Distributed Systems - Tutorial 1

badwords.py

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

import socket
import threading

class BadWordsServer:
    """Represent a server which is designed to filter "bad words"
       from data which is send to the network, and then resend
       the cleaned data to the client."""

    def __init__(self, bad_words, host='127.0.0.1', port=8000):
        """bad_words: words to ban
           host: host to use for the server
           port: port to use for the server"""

        self.bad_words = bad_words
        self.bad_replace = ['*' * len(x) for x in bad_words]

        self.s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        self.s.bind((host, port))
        self.s.listen(1)

    def serve(self, buf_size=1024, encoding='UTF-8'):
        """Run the server. It can handle multiple connection concurrently.

           buf_size: size of the buffer to use for recieving data
           encoding: encoding of input/output data"""

        while True:
            conn, _ = self.s.accept()
            threading.Thread(None, self.serve_one, None,\
                             (conn, buf_size, encoding)).start()

    def serve_one(self, conn, buf_size=1024, encoding='UTF-8'):
        """Serve only one connection. Close it recieve an empty data."""

        while True:
            data = conn.recv(buf_size).decode(encoding)

            if not data:
                break

            print('Raw data:', data)
            data = self.filter_bad_words(data)
            print('Cleaned data:', data)
            conn.send(bytes(data, encoding))

        conn.close()

    def filter_bad_words(self, data):
        """Remove the "bad words" from data, and return the cleaned data."""

        for w, r in zip(self.bad_words, self.bad_replace):
            data = data.replace(w, r)

        return data

class BadWordsClient:
    """Represent a client which ask user for input, and which send that
       data to a BadWordsServer for cleaning it. Once the client recieve
       the cleaned data it prints it."""

    def __init__(self, host='127.0.0.1', port=8000):
        """host: host of the server
           port: port of the server"""

        self.s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
        self.s.connect((host, port))

    def run(self, buf_size=1024, encoding='UTF-8'):
        """Ask the user for data, clean it through the BadWordsServer,
           and then print the result.

           buf_size: size of the buffer to use for recieving data
           encoding: encoding of input/output data"""

        while True:
            data = input('Data to filter? ')

            if not data:
                break

            self.s.send(bytes(data, encoding))

            data = self.s.recv(buf_size).decode(encoding)
            print('Cleaned data:', data)

        self.s.close()

if __name__ == '__main__':
    import sys

    def yes_no(prompt, default_yes=True):
        """Ask the user for a yes/no question.

           prompt: the question to ask
           default_yes: the default choice is yes if True no otherwise

           Return True for yes, False for no"""

        y, n = 'y', 'n'

        if default_yes:
            y = y.upper()
        else:
            n = n.upper()

        i = None

        while i not in ['', 'y', 'n']:
            i = input('%s (%c/%c)? ' % (prompt, y, n)).lower()

        if i == '':
            return default_yes
        elif i == 'y':
            return True
        else:
            return False

    def get_host_port():
        """Ask the user for an host and a port, and return them."""

        host = input('host? ')
        port = None

        while port == None:
            try:
                port = int(input('port? '))
            except ValueError:
                print('port should be an integer.', file=sys.stderr)

        return host, port

    arg_error = 'This program takes exactly one argument serve or client.'

    if len(sys.argv) != 2:
        print(arg_error, file=sys.stderr)
    elif sys.argv[1] == 'serve':
        bad_words = ['fuck', 'shit', 'dumb']

        if yes_no('Would you like to serve on localhost:8000'):
            host, port = 'localhost', 8000
        else:
            host, port = get_host_port()

        try:
            BadWordsServer(bad_words, host, port).serve()
        except socket.error as e:
            print('Error occured:', e, file=sys.stderr)
    elif sys.argv[1] == 'client':
        if yes_no('Would you like to use localhost:8000 as server'):
            host, port = 'localhost', 8000
        else:
            host, port = get_host_port()

        try:
            BadWordsClient(host, port).run()
        except socket.error as e:
            print('Error occured:', e, file=sys.stderr)
    else:
        print(arg_error, file=sys.stderr)

# Q4:

# One way of passing more information in a message, like encoding, language
# or degree of cleaning would be to use "headers" in the message, the first
# one would be the encoding, followed by (for instance) a semicolon, then
# the language followed by a semicolon, then the degree of cleaning followed
# by a semicolon and finally the message.

# e.g.
# UTF-8;french;moderate;<the actual message>

# the encoding and first semicolon should be in UTF-8 to be sure that it
# will be well undestand by the server, all the following data in the
# message (including the two remaining semicolons) should be encoded
# with the "new" encoding.