def receive(self):
data = self.channel['class'].receive()
# reassemble packets into segments
segments = []
buf = []
# reverse the response so we get oldest first
for msg in reversed(data):
# strip and decode the headers
tokenized = msg.split(" ", 2)
# if there aren't at least 3 tokens, the packet is not correct
# packet is: (packet number) (total packets in message) (msg body)
if len(tokenized) < 3:
continue
packet_no = base94.decode(str(tokenized[0]))
packet_total = base94.decode(str(tokenized[1]))
msg = ''.join(tokenized[2:])
if len(buf) == 0 and packet_no != 0:
# we caught the end of a segment, drop this packet
continue
# if the packet number isn't equal to the total packets in the segment
if packet_no + 1 != packet_total:
# we haven't found the last packet yet, so append this one to the buffer
buf.append(msg)
else:
# the buffer plus what we have here is a full sequence of packets
buf.append(msg)
# remove the headers and append all the data
segment = ''.join(buf)
segments.append(segment)
buf = []
if len(buf) != 0:
raise AssertionError(
"Buffer is not empty: packet must be missing.")
if not isinstance(segments, list):
raise TypeError(
'Data must be returned from channel receive function as an array.'
)
# the array is of individual 'packets' - i.e. metadata-wrapped bits of data
# this allows for timestamping messages, sending large messages as multiple
# fragments, etc.
output = []
# segments is an array of individual messages - decode them all one at a time
for datum in segments:
for encoder in reversed(self.encoders):
# go through the encoder chain in reverse to decode the data
# This allows decoders to be specified in the same order on both ends, and still work.
datum = encoder['class'].decode(datum)
# remove trailing newlines and spaces - they're added by some channels
output.append(datum.rstrip())
# write out the results
return output
def receive(self):
data = self.channel['class'].receive()
# reassemble packets into segments
segments = []
buf = []
# reverse the response so we get oldest first
for msg in reversed(data):
# strip and decode the headers
tokenized = msg.split(" ", 2)
# if there aren't at least 3 tokens, the packet is not correct
# packet is: (packet number) (total packets in message) (msg body)
if len(tokenized) < 3:
continue
packet_no = base94.decode(str(tokenized[0]))
packet_total = base94.decode(str(tokenized[1]))
msg = ''.join(tokenized[2:])
if len(buf) == 0 and packet_no != 0:
# we caught the end of a segment, drop this packet
continue
# if the packet number isn't equal to the total packets in the segment
if packet_no + 1 != packet_total:
# we haven't found the last packet yet, so append this one to the buffer
buf.append(msg)
else:
# the buffer plus what we have here is a full sequence of packets
buf.append(msg)
# remove the headers and append all the data
segment = ''.join(buf)
segments.append(segment)
buf = []
if len(buf) != 0:
raise AssertionError("Buffer is not empty: packet must be missing.")
if not isinstance(segments, list):
raise TypeError('Data must be returned from channel receive function as an array.')
# the array is of individual 'packets' - i.e. metadata-wrapped bits of data
# this allows for timestamping messages, sending large messages as multiple
# fragments, etc.
output = []
# segments is an array of individual messages - decode them all one at a time
for datum in segments:
for encoder in reversed(self.encoders):
# go through the encoder chain in reverse to decode the data
# This allows decoders to be specified in the same order on both ends, and still work.
datum = encoder['class'].decode(datum)
# remove trailing newlines and spaces - they're added by some channels
output.append(datum.rstrip())
# write out the results
return output
def send(self, data):
# header format: "lll nnn <data>"
# where "lll" is a 3 character base94 number representing this packet's index
# "nnn" is a 3 character base94 number representing the total number of packets in the fragment
# 3 digits in base 94 = max of 830583 packets per fragment
# this is 109,636,956 characters sent using Twitter,
# or ~100ish MB assuming each character is one byte
# pretty sure nobody will ever need more than 640k of ram
encoded = data
for encoder in self.encoders:
encoded = encoder['class'].encode(encoded)
chan = self.channel['class']
header_length = 8 # reserve characters for the headers
actual_length = chan.maxLength - header_length
# determine the number of packets we'll need to send
num_packets = int(math.ceil(len(encoded) / float(actual_length)))
# '~~~' is the largest three digit number in base94
max_packets = base94.decode("~~~")
if num_packets > max_packets:
raise ValueError("you can only send " + max_packets + " packets at a time")
for i in range(num_packets):
# wrap the data in headers
packet = base94.encode(i) + " " + base94.encode(num_packets) + " " + ''.join(encoded[i * actual_length:(i+1)*actual_length])
# double check that nothing went wrong
if len(packet) > chan.maxLength:
raise ValueError(self.channel['name'] + " cannot send more than " + str(chan.maxLength) + " characters")
# send it off
chan.send(packet)
def send(self, data):
# header format: "lll nnn <data>"
# where "lll" is a 3 character base94 number representing this packet's
# index "nnn" is a 3 character base94 number representing the total
# number of packets in the fragment
# 3 digits in base 94 = max of 830583 packets per fragment
# this is 109,636,956 characters sent using Twitter,
# or ~100ish MB assuming each character is one byte
# pretty sure nobody will ever need more than 640k of ram
encoded = data
for encoder in self.encoders:
encoded = encoder['class'].encode(encoded)
chan = self.channel['class']
header_length = 8 # reserve characters for the headers
if type(chan.maxLength) != int:
raise TypeError("Channel's maximum length must be an integer.")
actual_length = chan.maxLength - header_length
# determine the number of packets we'll need to send
num_packets = int(math.ceil(len(encoded) / float(actual_length)))
# '~~~' is the largest three digit number in base94
max_packets = base94.decode("~~~")
if num_packets > max_packets:
raise ValueError("you can only send " + max_packets +
" packets at a time")
for i in range(num_packets):
# wrap the data in headers
data_start = int(i * actual_length)
data_end = int((i + 1) * actual_length)
packet = base94.encode(i) + " " + base94.encode(
num_packets) + " " + ''.join(encoded[data_start:data_end])
# double check that nothing went wrong
if len(packet) > chan.maxLength:
raise ValueError(self.channel['name'] +
" cannot send more than " +
str(chan.maxLength) + " characters")
# send it off
chan.send(packet)