Upload folder using huggingface_hub

server/finenv_environment.py

@@ -73,7 +73,7 @@ class FinenvEnvironment(Environment):
 
         return FinenvObservation(
             stock=self.stock,
-            market=self.market.value,
+            market=self.market.value if self.market else None,
             price=self.stock_price,
             shares=self.shares,
             cash=self.cash,
@@ -96,12 +96,17 @@ class FinenvEnvironment(Environment):
 
         
         if action.type == "init":
+            if not action.stock or not action.market:
+                raise ValueError("Stock and market required for init")
+            
             self.stock = action.stock
-            self.market = StockExchangeMarket(action.market.value.lower())
+            self.market = StockExchangeMarket(action.market.lower())
+
+            
 
             self.initial_cash = action.initial_cash or 10000.0
             self.max_steps = action.max_steps or 100
-            self.task = action.task_type.value if action.task_type else "easy"
+            self.task = action.task_type.lower() if action.task_type else "easy"
 
             self.cash = self.initial_cash
             self.shares = 0
@@ -162,7 +167,7 @@ class FinenvEnvironment(Environment):
         new_value = self.cash + self.shares * self.stock_price
         reward = (new_value - old_value) / self.initial_cash
         reward = max(min(reward, 1.0), -1.0)
-
+        reward=float(reward)
        
         done = self._state.step_count >= self.max_steps
         if done:
@@ -170,15 +175,16 @@ class FinenvEnvironment(Environment):
             profit = final_value - self.initial_cash
 
             if self.task == "easy":
-                score = min(max(profit / 0.01*self.initial_cash, 0), 1)
+                score = min(max(profit / (0.01 * self.initial_cash), 0), 1)
 
             elif self.task == "medium":
-                score = min(max(profit / 0.05*self.initial_cash, 0), 1)
+                score = min(max(profit / (0.05 * self.initial_cash), 0), 1)
 
             elif self.task == "hard":
                 score = min(max((profit / self.initial_cash) / 0.1, 0), 1)
 
             reward = score
+            reward=float(reward)
         
         return FinenvObservation(
             stock=self.stock,