I am trying to make a simple image-sharing app in Python such that images can be shared from clients to a server. I am doing that using socket and sending images in form of numpy arrays. My problem is that when the app loads, I have made such that 6 images or less(if less than 6 present for an account) are sent from server to the client. Each image is stored in the server with a unique name identifying it. This name is of variable length and is decided by the user and is sent before the image array is sent. Also, the shape of the image is sent to reproduce it from the bytes at the client. But as the name length is not fixed, I am reading 10 buffer size. But if the name is smaller, it is reading the shape of the are that I am sending later as well. How do I fix this? Here is the sending image code:
def send_multiple_imgs(send_socket, imgs):
num = len(imgs)
send_socket.send(str(num).encode())
for img in imgs:
print(img)
send_socket.send(img.encode())
send_img(send_socket,imgs[img])
part of send_img function:
def send_img(send_socket,img):
send_socket.send(str(img.shape).encode())
.....
The later part is not important, I think.
Here is the receiving part:
def receive_multiple_img(recv_socket):
num = int(recv_socket.recv(1).decode())
imgs = {}
for i in range(num):
img_name = recv_socket.recv(10).decode()
print(img_name)
imgs[img_name] = recieve_image(recv_socket)
return imgs
What is happening is, I have an image named 'ds' of shape (200,200,4), but the img_name reads:
'ds(200, 20' and then it messes up the further sending and receiving as well. How do I fix this? I am using TCP protocol:
s = socket.socket(socket.AF_INET6, socket.SOCK_STREAM)
I am new to networking in python. So, please consider if this is a silly question
TCP, unlike UDP, is a stream protocol. That means that a single send() doesn't correspond to a single read(). A call to send() simply places the data to be sent into the TCP send buffer, and a call to read() simply returns the bytes from the TCP receive buffer. The bytes in the buffer could have come from a single send() on the other side or a hundered.
This is a fairy common misunderstanding that leads to many bugs. Here's me explaining this again, and again, and again, and again.
If you want to send() several separate messages or pieces of data, the reader must be able to tell the messages apart. There are many ways to do that, such as using fixed-length messages, prefixing the message's length before each message, or using a delimiter.
Related
I setted up a secure socket using Tor and socks, but i'm facing a problem when sending large amount of data
Sender:
socket.send(message.encode())
Receiver:
chunks = []
while 1:
part = connection.recv(4096)
chunks.append(part.decode())
if len(part) < 4096:
break
response = "".join(chunks)
Since the network speed is not consistent in a loop i don't always fill the 4096b buffer, so the loop breaks and i don't receive the full data.
Lowering the buffer size doesn't seem an option because the "packet" size can be as low as 20b sometimes
TCP can split your package data in any amount of pieces it wants. So you should never rely on other end of a socket on the size of the packet received. You have to invent another mechanism for detecting end of message/end of file.
If you are going to send only one blob and close socket, then on server side you just read until you get False value:
while True:
data = sock.recv(1024)
if data:
print(data)
# continue
else:
sock.close()
break
If you are going to send multiple messages, you have to decide, what will be the separator between them. For text protocols it is a good idea to use lineending. You can then enjoy the power of Twisted LineReceiver protocol and others.
If you are doing a binary protocol, it's a common practice to preface your each message with size byte/word/dword.
Try using structure to pass the length of the incoming data first to the receiver, "import struct". That way the receiving end knows exactly how much data to receive. In this example bytes are being sent over the socket, the examples here I've borrowed from my github upload github.com/nsk89/netcrypt for reference and cut out the encryption steps from the send function as well as it sending a serialised dictionary.
Edit I should also clarify that when you send data over the socket especially if your sending multiple messages they all sit in the stream as one long message. Not every message is 4096 bytes in length. If one is 2048 in length and the next 4096 and you receive 4096 on your buffers you'll receive the first message plus half of the next message or completely hang waiting for more data that doesn't exist.
data_to_send = struct.pack('>I', len(data_to_send)) + data_to_send # pack the length of data in the first four bytes of data stream, >I indicates internet byte order
socket_object.sendall(data_to_send) # transport data
def recv_message(socket_object):
raw_msg_length = recv_all(socket_object, 4) # receive first 4 bytes of data in stream
if not raw_msg_length:
return None
# unpack first 4 bytes using network byte order to retrieve incoming message length
msg_length = struct.unpack('>I', raw_msg_length)[0]
return recv_all(socket_object, msg_length) # recv rest of stream up to message length
def recv_all(socket_object, num_bytes):
data = b''
while len(data) < num_bytes: # while amount of data recv is less than message length passed
packet = socket_object.recv(num_bytes - len(data)) # recv remaining bytes/message
if not packet:
return None
data += packet
return data
By the way, no need to decode the every part before combine them to a chunk, combine all the parts to a chunk and then decode the chunk.
For your situation, the better way is using 2 steps.
Step1: sender send the size of the message, receiver take this size and ready to receive the message.
Step2: sender send the message, receiver combine the data if necessary.
Sender
# Step 1
socket.send( str(len(message.encode())).encode() )
# Step 2
socket.send(message.encode("utf-8"))
Receiver
# Step 1
message_size = connection.recv(1024)
print("Will receive message size：",message_size.decode())
# Step 2
recevied_size = 0
recevied_data = b''
while recevied_size < int(message_size.decode()):
part = connection.recv(1024)
recevied_size += len(part)
recevied_data += part
else:
print(recevied_data.decode("utf-8", "ignore"))
print("message receive done ....",recevied_size)
I am using TCP with Python sockets, transfering data from one computer to another. However the recv command reads more than it should in the serverside, I could not find the issue.
client.py
while rval:
image_string = frame.tostring()
sock.sendall(image_string)
rval, frame = vc.read()
server.py
while True:
image_string = ""
while len(image_string) < message_size:
data = conn.recv(message_size)
image_string += data
The length of the message is 921600 (message_size) so it is sent with sendall, however when recieved, when I print the length of the arrived messages, the lengths are sometimes wrong, and sometimes correct.
921600
921600
921923 # wrong
922601 # wrong
921682 # wrong
921600
921600
921780 # wrong
As you see, the wrong arrivals have no pattern. As I use TCP, I expected more consistency, however it seems the buffers are mixed up and somehow recieving a part of the next message, therefore producing a longer message. What is the issue here ?
I tried to add just the relevant part of the code, I can add more if you wish, but the code performs well on localhost but fails on two computers, so there should be no errors besides the transmitting part.
Edit1: I inspected this question a bit, it mentions that all send commands in the client may not be recieved by a single recv in the server, but I could not understand how to apply this to practice.
TCP is a stream protocol. There is ABSOLUTELY NO CONNECTION between the sizes of the chunks of data you send, and the chunks of data you receive. If you want to receive data of a known size, it's entirely up to you to only request that much data: you're currently requesting the total length of the data each time, which is going to try to read too much except in the unlikely event of the entire data being retrieved by the first .recv() call. Basically, you need to do something like data = conn.recv(message_size - len(image_string)) to reflect the fact that the amount of remaining data is decreasing.
Think of TCP as a raw stream of bytes. It is your responsibility to track where you are in the stream and interpret it correctly. Buffer what you read and only extract what you currently need.
Here's an (untested) class to illustrate:
class Buffer:
def __init__(self,socket):
self.socket = socket
self.buffer = b''
def recv_exactly(self,count):
# Could return less if socket closes early...
while len(self.buffer) < count:
data = self.socket.recv(4096)
if not data: break
self.buffer += data
ret,self.buffer = self.buffer[:count],self.buffer[count:]
return ret
The recv always requests the same amount of data and queues it in a buffer. recv_exactly only returns the number of bytes requested and leaves any extra in the buffer.
Situation:
I have a sendersocket which is bound to localhost UDP port 33100.
I have a receiversocket socket bound to localhost UDP port 33101.
The sender socket sends 4500 bytes of data (the string "hello man" * 500). On the receiver side, I have an epoll object which waits for the EPOLLIN event on the receiversocket. When there's an event, I do:
while True:
msg_received = receiver_socket.recv(128)
if msg_received.decode("UTF-8") == "":
break
else:
print msg_received.decode("UTF-8")
Problem:
The main problem is that I cannot read again after I have read the first 128 bytes of data. The sender side says it sent 4500 bytes of data as expected.
If the sender sends the same 4500 bytes of data again, the EPOLLIN event is registered again and I read the new string. Somehow, the buffer get's cleared after my first read.
Now even though the sender just sent me 4500 bytes of data, the first recv gives me 128 bytes of data and then nothing is recved after that.
I am probably doing something really stupid so please enlighten me. I want to receive all the 4500 bytes of data.
You should always call recv with 65535 (the maximum size of a UDP packet) if you don't already know the size of the packet. Otherwise, when you call recv, the entire packet is marked as read and cleared from the buffer, but then only the first 128 bytes are fed to msg_received.
edit: when(if) you transition to receiving data only across a network, you can use a smaller number for recv, as noted in the Docs
If you know that you'll be getting 4500 bytes of data, you can call receiver_socket.recv(4500)
What your code is saying is that the maximum bytes to read is 128 with receiver_socket.recv(128)
See the python documentation for sockets
I'm trying to send the contents of a file over to the client 1 line at a time
so that the client (written in Objective-C) can process each line individually.
