We Keep Coding

python socket - how to complete/close the connection on the client side?

server.py:
json files from NVD are used here
import socket, json, random, threading, zipfile, requests, re, zipfile
from bs4 import BeautifulSoup
from zipfile import *
def listen_user(user):
for x in range(2018,2021,1):
filename = "nvdcve-1.1-" + str(x) + ".json"
print(filename)
with open(filename, 'rb') as file:
sendfile = file.read()
user.sendall(sendfile)
print('file sent' + str(x))
def start_server():
while True:
user_socket, address = server.accept()
print(f"User <{address[0]}> connected!")
users.append(user_socket)
listen_accepted_user = threading.Thread(
target=listen_user,
args=(user_socket,)
)
listen_accepted_user.start()
if __name__ == '__main__':
users = []
server = socket.socket(
socket.AF_INET,
socket.SOCK_STREAM,
)
server.bind(
("127.0.0.1", 100)
)
server.listen(5)
print('waiting for connection...')
start_server()
client.py
import socket, json, random
from threading import Thread
def start_client(client):
savefilename = str(random.randint(1,10)) + 'new.json'
print(savefilename)
with client,open(savefilename,'wb') as file:
while True:
recvfile = client.recv(4096)
if not recvfile:
print('1 client')
break
file.write(recvfile)
file.close()
print('2 client')
client.close()
if __name__ == '__main__':
client = socket.socket(
socket.AF_INET,
socket.SOCK_STREAM,
)
client.connect(
("127.0.0.1", 100)
)
start_client(client)
when I send files - they are sent almost in full, but the program does not reach the line "print ('1 client')" or "print ('2 client')"
and the *new file contains all lines except a few dozen of the last
please help - how to fix the code?

recvfile = client.recv(4096) is inside the while loop and it is continuously waiting for the next bytes to receive. The client doesn't know the files are sent, so it waits for the next 4096 bytes and doesn't exit the loop.
To let the client know that the file transfer is completed, you can send a message from the server.py which you can validate in the client and break the loop as shown below.
server.py
def listen_user(user):
for x in ["f.json","g.json"]:
filename = x
print(filename)
with open(filename, 'rb') as file:
sendfile = file.read()
user.sendall(sendfile)
print('file sent' + str(x))
user.send(b"Done")
Client.py
def start_client(client):
savefilename = str(random.randint(1,10)) + 'new.json'
print(savefilename)
with client,open(savefilename,'wb') as file:
while True:
recvfile = client.recv(4096)
if recvfile.decode("utf-8") =="Done":
print('1 client')
file.close()
break
file.write(recvfile)
print('2 client')
client.close()

The call client.recv(4096) means that you are waiting for 4096 bytes to be received, then doing something with those bytes. What's likely happening in this case is that you're writing out all of the bytes, minus those that don't quite fill up the buffer at the end. This leaves the client waiting with a buffer with space that is doesn't think it is ready to write out yet.
I'm guessing that you're assuming that client.recv() will return an empty string once you've gotten all the data; this is not the case based on your code. If you want the client to be able to terminate the connection, you're going to need to send some kind of control sequence or try to otherwise assess the bytes received from the server to determined when it's time to close the connection. If you do this, you will probably want to set bufsize when calling client.recv() to 1, and instead use some other method to buffer before you write to a file.
For instance, since you're sending JSON data, you could concatenate the bytes to a variable and then repeatedly try to parse JSON. Once you have managed to successfully parse JSON, you can terminate the connection on the client side (though this would mean you have to open a new connection per file you're sending).
However, that raises the question: why do you need to close from the client side? Usually the server will just close the connection once it is done sending all of the relevant data.

