题意：从地图数据创建 OpenAI Gym 环境

问题背景：

I am just starting out with reinforcement learning and trying to create a custom environment with OpenAI gym. However, I am stumped with trying to create an environment (with roads and intersections) from map data, say Google Maps.

我刚刚开始学习强化学习，并尝试使用 OpenAI Gym 创建一个自定义环境。然而，我在尝试从地图数据（比如 Google 地图）创建一个包含道路和交叉口的环境时遇到了困难。

Would appreciate any help I can get.

任何帮助我都会非常感激。

问题解决：

The best way that I have found would be to use osmnx to have your map data ready in a graph format. It allows easy access of nodes that make up a street map in a graph or geopandas form. This data can then be used to drive the game logic. With regards to rendering, external rendering tools like PyGame can then be used.

我发现的最佳方法是使用 `osmnx` 将地图数据准备成图形格式。它允许以图形或 geopandas 形式轻松访问构成街道地图的节点。然后，这些数据可以用于驱动游戏逻辑。至于渲染，可以使用像 PyGame 这样的外部渲染工具。