change create_agent interface

agent/__init__.py

@@ -1 +1 @@
-from .util import create_create_agent
+from .util import create_agent_helper

agent/util.py

@@ -1,12 +1,12 @@
 from recsim.agents import full_slate_q_agent
 
-def create_create_agent(agent=full_slate_q_agent.FullSlateQAgent):
+def create_agent_helper(agent=full_slate_q_agent.FullSlateQAgent, **kwargs):
     def create_agent(sess, environment, eval_mode, summary_writer=None):
-        kwargs = {
-            'observation_space': environment.observation_space,
-            'action_space': environment.action_space,
-            'summary_writer': summary_writer,
-            'eval_mode': eval_mode,
-        }
+        print(f"using {agent.__name__}")
+        kwargs['observation_space'] = environment.observation_space
+        kwargs['action_space'] = environment.action_space
+        kwargs['summary_writer'] = summary_writer
+        kwargs['eval_mode'] = eval_mode
+
         return agent(sess, **kwargs)
     return create_agent

main.py

@@ -5,7 +5,7 @@ from document import FlashcardDocumentSampler
 from recsim.simulator import recsim_gym
 from recsim.agents import full_slate_q_agent
 from recsim.simulator import runner_lib
-from agent import create_create_agent
+from agent import create_agent_helper
 from util import reward, update_metrics
 
 slate_size = 1
@@ -14,7 +14,7 @@ time_budget = 60
 
 tf.compat.v1.disable_eager_execution()
 
-create_agent_fn = create_create_agent(full_slate_q_agent.FullSlateQAgent)
+create_agent_fn = create_agent_helper(full_slate_q_agent.FullSlateQAgent)
 
 ltsenv = environment.Environment(
   FlashcardUserModel(num_candidates, time_budget, slate_size),
@@ -31,7 +31,7 @@ runner = runner_lib.TrainRunner(
     base_dir=tmp_base_dir,
     create_agent_fn=create_agent_fn,
     env=lts_gym_env,
-    episode_log_file="",
+    episode_log_file="episode.log",
     max_training_steps=5,
     num_iterations=1
 )