Transfer contents of a folder over network by python

I am facing a problem writing a program to send contents of a folder over the network by using Python. There are a lot of examples out there, all the examples I found are assuming the receiver side knew name of the file he want to receive. The program I am trying to do assuming that the receiver side agree to receive a files and there is no need to request a file by its name from the server. Once the connection established between the server and the client, the server start send all files inside particular folder to the client. Here is a image to show more explanation:example here
Here are some programs that do client server but they send one file and assume the receiver side knew files names, so the client should request a file by its name in order to receive it.
Note: I apologies for English grammar mistakes.
https://www.youtube.com/watch?v=LJTaPaFGmM4
http://www.bogotobogo.com/python/python_network_programming_server_client_file_transfer.php
python socket file transfer
Here is best example I found:
Server side:
import sys
import socket
import os
workingdir = "/home/SomeFilesFolder"
host = ''
skServer = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
skServer.bind((host, 1000))
skServer.listen(10)
print "Server Active"
bFileFound = 0
while True:
Content, Address = skServer.accept()
print Address
sFileName = Content.recv(1024)
for file in os.listdir(workingdir):
if file == sFileName:
bFileFound = 1
break
if bFileFound == 0:
print sFileName + " Not Found On Server"
else:
print sFileName + " File Found"
fUploadFile = open("files/" + sFileName, "rb")
sRead = fUploadFile.read(1024)
while sRead:
Content.send(sRead)
sRead = fUploadFile.read(1024)
print "Sending Completed"
break
Content.close()
skServer.close()
Client side:
import sys
import socket
skClient = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
skClient.connect(("ip address", 1000))
sFileName = raw_input("Enter Filename to download from server : ")
sData = "Temp"
while True:
skClient.send(sFileName)
sData = skClient.recv(1024)
fDownloadFile = open(sFileName, "wb")
while sData:
fDownloadFile.write(sData)
sData = skClient.recv(1024)
print "Download Completed"
break
skClient.close()
if there is a way to eliminate this statement from the client side:
sFileName = raw_input("Enter Filename to download from server : ")
and make the server side send all files one by one without waiting for the client to pick a file.

Here's an example that recursively sends anything in the "server" subdirectory to a client. The client will save anything received in a "client" subdirectory. The server sends for each file:
The path and filename relative to the server subdirectory, UTF-8-encoded and terminated with a newline.
The file size in decimal as a UTF-8-encoded string terminated with a newline.
Exactly "file size" bytes of file data.
When all files are transmitted the server closes the connection.
server.py
from socket import *
import os
CHUNKSIZE = 1_000_000
sock = socket()
sock.bind(('',5000))
sock.listen(1)
while True:
print('Waiting for a client...')
client,address = sock.accept()
print(f'Client joined from {address}')
with client:
for path,dirs,files in os.walk('server'):
for file in files:
filename = os.path.join(path,file)
relpath = os.path.relpath(filename,'server')
filesize = os.path.getsize(filename)
print(f'Sending {relpath}')
with open(filename,'rb') as f:
client.sendall(relpath.encode() + b'\n')
client.sendall(str(filesize).encode() + b'\n')
# Send the file in chunks so large files can be handled.
while True:
data = f.read(CHUNKSIZE)
if not data: break
client.sendall(data)
print('Done.')
The client creates a "client" subdirectory and connects to the server. Until the server closes the connection, the client receives the path and filename, the file size, and the file contents and creates the file in the path under the "client" subdirectory.
client.py
from socket import *
import os
CHUNKSIZE = 1_000_000
# Make a directory for the received files.
os.makedirs('client',exist_ok=True)
sock = socket()
sock.connect(('localhost',5000))
with sock,sock.makefile('rb') as clientfile:
while True:
raw = clientfile.readline()
if not raw: break # no more files, server closed connection.
filename = raw.strip().decode()
length = int(clientfile.readline())
print(f'Downloading {filename}...\n Expecting {length:,} bytes...',end='',flush=True)
path = os.path.join('client',filename)
os.makedirs(os.path.dirname(path),exist_ok=True)
# Read the data in chunks so it can handle large files.
with open(path,'wb') as f:
while length:
chunk = min(length,CHUNKSIZE)
data = clientfile.read(chunk)
if not data: break
f.write(data)
length -= len(data)
else: # only runs if while doesn't break and length==0
print('Complete')
continue
# socket was closed early.
print('Incomplete')
break
Put any number of files and subdirectories under a "server" subdirectory in the same directory as server.py. Run the server, then in another terminal run client.py. A client subdirectory will be created and the files under "server" copied to it.

So... I've decided I've posted enough in comments and I might as well post a real answer. I see three ways to do this: push, pull, and indexing.
Push
Recall the HTTP protocol. The client asks for a file, the server locates it, and sends it. So get a list of all the files in a directory and send them all together. Better yet, tar them all together, zip them with some compression algorithm, and send that ONE file. This method is actually pretty much industry standard among Linux users.
Pull
I identifed this in the comments, but it works like this:
Client asks for directory
Server returns a text file containing the names of all the files.
Client asks for each file.
Index
This technique is the least mutable of the three. Keep an index of all the files in the directory, named INDEX.xml (funny enough, you could model the entire directory tree in xml.) your client will request the xml file, then walk the tree requesting other files.

you need to send os.listdir() by using json.dumps() and encode it as utf-8
at client side you need to decode and use json.loads() so that list will be transfer to client
place sData = skClient.recv(1024) before sFileName = raw_input("Enter Filename to download from server : ") so that the server file list can be display
you can find at here its a interesting tool
https://github.com/manoharkakumani/mano

FTP server breaks connection before download is complete

Sometimes, FTP server closes connection before file is completely downloaded..
Here is my code:
ftp = ftplib.FTP(site)
ftp.login(user, pw)
ftp.cwd(dir)
remotefiles = ftp.nlst()
for file in remotefiles:
if fnmatch.fnmatch(file, match_text):
if os.path.exists(file):
if True: print file, 'already fetched'
else:
if True: print 'Downloading', file
local = open(file, 'wb')
try:
ftp.retrbinary('RETR ' + file, local.write)
finally:
local.close()
if True: print 'Download done.'

