update

app.py

@@ -1,15 +1,22 @@
-from fastapi import FastAPI
+from fastapi import Body, FastAPI
+from pydantic import BaseModel, ConfigDict
 from env.core import CacheEnv
 
 app = FastAPI()
 env = CacheEnv()
 
+
+class ResetBody(BaseModel):
+    model_config = ConfigDict(extra="ignore")
+    task_id: str | None = None
+
+
 @app.post("/reset")
-def reset():
-    state = env.reset()
+def reset(body: ResetBody = Body(default_factory=ResetBody)):
+    state = env.reset(task_id=body.task_id)
     return {
         "state": state,
-        "task_id": state.get("task_id")
+        "task_id": state.get("task_id"),
     }
 @app.post("/step")
 def step(action: dict):

cache_invalidation_env.egg-info/PKG-INFO

@@ -0,0 +1,12 @@
+Metadata-Version: 2.4
+Name: cache-invalidation-env
+Version: 0.1.0
+Summary: Cache invalidation decision environment for OpenEnv
+Requires-Python: >=3.10
+Requires-Dist: openenv-core[core]>=0.2.2
+Requires-Dist: fastapi>=0.100.0
+Requires-Dist: uvicorn[standard]>=0.22.0
+Requires-Dist: pydantic>=2.0.0
+Requires-Dist: requests>=2.28.0
+Requires-Dist: openai>=1.0.0
+Requires-Dist: python-dotenv>=1.0.0

cache_invalidation_env.egg-info/SOURCES.txt

@@ -0,0 +1,16 @@
+README.md
+pyproject.toml
+cache_invalidation_env.egg-info/PKG-INFO
+cache_invalidation_env.egg-info/SOURCES.txt
+cache_invalidation_env.egg-info/dependency_links.txt
+cache_invalidation_env.egg-info/entry_points.txt
+cache_invalidation_env.egg-info/requires.txt
+cache_invalidation_env.egg-info/top_level.txt
+env/__init__.py
+env/core.py
+env/generator.py
+env/grader.py
+env/models.py
+env/tasks.py
+server/__init__.py
+server/app.py

cache_invalidation_env.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

cache_invalidation_env.egg-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+server = server.app:main

cache_invalidation_env.egg-info/requires.txt

@@ -0,0 +1,7 @@
+openenv-core[core]>=0.2.2
+fastapi>=0.100.0
+uvicorn[standard]>=0.22.0
+pydantic>=2.0.0
+requests>=2.28.0
+openai>=1.0.0
+python-dotenv>=1.0.0

cache_invalidation_env.egg-info/top_level.txt

@@ -0,0 +1,2 @@
+env
+server

env/core.py

@@ -2,16 +2,17 @@ import random
 from env.generator import generate_env
 from env.grader import compute_step_reward
 from env.tasks import sample_task
-from env.grader import normalize_episode_score
-
 class CacheEnv:
 
     def __init__(self):
         self.reset()
 
-    def reset(self):
+    def reset(self, task_id=None):
         self.history = []
-        self.task_id = sample_task()
+        if task_id in ("easy", "medium", "hard"):
+            self.task_id = task_id
+        else:
+            self.task_id = sample_task()
         items, hidden, current_time = generate_env(self.task_id)
 
         self.state = {

env/grader.py

@@ -1,3 +1,11 @@
+# Submission validators require final scores strictly in (0, 1), not at the endpoints.
+_SCORE_EPS = 1e-4
+
+
+def clamp_strict_unit_interval(x: float) -> float:
+    return float(min(1.0 - _SCORE_EPS, max(_SCORE_EPS, x)))
+
+
 def compute_step_reward(action_type, is_stale):
     reward = 0
 
@@ -15,7 +23,7 @@ def compute_step_reward(action_type, is_stale):
 def normalize_episode_score(total_reward, max_steps=10):
     # expected max ≈ 1.0 per step
     score = total_reward / max_steps
-    return max(0.0, min(1.0, score))
+    return clamp_strict_unit_interval(max(0.0, min(1.0, score)))
 
 
 
@@ -31,7 +39,7 @@ def evaluate_episode(history):
     total_steps = len(history)
 
     if total_steps == 0:
-        return 0.0
+        return clamp_strict_unit_interval(0.0)
 
     correct_decisions = 0
     unnecessary_invalidations = 0
@@ -72,4 +80,4 @@ def evaluate_episode(history):
         0.2 * stability
     )
 
-    return max(0.0, min(1.0, score))
+    return clamp_strict_unit_interval(max(0.0, min(1.0, score)))

inference.py

@@ -8,6 +8,8 @@ from typing import List, Optional
 import requests
 from openai import OpenAI
 
+from env.grader import clamp_strict_unit_interval
+
 # Load .env from repo root so HF_TOKEN / API_BASE_URL work when you run: python inference.py
 try:
     from dotenv import load_dotenv
@@ -233,6 +235,7 @@ def run() -> None:
         success = False
         print(f"[RUN] fatal: {exc}", file=sys.stderr)
     finally:
+        episode_score = clamp_strict_unit_interval(episode_score)
         log_end(
             success=success,
             steps=steps_taken,

openenv.yaml

@@ -1,6 +1,11 @@
 name: cache_invalidation_env
 description: Decision-making environment for cache invalidation under uncertainty
 
+tasks:
+  - easy
+  - medium
+  - hard
+
 actions:
   type: object
   properties: