import math
import numpy
import lab5_wnet
import matplotlib.pyplot as p

# Stuff about each packet in the simulator, to keep track of start and end 
# times, identity of sender, whether collision happened or not.
class Packet:
    def __init__(self,starttime,sender,receiver=None,ptime=1):
	self.start = starttime
	self.sender = sender
        self.receiver = receiver
        self.ptime = ptime  # size of packet in time slots
	self.end = -1		# will get initialized later when xmitting
        if receiver == None:
            self.coll_flag = lab5_wnet.NO_COLLISION

#
# Various performance statistics that we want to maintain and possibly plot.
# Attach one of these objects to each node, and also to the entire network.
#
class Stats:
    def __init__(self,simtime):
	self.reset(simtime)

    def reset(self,simtime):
	self.simtime = simtime	# total simulation time
	self.attempts = 0	# number of attempts by network or node
	self.success = 0	# number that succeeded
	self.collisions = 0	# number that failed (collided)
	self.latency = 0     # average latency; most useful for a node
	self.backoffs = 0    # number of backoffs; generally useful at a node
        self.pending = 0     # only makes much sense at a node
	self.numbackoffs = 0	# number of backoffs at node
        # downlink stats for a node
        self.downrecd = 0       # number of packets received on downlink to node
        self.downq = 0          # current length of downlink queue to a node
        self.plist = []

    # _print() is a bit hacky.  we have one kind of Stats object that
    # we attach to both the network and each node.  The info we want printed
    # depends on whether we're printing it from the network or a node.  When 
    # it's a node, the "type" is the node ID, otherwise, it's 'net'
    def _print(self,time,ptime,type):
	if time == 0: u = 0
	else: u = (1.0*self.success*ptime)/time
	if type == 'net':
#	    print "Time %d attempts %d success %d coll %d util %.2f downloaded %d" % (time,self.attempts,self.success,self.collisions,u,self.downrecd)
	    print "Time %d attempts %d success %d util %.2f" % (time,self.attempts,self.success,u)
	else:
#	    print "  Node %d attempts %d success %d coll %d lat %d backoffs %d downrecd %d" % (int(type), self.attempts, self.success, self.collisions, self.latency, self.numbackoffs, self.downrecd)
#	    print "  Node %d attempts %d success %d coll %d lat %d" % (int(type), self.attempts, self.success, self.collisions, self.latency)
	    print "  Node %d attempts %d success %d coll %d" % (int(type), self.attempts, self.success, self.collisions)

##########################################################################
# Plot scatter plot of successful transmissions and collisions.
# Plot bar graph of per-node throughput (# successful transmissions
##########################################################################
def plot_data(wnet):
        succ = []
        x = []
        y = []
        xcoll = []
        ycoll = []
        for node in wnet.nlist:
            succ.append(node.stats.success)
            y = y + [node.get_id()] * len(node.sent)
            x = x + node.sent
            ycoll = ycoll + [node.get_id()] * len(node.coll)
            xcoll = xcoll + node.coll
        p.subplots_adjust(hspace = 0.1)
        p.subplot(2,1,1)
        p.ylabel('Node (blue=success; red=collision)')
        p.xlabel('Time sent')
        if x != [] and y != []:
            p.scatter(x, y, c='b')
        if xcoll != [] and ycoll != []:
            ycoll = numpy.array(ycoll)
            p.scatter(xcoll, ycoll-0.4, c='r')

        p.subplot(2,1,2)
        ind = numpy.arange(len(wnet.nlist))
        width = 0.35
        p.bar(ind, succ, width, color = 'g')
        p.ylabel('# successful receptions')
        p.xlabel('Node #')
        p.show()