You can specify a timeout parameter in the FTP constructor and set it to 0 or something very large value like sys.maxint.
class ftplib.FTP([host[, user[, passwd[, acct[, timeout]]]]])
Additionally, you can turn on debugging to see what's going on behind the scenes.
ftp = ftplib.FTP(site, user, pw, timeout=0)
ftp.set_debuglevel(2)
Hope this helps.

Not able to receive file from client in python

I am trying to program compilation server which compiles a C program sent by client and returns an object file which can then be linked and executed at the client. Here are my client and server programs respectively
client.py:
# Compilation client program
import sys, socket, string
File = raw_input("Enter the file name:")
ssock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
ssock.connect(('localhost', 5000))
csock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
csock.connect(('localhost', 5001))
f = open(File, "rb")
data = f.read()
f.close()
ssock.send(File) #send filename
ssock.send(data) #send file
fd=raw_input("Enter a key to start recieving object file:")
data=csock.recv(1024) #receive status
if data=="sucess\n":
File=File.replace(".c",".o") #objectfile name
print "Object file, "+File+", recieved sucessfully"
else:
print "There are compilation errors in " + File
File="error.txt" #errorfile name
print "Errors are reported in the file error.txt"
fobj=open(File,"wb")
while 1:
data=ssock.recv(1024) # if any error in c sourcefile then error gets
# eported in errorfile "error.txt" else objectfile is
# returned from server
if not data:break
fobj.write(data)
fobj.close()
ssock.close()
csock.close()
server.py
#Compilation Server program
import subprocess
import socket, time, string, sys, urlparse, os
ssock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
ssock.bind(('', 5000))
ssock.listen(2)
print 'Server Listening on port 5000'
csock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
csock.bind(('', 5001))
csock.listen(2)
print 'Control server listening on port 5001'
client, claddr = ssock.accept()
controlsoc, caddr = csock.accept()
filename=client.recv(1024) #receive filename
print filename
############### This code is not working, i'm not getting the reason #######
############### I want to receive a file more than 1KB from client #######
f = open(filename,"wb") #receive file======
while 1:
data = client.recv(1024)
if not data: break
f.write(data)
f.close()
###############
###############
data="gcc -c " + filename + " 2> error.txt" #shell command to execute c source file
#report errors if any to error.txt
from subprocess import call
call(data,shell=True) #executes the above shell command
fil = filename.replace(".c",".o")
if (os.path.isfile(fil))== True: #test for existence of objectfile
data = "sucess\n" #no objectfile => error in compilation
filename = filename.replace(".c",".o")
else:
data = "unsucessful\n"
print data+"hi"
filename = "error.txt"
controlsoc.send(data)
f = open(filename,"rb")
data=f.read()
f.close()
print data
client.send(data)
client.close()
controlsoc.close()
I'm not able to recieve files of multiple KB. Is there any flaw in my code or how should i modify my code in order to achieve my objective of coding a compilation server.
Please help me with this regard..Thanks in advance

The problem here is you assume that ssock.send(File) will result in filename=client.recv(1024) reading exactly the filename and not more, but in fact the receiving side has no idea where the filename ends and you end up getting the file name and part of the data in the filename variable.
TCP connection is a bi-directional stream of bytes. It doesn't know about boundaries of your messages. One send might correspond to more then one recv on the other side (and the other way around). You need an application-level protocol on top of raw TCP.
The easiest in your case would be to send a text line in the form file-size file-name\n as a header. This way your server would be able to not only separate header from file data (via newline) but also know how many bytes of file content to expect, and reuse same TCP connection for multiple files.

