Upload folder using huggingface_hub

inference.py

@@ -1,55 +1,33 @@
 import os
-import sys
 from openai import OpenAI
 from client import SupportEnvClient, SupportAction
 
-# Hackathon defined variables
 API_BASE_URL = os.getenv("API_BASE_URL", "https://router.huggingface.co/v1")
 MODEL_NAME = os.getenv("MODEL_NAME", "Qwen/Qwen2.5-72B-Instruct")
-API_KEY = os.getenv("HF_TOKEN") or os.getenv("API_KEY")
-BENCHMARK = "support_env"
+API_KEY = os.getenv("HF_TOKEN") or os.getenv("OPENAI_API_KEY")
+ENV_URL = os.getenv("ENV_URL", "https://swapnilpatil28-support-env.hf.space")
 
 client = OpenAI(base_url=API_BASE_URL, api_key=API_KEY)
 
 def run_task(task_name: str):
-    env_url = os.getenv("ENV_URL", "https://swapnilpatil28-support-env.hf.space")
-    env = SupportEnvClient(base_url=env_url).sync()
+    env = SupportEnvClient(base_url=ENV_URL).sync()
+    obs = env.reset(task_name=task_name)
+    print(f"[START] task={task_name} env=support_env model={MODEL_NAME}", flush=True)
     
-    try:
-        result = env.reset(task_name=task_name)
-        print(f"[START] task={task_name} env={BENCHMARK} model={MODEL_NAME}", flush=True)
-
-        step = 1
-        rewards = []
-
-        while not result.done:
-            prompt = f"Ticket: '{result.observation.ticket_text}'. Reply ONLY with one word: Billing, Tech, or Sales."
-            try:
-                completion = client.chat.completions.create(
-                    model=MODEL_NAME,
-                    messages=[{"role": "user", "content": prompt}],
-                    max_tokens=10
-                )
-                action_str = completion.choices[0].message.content.strip().replace('.', '')
-            except Exception as e:
-                action_str = "Tech" # Fallback
-
-            result = env.step(SupportAction(department=action_str))
-            reward_val = float(result.reward or 0.0)
-            rewards.append(reward_val)
-
-            print(f"[STEP] step={step} action={action_str} reward={reward_val:.2f} done={str(result.done).lower()} error=null", flush=True)
-            step += 1
-            if step > 10: break # Safety break
-
-        score = sum(rewards)
-        success = "true" if score >= 0.99 else "false"
-        rewards_str = ",".join([f"{r:.2f}" for r in rewards])
-
-        print(f"[END] success={success} steps={step-1} score={score:.2f} rewards={rewards_str}", flush=True)
-    finally:
-        env.close()
+    step, rewards = 1, []
+    while not obs.done:
+        prompt = f"Ticket: {obs.observation.content}. Details: {obs.observation.search_result or 'None'}. Output one word: Billing, Tech, or Sales."
+        completion = client.chat.completions.create(model=MODEL_NAME, messages=[{"role": "user", "content": prompt}], max_tokens=10)
+        dept = completion.choices[0].message.content.strip().strip('.')
+        
+        res = env.step(SupportAction(action_type="route", department=dept))
+        r = float(res.reward or 0.0)
+        rewards.append(r)
+        print(f"[STEP] step={step} action={dept} reward={r:.2f} done={str(res.done).lower()} error=null", flush=True)
+        obs, step = res, step + 1
+
+    print(f"[END] success={str(sum(rewards)>0).lower()} steps={step-1} score={sum(rewards)/len(rewards):.2f} rewards={','.join([f'{r:.2f}' for r in rewards])}", flush=True)
+    env.close()
 
 if __name__ == "__main__":
-    for task in ["easy", "medium", "hard"]:
-        run_task(task)
+    for task in ["easy", "medium", "hard"]: run_task(task)

models.py

@@ -1,15 +1,17 @@
-from typing import List, Optional, Dict
+from typing import List, Optional, Literal
 from openenv.core.env_server import Action, Observation, State
+from pydantic import Field
 
 class SupportAction(Action):
-    department: str
+    action_type: Literal["search", "route"] = Field(..., description="Action to take")
+    department: Optional[str] = Field(None, description="Required for route action")
 
 class SupportObservation(Observation):
     ticket_id: str
-    ticket_text: str
-    message: str
+    content: str
+    search_result: Optional[str] = None
+    available_departments: List[str] = ["Billing", "Tech", "Sales"]
 
 class SupportState(State):
-    task_name: str = "easy"
-    total_tickets: int = 1
-    tickets_routed: int = 0
+    task_id: str = "easy"
+    current_ticket_index: int = 0

server/app.py

@@ -3,7 +3,6 @@ from models import SupportAction, SupportObservation
 from server.environment import SupportEnvironment
 import uvicorn
 
