelta(hours = 4)).days < 0:
info += '\nFile is old - rechecking.'
recheck = True
print(info)
if reload or recheck:
try:
request = config_host['proto'] + '://' + config_host['root'] + self.path
my_headers = {}
for header in ('Accept', 'Cache-Control', 'Cookie', 'Referer', 'User-Agent'):
if header in self.headers:
my_headers[header] = self.headers[header]
needed = None
if self.command not in ('HEAD'):
if '_parts' in record and record['_parts'] != None:
if config_host['noparts'] != 'no' or requested_ranges == None or requested_ranges == spacemap.SpaceMap():
needed = record['_parts']
else:
needed = record['_parts'] & requested_ranges
elif config_host['noparts'] =='no' and requested_ranges != None and requested_ranges != spacemap.SpaceMap():
needed = requested_ranges
ranges = ()
print('Missing ranges: {}, requested ranges: {}, needed ranges: {}.'.format(record['_parts'], requested_ranges, needed))
if needed != None and len(needed) > 0:
needed.rewind()
while True:
range = needed.pop()
if range[0] == None:
break
ranges += '{}-{}'.format(range[0], range[1] - 1),
my_headers['Range'] = 'bytes=' + ','.join(ranges)
#my_headers['Accept-Encoding'] = 'gzip, compress, deflate, identity; q=0'
request = urllib.request.Request(request, headers = my_headers)
source = urllib.request.urlopen(request, timeout = 60)
new_record = {}
new_record['_parts'] = record['_parts']
headers = source.info()
if 'Content-Encoding' in headers and headers['Content-Encoding'] == 'gzip':
import gzip
source = gzip.GzipFile(fileobj=source)
# stripping unneeded headers (XXX make this inplace?)
for header in headers:
if header in desc_fields:
#if header == 'Pragma' and headers[header] != 'no-cache':
if header == 'Content-Length':
if 'Content-Range' not in headers:
new_record[header] = int(headers[header])
else:
new_record[header] = headers[header]
elif header == 'Content-Range':
range = re.compile('^bytes (\d+)-(\d+)/(\d+)$').match(headers[header])
if range:
new_record['Content-Length'] = int(range.group(3))
else:
assert False, 'Content-Range unrecognized.'
elif not header in ignore_fields:
print('Undefined header "', header, '": ', headers[header], sep='')
# comparing headers with data found in index
# if any header has changed (except Pragma) file is fully downloaded
# same if we get more or less headers
old_keys = set(record.keys())
old_keys.discard('_time')
old_keys.discard('Pragma')
more_keys = set(new_record.keys()) - old_keys
more_keys.discard('Pragma')
less_keys = old_keys - set(new_record.keys())
if len(more_keys) > 0:
if len(old_keys) != 0:
print('More headers appear:', more_keys)
reload = True
elif len(less_keys) > 0:
print('Less headers appear:', less_keys)
else:
for key in record.keys():
if key[0] != '_' and key != 'Pragma' and record[key] != new_record[key]:
print('Header "', key, '" changed from [', record[key], '] to [', new_record[key], ']', sep='')
print(type(record[key]), type(new_record[key]))
reload = True
if reload:
print('Reloading.')
if os.access(temp_name, os.R_OK):
os.unlink(temp_name)
if os.access(file_name, os.R_OK):
os.unlink(file_name)
if 'Content-Length' in new_record:
new_record['_parts'] = spacemap.SpaceMap({0: int(new_record['Content-Length'])})
if not new_record['_parts']:
new_record['_parts'] = spacemap.SpaceMap()
print(new_record)
# downloading file or segment
if 'Content-Length' in new_record:
if needed == None:
needed = new_record['_parts']
else:
if len(needed) > 1:
print("Multipart requests currently not supported.")
assert False, 'Skip this one for now.'