python socket file transfer

I'm trying to write transfer files or chunks of data over a socket. I feel as if I'm reinventing the wheel, but my searches for a simple solution have failed (everything I find is either too simple or too complex). The server would run on a phone running python 2.5.4. The intended application would be to sync music files between the phone and a host computer.
This is the guts of what I have, which appears to work. I send and receive 'ok' to break up streams.
Is sending 'ok' back and forth essentially as stop bits to break up streams of data a reasonable technique?
Is there a standard way to do this?
Running any sort of library server (ftp, http) on the phone is not a useful solution given the limits of the phone's memory and processing power.
server:
import socket
c = socket.socket(socket.AF_INET,socket.SOCK_STREAM)
c.bind(('', 1234))
c.listen(1)
s,a = c.accept()
while True:
data = s.recv(1024)
cmd = data[:data.find('\n')]
if cmd == 'get':
x, file_name, x = data.split('\n', 2)
s.sendall('ok')
with open(file_name, 'rb') as f:
data = f.read()
s.sendall('%16d' % len(data))
s.sendall(data)
s.recv(2)
if cmd == 'end':
s.close()
c.close()
break
client:
import socket
s = socket.socket()
s.connect(('192.168.1.2', 1234))
def get_file(s, file_name):
cmd = 'get\n%s\n' % (file_name)
s.sendall(cmd)
r = s.recv(2)
size = int(s.recv(16))
recvd = ''
while size > len(recvd):
data = s.recv(1024)
if not data:
break
recvd += data
s.sendall('ok')
return recvd
print get_file(s, 'file1')
print get_file(s, 'file2')
s.sendall('end\n')

Is sending 'ok' back and forth essentially as stop bits to break up
streams of data a reasonable technique?
Most protocols use some terminator or another. Popular alternatives are '\r\n', '\r\n\r\n' or EOF (ctrl+d), but these are just arbitrarily chosen and no worse or better than your 'ok', as long as your client and server know how to handle it.

Your code looks good.
You don't actually need to send across the size of the file. You can use while True, as the check if not data: break will stop the loop.
while True:
data = s.recv(1024)
if not data: print " Done "; break
recvd += data
Also, why are you sending 'ok' is the other side doesn't check for it? You are just skipping 2 bytes at each side.
Don't you need to cater to multiple clients? No need for multi-threading?

Is there a standard way to do this?
Yes. http://www.faqs.org/rfcs/rfc959.html
Describes the standard way to do this.
Here is an implementation: http://docs.python.org/library/ftplib.html

U may look at this implementation. It also take care of if the file is in a sub-directory. Here is the link!
server
import socket
import os
print('Waiting for clinet to connect...')
c = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
c.bind(('', 1234))
c.listen(1)
s, a = c.accept()
print('Connected. Going to receive file.')
s.sendall('getfilename')
filename = s.recv(1024)
if '/' in filename:
dir = os.path.dirname(filename)
try:
os.stat(dir)
except:
print('Directory does not exist. Creating directory.')
os.mkdir(dir)
f = open(filename, 'wb')
print('Filename: ' + filename)
while True:
s.sendall('getfile')
size = int(s.recv(16))
print('Total size: ' + str(size))
recvd = ''
while size > len(recvd):
data = s.recv(1024)
if not data:
break
recvd += data
f.write(data)
#print(len(recvd))
break
s.sendall('end')
print('File received.')
s.close()
c.close()
f.close()
client
import socket
import sys
if len(sys.argv) > 1 :
print('Trying to connect...')
s = socket.socket()
s.connect(('127.0.0.1', 1234))
print('Connected. Wating for command.')
while True:
cmd = s.recv(32)
if cmd == 'getfilename':
print('"getfilename" command received.')
s.sendall(sys.argv[1])
if cmd == 'getfile':
print('"getfile" command received. Going to send file.')
with open(sys.argv[1], 'rb') as f:
data = f.read()
s.sendall('%16d' % len(data))
s.sendall(data)
print('File transmission done.')
if cmd == 'end':
print('"end" command received. Teminate.')
break

rsync is the standard way to sync files between two computers. You could write it in Python like this http://code.activestate.com/recipes/577518-rsync-algorithm/ or you could wrap the C library like this http://freshmeat.net/projects/pysync/ with some tweaks like replacing MD4 with MD5.
Or, if you want to do this at the socket level, you really should be using asynchat with asyncore. Here is an FTP server written with asynchat http://pyftpdlib.googlecode.com/svn-history/r20/trunk/pyftpdlib/FTPServer.py but you should start by reading http://www.doughellmann.com/PyMOTW/asynchat/ Pay attention to the part about Message Terminators point 2. A lot of network protocols do odd stuff like this, i.e. sometimes they send and receive full line commands and responses, and sometimes they send and receive chunks of arbitrary data preceded by the count of how many bytes are in the chunk. You can handle this much more easily with asynchat, and your program will scale much better too.