However the log of the client shows that the data being send over from the server
is all coming through as 1 line and is apparently too large so the it cuts off mid
way through one line causing the client to crash because of the unexpected syntax.
Is there something i'm doing on the server(written in python with twisted) that is
causing the lines to not be sent separately?
Here is the particular code in the server that is holding me up at the moment.
def sendLine(self, line):
self.transport.write(line + '\r\n')
def updateShiftList(self):
#open the datesRequested file for the appropriate store and load the dates into a list
fob = open('stores/'+self.storeName+'/requests/datesRequested','r')
DATES_REQUESTED = fob.read()
datesRequested = DATES_REQUESTED.split('\n')
#open each date file that is listed in datesRequested
for date in datesRequested:
if os.path.isfile('stores/'+self.storeName+'/requests/' + date):
fob2 = open('stores/'+self.storeName+'/requests/' + date,'r')
#load the file into memory and split the individual requests up
THE_REQUESTS = fob2.read()
thedaysRequests = THE_REQUESTS.split('\n')
for oneRequest in thedaysRequests:
if len(oneRequest) > 4:
print "*)[*_-b4.New_REQUEST:"+oneRequest
self.sendLine('*)[*_-b4.New_REQUEST:'+oneRequest)
fob2.close()
fob.close()
So frustrating and i'm sure it's something easy. Thanks.
This question concerns a topic that is frequently raised. There are a number of questions on stackoverflow about the same issue:
TCP messages getting coalesced
gen_tcp smushed messages
Broken TCP messages
How to separate TCP socket messages in node.js
etc.
TCP sends an ordered, reliable stream of data.
Ordering means that bytes sent first arrive first.
Reliability means bytes sent will be delivered, or the connection will break. Data is not silently dropped.
Streaming is what mostly concerns your question. A stream is not divided up into separate messages. It consists of bytes, and the "boundary" between those bytes can move arbitrarily.
If you send "hello, " and then you send "world", the boundary between those two strings in the stream may disappear. The peer may receive "hello, world", or "h", "ello, world", or "he", "ll", "o,", " w", "or", "ld".
This is the very reason people use line-oriented protocols. The "\r\n" at the end of each logical message lets the receiver buffer and split the stream up into those original logical messages.
For a deeper dive, I recommend this video of a recent PyCon presentation: http://pyvideo.org/speaker/417/glyph
This all points towards the other end of your connection, the ObjC client, as the source of your misbehavior.
See edits below.
I have two programs that communicate through sockets. I'm trying to send a block of data from one to the other. This has been working with some test data, but is failing with others.
s.sendall('%16d' % len(data))
s.sendall(data)
print(len(data))
sends to
size = int(s.recv(16))
recvd = ''
while size > len(recvd):
data = s.recv(1024)
if not data:
break
recvd += data
print(size, len(recvd))
At one end:
s = socket.socket()
s.connect((server_ip, port))
and the other:
c = socket.socket()
c.bind(('', port))
c.listen(1)
s,a = c.accept()
In my latest test, I sent a 7973903 byte block and the receiver reports size as 7973930.
Why is the data block received off by 27 bytes?
Any other issues?
Python 2.7 or 2.5.4 if that matters.
EDIT: Aha - I'm probably reading past the end of the send buffer. If remaining bytes is less than 1024, I should only read the number of remaining bytes. Is there a standard technique for this sort of data transfer? I have the feeling I'm reinventing the wheel.
EDIT2: I'm screwing up by reading the next file in the series. I'm sending file1 and the last block is 997 bytes. Then I send file2, so the recv(1024) at the end of file1 reads the first 27 bytes of file2.
I'll start another question on how to do this better.
Thanks everyone. Asking and reading comments helped me focus.
First, the line
size = int(s.recv(16))
might read less than 16 bytes — it is unlikely, I will grant, but possible depending on how the network buffers align. The recv() call argument is a maximum value, a limit on how much data you are willing to receive. But you might only receive one byte. The operating system will generally give you control back once at least one byte has arrived, maybe (depending on the OS and on how busy the CPU is) after waiting another few milliseconds in case a second packet arrives with some further data, so that it only has to wake you up once instead of twice.
So you would want to say instead (to do the simplest possible loop; other variants are possible):
data = ''
while len(data) < 16:
more = s.recv(16 - len(data))
if not more:
raise EOFError()
data += more
This is indeed a wheel nearly everyone re-invents because it is so often needed. And your own code needs it a second time: your while loop needs its recv() to count down, asking for smaller and smaller limits until finally it has received exactly the number of bytes that were promised, and no more.