X-Git-Url: https://git.ucc.asn.au/?p=progcomp10.git;a=blobdiff_plain;f=src%2FSampleAgents.py.old;fp=src%2FSampleAgents.py.old;h=d307214ac53e8e7e3a4677cd6b3126cec7980ae2;hp=0000000000000000000000000000000000000000;hb=e9a8105a8f22404f4ac550d79954eaa6b7f5d8ff;hpb=4597b27a44522a5c2785cc23029435f44f60ea55

diff --git a/src/SampleAgents.py.old b/src/SampleAgents.py.old
new file mode 100644
index 0000000..d307214
--- /dev/null
+++ b/src/SampleAgents.py.old
@@ -0,0 +1,89 @@
+'''SampleAgents.py - A collection of sample agents for playing Rock Paper Scissors.
+Written by Luke Williams <shmookey@ucc.asn.au> for the UCC Programming Competition in 2008.
+Requires Python 2.5.
+
+Licensed under an MIT-style license: see the LICENSE file for details.
+'''
+
+from uccProgComp import BaseAgent, LearningAgent, RandomAttack
+from rpsconst import *
+
+# Angel is a very simple bot that always tells the truth and expects others to do the same.
+class Dummy (BaseAgent):
+	def Attack (self, foe):
+		return Paper, Paper
+	def Defend (self, foe, bluff):
+		return bluff
+
+class Angel (BaseAgent):
+	def Attack (self, foe):
+		attack = RandomAttack ()	# Chooses randomly from Rock, Paper, Scissors
+		return attack, attack		# Tells the truth for its bluff.
+	def Defend (self, foe, bluff):
+		return bluff			# Trusts them to be going for a tie.
+
+# Lucifer here is the opposite. He always lies expecting people to be good and always goes for the kill.
+class Lucifer (BaseAgent):
+	def Attack (self, foe):
+		attack = RandomAttack ()
+		if attack == Rock: bluff = Scissors	# Here we choose the thing
+		elif attack == Paper: bluff = Rock	# that will hurt them
+		else: bluff = Paper			# if they go for a tie.
+		return attack, bluff
+	def Defend (self, foe, bluff):
+		if bluff == Rock: attack = Paper	# Here we trust that they
+		elif bluff == Paper: attack = Scissors	# are telling the truth.
+		else: attack = Rock			# And we try to kill them.
+		return attack
+#	def Results (self, foeName, wasAttacker, winner, attItem, defItem, bluffItem, pointDelta):
+#		BaseAgent.Results (self, foeName, wasAttacker, winner, attItem, defItem, bluffItem, pointDelta)
+#		print "I just scored " + str(pointDelta) + " points!"
+
+
+# Streetfighter assumes everyone has it in for him.
+class Streetfighter (BaseAgent):
+	def Attack (self, foe):
+		attack = RandomAttack ()
+		if attack == Rock: bluff = Paper	# Here we choose the thing
+		elif attack == Paper: bluff = Scissors	# that will hurt them
+		else: bluff = Rock			# if they go for a tie.
+		return attack, bluff
+	def Defend (self, foe, bluff):
+		if bluff == Rock: attack = Paper	# Here we trust that they
+		elif bluff == Paper: attack = Scissors	# are telling the truth.
+		else: attack = Rock			# And we try to kill them.
+		return attack
+
+# This is our first bot with any sort of learning capability, based on the LearningAgent base.
+# Experienced programmers might opt to write their own learning code based on BaseAgent, but it's up to you.
+# Frenchie is a simple bot that is by default nice but will permanently turn against any agent that betrays it.
+class Frenchie (LearningAgent):
+	def Attack (self, foe):
+		attack = RandomAttack ()
+		if Loss in LearningAgent.GetWinHistory (self, foe):
+			if attack == Rock: bluff = Scissors
+			elif attack == Paper: bluff = Rock
+			else: bluff = Paper
+		else:
+			bluff = attack
+		return attack, bluff
+	def Defend (self, foe, bluff):
+		if Loss in LearningAgent.GetWinHistory (self, foe):
+			if bluff == Rock: attack = Scissors	# They've fucked us in the past,
+			elif bluff == Paper: attack = Rock	# so we assume they're lying and
+			else: attack = Paper			# hoping we go for a kill.
+		else:
+			attack = bluff
+		return attack
+
+
+# If you want to implement your own Results () callback, you have to call the parent class's first:
+class Blank (BaseAgent):
+	def Attack (self, foe):
+		return Paper, Paper
+	def Defend (self, foe, bluff):
+		return bluff
+	def Results (self, foeName, wasAttacker, winner, attItem, defItem, bluffItem, pointDelta):
+		BaseAgent.Results (self, foeName, wasAttacker, winner, attItem, defItem, bluffItem, pointDelta)
+		# Now you can do your own thing
+