#else:
#assert False, 'No content-length or Content-Range header.'
new_record['_time'] = datetime.datetime.now()
if self.command not in ('HEAD'):
# file is created at temporary location and moved in place only when download completes
if not os.access(temp_name, os.R_OK):
empty_name = config_host['dir'] + os.sep + '.tmp'
with open(empty_name, 'w+b') as some_file:
pass
os.renames(empty_name, temp_name)
temp_file = open(temp_name, 'r+b')
if requested_ranges == None and needed == None:
needed = new_record['_parts']
needed.rewind()
countdown = 16
while True:
# XXX can make this implicit - one request per range
(start, end) = needed.pop()
if start == None:
break
stream_last = start
old_record = copy.copy(new_record)
if end - start < block_size:
req_block_size = end - start
else:
req_block_size = block_size
buffer = source.read(req_block_size)
length = len(buffer)
while length > 0 and stream_last < end:
stream_pos = stream_last + length
assert stream_pos <= end, 'Received more data then requested: pos:{} start:{} end:{}.'.format(stream_pos, start, end)
temp_file.seek(stream_last)
temp_file.write(buffer)
x = new_record['_parts'] - spacemap.SpaceMap({stream_last: stream_pos})
new_record['_parts'] = new_record['_parts'] - spacemap.SpaceMap({stream_last: stream_pos})
countdown -= 1
if countdown == 0:
index[my_path_b] = old_record
index.sync()
countdown = 16
old_record = copy.copy(new_record)
stream_last = stream_pos
if end - stream_last < block_size:
req_block_size = end - stream_last
buffer = source.read(req_block_size)
length = len(buffer)
# moving downloaded data to real file
temp_file.close()
index[my_path_b] = new_record
index.sync()
except urllib.error.HTTPError as error:
# in case of error we don't need to do anything actually,
# if file download stalls or fails the file would not be moved to it's location
self.send_response(error.code)
self.end_headers()
print(error, repr(my_headers))
return
#print(index[my_path_b])
if not os.access(file_name, os.R_OK) and os.access(temp_name, os.R_OK) and '_parts' in index[my_path_b] and index[my_path_b]['_parts'] == spacemap.SpaceMap():
# just moving
# drop old dirs XXX
print('Moving temporary file to new destination.')
os.renames(temp_name, file_name)
if not my_path_b in index:
self.send_response(502)
self.end_headers()
return
if self.command == 'HEAD':
self.send_response(200)
if 'Content-Length' in index[my_path_b]:
self.send_header('Content-Length', index[my_path_b]['Content-Length'])
self.send_header('Accept-Ranges', 'bytes')
self.send_header('Content-Type', 'application/octet-stream')
if 'Last-Modified' in index[my_path_b]:
self.send_header('Last-Modified', index[my_path_b]['Last-Modified'])
self.end_headers()
else:
if ('_parts' in index[my_path_b] and index[my_path_b]['_parts'] != spacemap.SpaceMap()) or not os.access(file_name, os.R_OK):
file_name = temp_name
with open(file_name, 'rb') as real_file:
file_stat = os.stat(file_name)
if 'Range' in self.headers:
self.send_response(206)
ranges = ()
requested_ranges.rewind()
while True:
pair = requested_ranges.pop()
if pair[0] == None:
break
ranges += '{}-{}'.format(pair[0], str(pair[1] - 1)),
self.send_header('Content-Range', 'bytes {}/{}'.format(','.join(ranges), index[my_path_b]['Content-Length']))
else:
self.send_response(200)
self.send_header('Content-Length', str(file_stat.st_size))
requested_ranges = spacemap.SpaceMap({0: file_stat.st_size})
if 'Last-Modified' in index[my_path_b]:
self.send_header('Last-Modified', index[my_path_b]['Last-Modified'])
self.send_header('Content-Type', 'application/octet-stream')
self.end_headers()
if self.command in ('GET'):
if len(requested_ranges) > 0:
requested_ranges.rewind()
(start, end) = requested_ranges.pop()
else:
start = 0
# XXX ugly hack
if 'Content-Length' in index[my_path_b]:
end = index[my_path_b]['Content-Length']
else:
end = 0
real_file.seek(start)
if block_size > end - start:
req_block_size = end - start
else:
req_block_size = block_size
buffer = real_file.read(req_block_size)
length = len(buffer)
while length > 0:
self.wfile.write(buffer)
start += len(buffer)
if req_block_size > end - start:
req_block_size = end - start
if req_block_size == 0:
break
buffer = real_file.read(req_block_size)
length = len(buffer)
def do_HEAD(self):
return self.__process()
def do_GET(self):
return self.__process()
server = http.server.HTTPServer(('127.0.0.1', int(config['DEFAULT']['port'])), MyRequestHandler)
server.serve_forever()
#gevent.joinall()