#!/usr/bin/python3
#
# Skeleton proxy server (in Python 3) for CS 125
#
"""Web proxy server.  Intercepts Web requests, caches them, and
optionally modifies them.

Usage:
    proxyserver.py [-d] SERVER-IP PORT

Options:
    -d  Enable debugging messages

SERVER-IP is the DNS name or IP address to listen on; usually
"localhost" will do just fine.

PORT is the port to listen on; usually the UID printed by the "id"
command will be appropriate."""

import docopt
import hashlib
from socket import *
import sys

args = None
debug = False
port = None

def main():
    global args
    global debug
    args = docopt.docopt(__doc__, version = 'proxyserver version 0.1')
    #
    # args['SERVER-IP'] contains the DNS name or IP address to bind to.
    # args['PORT'] contains the port number (as a string).
    #
    if args['-d']:
        debug = True

    #
    # Invoke the proxy
    #
    proxy()

def proxy():
    #
    # Convert the port number to an integer.
    #
    port = int(args['PORT'])
    # Create a server socket, bind it to a port and start listening
    tcpServerSocket = socket(AF_INET, SOCK_STREAM)
    # The following line allows you to rerun the server right away after killing it.
    # That’s useful for testing; otherwise you may have to wait 15-30 seconds each time.
    tcpServerSocket.setsockopt(SOL_SOCKET, SO_REUSEADDR, 1)
    # FILL IN HERE
    while True:
        # Start receiving data from the client
        if debug:
            print('Ready to serve...')
        tcpClientSocket, clientAddress = tcpServerSocket.accept()
        print('Received a connection from:', clientAddress)
        clientMessage = ''# FILL IN HERE
        if debug:
            print('Request:', clientMessage)
        #
        # Extract the filename from the given message.  Note that we
        # are very sloppy about error checking; this should be
        # improved.  However, this will work for syntactically correct
        # requests and thus is handy for initial development.  Error
        # handling can be added later.
        #
        # FILL IN HERE
        url = clientMessage.split()[1]
        if debug:
            print('Parsed URL:', url)
        #
        # Again, we are sloppy about parsing the requested URL, and
        # that should be improved.
        #
        # FILL IN HERE
        head, slash, filename = url.partition('/')
        if debug:
            print('Parsed filename:', filename)
        fileExists = False
        #
        # Figure out a name to use in the cache.  Note that it is both
        # dangerous and incorrect to simply use the name requested by
        # the client.  The name provided might include strange
        # characters that could break your file system, and if the
        # client requests a file with the same name from two different
        # Web servers, the proxy will serve up the cached file from
        # the first server without noticing the problem.  (This is
        # especially troublesome for common names like "index.html".)
        # Correcting this flaw is left as an exercise for the student
        # (it can be useful to calculate a SHA-1 hash [see
        # hashlib.sha1] of the entire URL and use that as the basis
        # for a file name).
        #
        # FILL IN HERE
        cachedName = "/" + filename
        if debug:
            print('Cached name:', cachedName)
        try:
            # Check whether the file exists in the cache
            cachedFile = open(cachedName[1:], "rb")
            #
            # It is incorrect to use read() here; if the file is
            # enormous the proxy server will die trying to swallow it.
            # However, this code will work for small files.
            #
            # FILL IN HERE
            outputdata = cachedFile.read()
            fileExists = True
            #
            # We found a cache hit, so we can generate a response message.
            # However, the code below is incorrect because it always
            # assumes that the result is text/html.  A correct
            # implementation will record the content-type of the
            # cached data and pass that on to the client.
            #
            # FILL IN HERE
            tcpClientSocket.send("HTTP/1.0 200 OK\r\n.encode()")
            tcpClientSocket.send("Content-Type: text/html\r\n".encode())
            # FILL IN HERE
            if debug:
                print('Successfully served data from cache')
            cachedFile.close()
        # Error handling for file not found in cache
        except IOError:
            if not fileExists:
                # Create a socket to connect to the main Web server
                socketToServer = ''# FILL IN HERE
                try:
                    # Connect to the server on port 80
                    # FILL IN HERE
                    socketToServer.send(
                      "GET http://" + filename + " HTTP/1.0\n\n".encode())
                    # Read the server's response
                    response = ''# FILL IN HERE
                    # Create a new file in the cache for the requested file.
                    # See above regarding file naming.
                    # FILL IN HERE
                    tmpFile = open("./" + filename, "wb")
                    # Also send the response the client socket.
                    # FILL IN HERE
                except:
                    # We couldn't connect to the server.  Generate an HTTP 500.
                    # FILL IN HERE
                    pass
                #
                # All done! Close the server socket
                #
                socketToServer.close()
            else:
                #
                # We failed to read from the cache; generate an HTTP 500
                # (internal server error).
                #
                # FILL IN HERE
                pass
        # Close the client socket
        tcpClientSocket.close()

if __name__ == '__main__':
    main()