Update server/app.py

server/app.py

@@ -1,96 +1,43 @@
 from fastapi import FastAPI
-from env import EmailEnv
-import os
-from openai import OpenAI
-import threading
-import time
-import requests
+from pydantic import BaseModel
+import random
 
 app = FastAPI()
-env = EmailEnv("hard")
 
+emails = [
+    {"email": "Refund my order", "label": "support"},
+    {"email": "Interested in pricing", "label": "sales"},
+    {"email": "Bug in product", "label": "support"},
+    {"email": "Partnership request", "label": "business"}
+]
 
-def make_llm_call(label: str = "LLM"):
-    base = os.environ.get("API_BASE_URL")
-    key = os.environ.get("API_KEY")
+state = {}
+done = False
 
-    if not base or not key:
-        print(f"[{label}] API_BASE_URL or API_KEY not set")
-        return None
-
-    try:
-        client = OpenAI(base_url=base, api_key=key)
-        response = client.chat.completions.create(
-            model="gpt-4o-mini",
-            messages=[{"role": "user", "content": "Reply OK"}]
-        )
-        print(f"[{label}] Call succeeded: {response.choices[0].message.content}")
-        return response
-    except Exception as e:
-        print(f"[{label}] Call failed: {e}")
-        return None
-
-
-def startup_llm_trigger():
-    time.sleep(3)
-    print("ENV CHECK →", {
-        "API_BASE_URL": os.environ.get("API_BASE_URL", "NOT SET"),
-        "API_KEY": "SET" if os.environ.get("API_KEY") else "NOT SET"
-    })
-    make_llm_call("startup-trigger")
-
-@app.on_event("startup")
-def startup():
-    # Fire startup LLM call in background after server is ready
-    threading.Thread(target=startup_llm_trigger, daemon=True).start()
+class Action(BaseModel):
+    action: str
 
+@app.get("/")
+def root():
+    return {"status": "ok"}
 
 @app.post("/reset")
 def reset():
-    state = env.reset()
+    global state, done
+    state = random.choice(emails)
+    done = False
     return {"state": state}
 
-
 @app.post("/step")
-def step(action: dict):
-    act = action.get("action")
-    state, reward, done, info = env.step(act)
+def step(action: Action):
+    global state, done
+
+    correct = action.action == state["label"]
+    reward = 1.0 if correct else 0.0
+    done = True
+
     return {
         "state": state,
         "reward": reward,
-        "done": done,
-        "info": info
-    }
-
-
-@app.get("/")
-def root():
-    base = os.environ.get("API_BASE_URL")
-    key = os.environ.get("API_KEY") or os.environ.get("HF_TOKEN")
-
-    print(f"API_BASE_URL: '{base}'")
-    print(f"API_KEY present: {bool(key)}")
-
-    if not key:
-        print("Missing API_KEY")
-        return {"message": "running", "llm": "unavailable"}
-
-    try:
-        client = OpenAI(base_url=base, api_key=key)  # base can be None, OpenAI handles it
-        response = client.chat.completions.create(
-            model=os.environ.get("MODEL_NAME", "gpt-4o-mini"),
-            messages=[{"role": "user", "content": "Reply OK"}]
-        )
-        print("LLM call succeeded:", response.choices[0].message.content)
-        return {"message": "running", "llm": response.choices[0].message.content}
-
-    except BaseException as e:
-        print("LLM call failed:", repr(e))
-        return {"message": "running", "llm": "unavailable"}
-def main():
-    import uvicorn
-    uvicorn.run("server:app", host="0.0.0.0", port=7860)
-
-
-if __name__ == "__main__":
-    main()
+        "done": done
+    }