to see the sample agents duke it out.
This is a brief overview of the included files and directories:
- - SampleAgents.py : the sample agents. This is the bit you care about!
+ - USINGPYTHON.txt: A guide to writing an agent in Python.
+ - SampleAgents.py : the sample agents.
+ - agents/ : where you put your agent.
- arenas/ : arenas define which agents play against which, and some other parameters.
- link/ : Other languages belong here.
- conf.py : has various useful things like the ability to turn on debug support.
--- /dev/null
+USINGPYTHON.txt: A 10 step guide to writing an agent in Python.
+
+1. Pick a name for your agent. Make sure the name is a valid python identifier. Be original.
+
+2. Open SampleAgents.py. Pick a sample agent to copy-paste (so you don't have to type out the function definitions).
+
+3. Copy the sample agent code into a new file in the "agents" directory. Make sure the file has the same name as the agent.
+
+4. Modify the code to do your bidding.
+
+4a. If you wish to implement your own leaning agent based on BaseAgent, rather than use the limited functionality of LearningAgent, insert the following code snippet inside your class:
+
+ def Results (self, foeName, wasAttacker, winner, attItem, defItem, bluffItem, pointDelta):
+ BaseAgent.Results (self, foeName, wasAttacker, winner, attItem, defItem, bluffItem, pointDelta)
+ # your own code goes here
+
+
+5. Create an arena in which your agent can battle:
+
+ 5.1. open arenas/MyArena.py
+
+ 5.2. add "from agents.<youragentname> import <youragentname>"
+
+ 5.3. modify the "Agents =" line to include your agent, and take out any agents you don't want to battle.
+
+6. Watch your agent in action: ./simulate -v -a MyArena
+
+7. Oh no, my agent dies very quickly: what's going on?
+
+ 7.1 insert print statements in your python module. (Hint: prefix them with self.id so you can identify different agents)
+
+ 7.2 Run "./simulate -v -n 1 -a MyArena" to start with only 1 of each agent.
+
+ 7.3 Change the agents against which you're battling in MyArena.py so that you have a predictable opponent.
+
+ 7.1 Edit conf.py, and set DEBUG=True. Don't forget to reset it when you're done.
+
+8. Once your agent works to your satisfaction, try it both on short and long durations (100 and 1000 rounds: see MAX_ITERATIONS in conf.py)
+
+9. If the rolling scoreboard has been opened on progcomp.ucc.asn.au/, submit it there! Otherwise, sit tight.
+
+10. Watch its progress on the scoreboard and adjust your strategy accordingly.
\ No newline at end of file
--- /dev/null
+#!/usr/bin/python2.5
+'''PythonSampleAgents.py
+An arena that runs the sample bots in python.
+Written by Daniel Axtens <
[email protected]> for the UCC Programming Competition in 2010.
+
+Licensed under an MIT-style license: see the LICENSE file for details.
+'''
+
+from SampleAgents import Angel, Lucifer, Dummy, Wash, Streetfighter
+
+# Import your agent here:
+# e.g. from agents.<youragentname> import <youragentname>
+
+class arena:
+ # Add you agent here:
+ # e.g. Agents=[<youragentname,Angel,..]
+ Agents = [Angel,Lucifer,Streetfighter,Wash]