-# OpenEnv factory: Pass the CLASS
 app = create_fastapi_app(SupportEnvironment, SupportAction, SupportObservation)
 
 def main():

server/environment.py

@@ -1,56 +1,46 @@
 import uuid
 from typing import List, Dict
-from openenv.core.env_server import Observation, Environment
+from openenv.core.env_server import Environment
 from models import SupportAction, SupportObservation, SupportState
 
 class SupportEnvironment(Environment):
     def __init__(self):
         super().__init__()
-        self._state = SupportState()
-        self._current_queue: List[Dict[str, str]] = []
-        self._current_ticket = None
+        self._db = {
+            "T1": "Customer History: Frequent refunder. Status: Valid invoice.",
+            "T4": "System Log: Payment gateway timed out. Card ending in 4242.",
+            "T6": "Dev Log: 500 error on /api/auth. Cluster: us-east-1."
+        }
         self.tasks = {
-            "easy": [{"id": "T1", "text": "Can I get a refund for my last invoice?", "dept": "Billing"}],
-            "medium": [{"id": "T2", "text": "My screen is flickering.", "dept": "Tech"},
-                       {"id": "T3", "text": "I want to upgrade.", "dept": "Sales"}],
-            "hard": [{"id": "T4", "text": "Update credit card?", "dept": "Billing"},
-                     {"id": "T5", "text": "Enterprise discount?", "dept": "Sales"},
-                     {"id": "T6", "text": "API 500 error on POST.", "dept": "Tech"}]
+            "easy": [{"id": "T1", "text": "Refund status?", "dept": "Billing"}],
+            "medium": [{"id": "T2", "text": "Upgrade account?", "dept": "Sales"}, 
+                       {"id": "T3", "text": "App keeps crashing.", "dept": "Tech"}],
+            "hard": [{"id": "T4", "text": "Payment failed!", "dept": "Billing"},
+                     {"id": "T5", "text": "Bulk pricing?", "dept": "Sales"},
+                     {"id": "T6", "text": "API Auth failure.", "dept": "Tech"}]
         }
 
     def reset(self, task_name: str = "easy") -> SupportObservation:
-        if task_name not in self.tasks:
-            task_name = "easy"
-        self._current_queue = self.tasks[task_name].copy()
-        self._state = SupportState(
-            episode_id=str(uuid.uuid4()), step_count=0, task_name=task_name,
-            total_tickets=len(self._current_queue), tickets_routed=0
-        )
-        self._current_ticket = self._current_queue.pop(0)
-        return SupportObservation(
-            done=False, reward=0.0, ticket_id=self._current_ticket["id"],
-            ticket_text=self._current_ticket["text"], message="New ticket arrived."
-        )
+        self.current_task = self.tasks.get(task_name, self.tasks["easy"])
+        self._state = SupportState(episode_id=str(uuid.uuid4()), task_id=task_name)
+        t = self.current_task[0]
+        return SupportObservation(done=False, reward=0.0, ticket_id=t["id"], content=t["text"])
 
     def step(self, action: SupportAction) -> SupportObservation:
         self._state.step_count += 1
-        is_correct = action.department.strip().lower() == self._current_ticket["dept"].lower()
+        ticket = self.current_task[self._state.current_ticket_index]
         
-        # Normalized reward: 1.0 / total_tickets for each correct step
-        reward = (1.0 / self._state.total_tickets) if is_correct else 0.0
-        self._state.tickets_routed += 1
+        if action.action_type == "search":
+            return SupportObservation(done=False, reward=-0.05, ticket_id=ticket["id"], content=ticket["text"], search_result=self._db.get(ticket["id"], "No record."))
 
-        if len(self._current_queue) > 0:
-            self._current_ticket = self._current_queue.pop(0)
-            return SupportObservation(
-                done=False, reward=reward, ticket_id=self._current_ticket["id"],
-                ticket_text=self._current_ticket["text"], message=f"Routed. Correct: {is_correct}"
-            )
-        else:
-            return SupportObservation(
-                done=True, reward=reward, ticket_id="None",
-                ticket_text="Queue is empty.", message="All tickets processed."
-            )
+        correct = action.department.strip().lower() == ticket["dept"].lower()
+        reward = 1.0 if correct else 0.0
+        self._state.current_ticket_index += 1
+        
+        if self._state.current_ticket_index < len(self.current_task):
+            t = self.current_task[self._state.current_ticket_index]
+            return SupportObservation(done=False, reward=reward, ticket_id=t["id"], content=t["text"])
+        return SupportObservation(done=True, reward=reward, ticket_id="EOF", content="Done.")
 
     @property
     def state(self) -> SupportState: