Fix Pydantic typing error for Request

README.md

@@ -122,17 +122,17 @@ Prompt --> LLM generates action (JSON) --> Environment executes
 ### Reward vs Training Step
 Shows reward improvement over GRPO training iterations compared to random baseline.
 
-![Reward vs Episode](https://raw.githubusercontent.com/joshua400/FairRecovery-PlusPlus/plots-branch/plots/reward_vs_episode.png)
+`plots/reward_vs_episode.png`
 
 ### Fairness: Baseline vs Trained
 Demonstrates that the trained agent achieves better fairness scores (closer to 0) than the random baseline.
 
-![Fairness vs Episode](https://raw.githubusercontent.com/joshua400/FairRecovery-PlusPlus/plots-branch/plots/fairness_vs_episode.png)
+`plots/fairness_vs_episode.png`
 
 ### Component Rewards Breakdown
 Shows how R_exec (utility) and R_fair (fairness) improve after training.
 
-![Component Rewards](https://raw.githubusercontent.com/joshua400/FairRecovery-PlusPlus/plots-branch/plots/component_rewards.png)
+`plots/component_rewards.png`
 
 ---
 

server/app.py

@@ -7,6 +7,7 @@ Mirrors hallucination-detector-gym server/app.py pattern exactly.
 from __future__ import annotations
 import os, sys
 from typing import Optional
+from fastapi import FastAPI, Request
 
 sys.path.insert(0, os.path.dirname(os.path.dirname(__file__)))
 
@@ -24,7 +25,6 @@ configure_logging(json_output=True, log_level="INFO")
 
 
 def _build_app():
-    from fastapi import FastAPI, Request
     from fastapi.responses import JSONResponse, RedirectResponse
     import gradio as gr