rebasing, confirming RT is running before refactor

research/test_methodology/RT.py

@@ -16,12 +16,12 @@
     agent1 = TrainedAgent("../blueprint_algo/results_2020_05_10_21_36_47_291425")
     # sorta hacky, but I loaded the average strategy above, now I'm replacing with
     # the better strategy
-    offline_strategy = joblib.load('/Users/colin/Downloads/offline_strategy_285800.gz')
+    #offline_strategy = joblib.load('/Users/colin/Downloads/offline_strategy_285800.gz')
     #print(sys.getsizeof(offline_strategy))
     # agent1.offline_strategy = offline_strategy
     # print(sys.getsizeof(agent1.offline_strategy))
-    action_sequence = ["raise", "call", "call", "call", "call", "call", "call", "call", "call", "call", "call"]
-    agent_output = train(offline_strategy, public_cards, action_sequence, 200, 60, 60, 3, 2, 60)  # TODO: back to 50
+    action_sequence = ["raise", "call", "call", "call","call"]
+    agent_output = train(agent1.offline_strategy, public_cards, action_sequence, 200, 60, 60, 3, 2, 60)  # TODO: back to 50
     with open('realtime-strategy-off-better-1500.pkl', 'wb') as file:
         pickle.dump(agent_output, file)
     import ipdb;