Update src/proxy_app/main.py

src/proxy_app/main.py

@@ -1515,6 +1515,57 @@ async def refresh_quota_stats(
         raise HTTPException(status_code=500, detail=str(e))
 
 
+# --- Log Viewing Endpoint ---
+@app.get("/logs/{log_type}")
+async def view_logs(
+    log_type: str,
+    lines: int = 50,
+    _: str = Depends(verify_api_key),
+):
+    """
+    View the last N lines of a log file.
+    Authenticated with PROXY_API_KEY.
+    
+    Path Params:
+      - log_type: 'proxy', 'debug', or 'failures'
+      
+    Query Params:
+      - lines: number of lines to return (default: 50)
+      
+    Example:
+      curl -H "Authorization: Bearer YOUR_KEY" "https://your-space.hf.space/logs/failures?lines=100"
+    """
+    allowed_logs = {
+        "failures": "failures.log",
+        "proxy": "proxy.log",
+        "debug": "proxy_debug.log",
+    }
+    
+    if log_type not in allowed_logs:
+        raise HTTPException(
+            status_code=400, 
+            detail=f"Invalid log type. Choose from: {list(allowed_logs.keys())}"
+        )
+        
+    log_file = LOG_DIR / allowed_logs[log_type]
+    
+    if not log_file.exists():
+        return JSONResponse(
+            status_code=404,
+            content={"error": "Log file not found", "path": str(log_file), "available_logs": list(allowed_logs.keys())}
+        )
+        
+    try:
+        # Simple/naive implementation - fine for small log files
+        with open(log_file, "r", encoding="utf-8", errors="replace") as f:
+            all_lines = f.readlines()
+            # Return last N lines as a single string
+            result = "".join(all_lines[-lines:])
+            return JSONResponse(content={"log": allowed_logs[log_type], "lines": lines, "content": result})
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Failed to read log file: {str(e)}")
+
+
 @app.post("/v1/token-count")
 async def token_count(
     request: Request,
@@ -1763,4 +1814,4 @@ if __name__ == "__main__":
             workers=workers,
             limit_concurrency=100,      # Max simultaneous connections per worker
             timeout_keep_alive=30,      # Keep-alive for connection reuse
-        )
+        )