Upload folder using huggingface_hub

models.py

@@ -12,6 +12,7 @@ The textarena environment is a simple test environment that echoes back messages
 
 from __future__ import annotations
 
+from pydantic import Field
 from typing import Any, Dict, List, Optional
 
 from pydantic import BaseModel, Field
@@ -19,7 +20,7 @@ from pydantic import BaseModel, Field
 from openenv.core.env_server.types import Action, Observation, State
 
 
-class TextArenaMessage(BaseModel):
+class TextArenaMessage:
     """Single message observed by a player."""
 
     sender_id: int
@@ -42,7 +43,7 @@ class TextArenaObservation(Observation):
     legal_players: List[int] = Field(default_factory=list)
     info: Dict[str, Any] = Field(default_factory=dict)
 
-
+    
 class TextArenaState(State):
     """Structured state snapshot for the server."""
 
@@ -53,3 +54,4 @@ class TextArenaState(State):
     last_reward: float = 0.0
     last_info: Dict[str, Any] = Field(default_factory=dict)
     raw_state: Dict[str, Any] = Field(default_factory=dict)
+

rewards.py

@@ -5,7 +5,10 @@ from __future__ import annotations
 import re
 from typing import Dict, List, Protocol, Tuple
 
-from .models import TextArenaAction, TextArenaObservation
+try:
+    from textarena_env.models import TextArenaAction, TextArenaObservation
+except ImportError:
+    from models import TextArenaAction, TextArenaObservation
 
 
 class RewardProvider(Protocol):

server/app.py

@@ -56,7 +56,12 @@ def create_textarena_environment():
 
 # Create the FastAPI app
 # Pass the factory function instead of an instance for WebSocket session support
-app = create_app(create_textarena_environment, TextArenaAction, TextArenaObservation, env_name="textarena_env")
+app = create_app(
+    create_textarena_environment,
+    TextArenaAction,
+    TextArenaObservation,
+    env_name="textarena_env",
+)
 
 
 def main(host: str = "0.0.0.0", port: int = 8000):

server/environment.py

@@ -104,6 +104,11 @@ class TextArenaEnvironment(Environment):
         self._reward_providers: List[RewardProvider] = build_reward_providers(env_id)
         self._last_reward_signals: Dict[str, float] = {}
 
+        # Initialize environment state - TextArena envs require reset() to be called
+        # before step() can be used, as the internal state object isn't created until reset.
+        # This ensures the environment is always in a valid state after construction.
+        self._ta_env.reset(num_players=self.num_players)
+
     # ------------------------------------------------------------------
     # Environment interface
     # ------------------------------------------------------------------