Source code for vmas.interactive_rendering

#  Copyright (c) 2022-2024.
#  ProrokLab (https://www.proroklab.org/)
#  All rights reserved.
"""
Use this script to interactively play with scenarios

You can change agent by pressing TAB
You can reset the environment by pressing R
You can move agents with the arrow keys
If you have more than 1 agent, you can control another one with W,A,S,D
and switch the agent with these controls using LSHIFT
"""

from argparse import ArgumentParser, BooleanOptionalAction
from operator import add
from typing import Dict, Union

import numpy as np
from torch import Tensor

from vmas.make_env import make_env
from vmas.simulator.environment.gym import GymWrapper
from vmas.simulator.scenario import BaseScenario
from vmas.simulator.utils import save_video

N_TEXT_LINES_INTERACTIVE = 6


class InteractiveEnv:
    """
    Use this script to interactively play with scenarios

    You can change agent by pressing TAB
    You can reset the environment by pressing R
    You can control agent actions with the arrow keys and M/N (left/right control the first action, up/down control the second, M/N controls the third)
    If you have more than 1 agent, you can control another one with W,A,S,D and Q,E in the same way.
    and switch the agent with these controls using LSHIFT
    """

    def __init__(
        self,
        env: GymWrapper,
        control_two_agents: bool = False,
        display_info: bool = True,
        save_render: bool = False,
        render_name: str = "interactive",
    ):
        self.env = env
        self.control_two_agents = control_two_agents
        # hard-coded keyboard events
        self.current_agent_index = 0
        self.current_agent_index2 = 1
        self.n_agents = self.env.unwrapped.n_agents
        self.agents = self.env.unwrapped.agents
        self.continuous = self.env.unwrapped.continuous_actions
        self.reset = False
        self.keys = np.array(
            [0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
        )  # up, down, left, right, rot+, rot-
        self.keys2 = np.array(
            [0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
        )  # up, down, left, right, rot+, rot-
        self.u = [0] * (3 if self.continuous else 2)
        self.u2 = [0] * (3 if self.continuous else 2)
        self.frame_list = []
        self.display_info = display_info
        self.save_render = save_render
        self.render_name = render_name

        if self.control_two_agents:
            assert (
                self.n_agents >= 2
            ), "Control_two_agents is true but not enough agents in scenario"

        self.text_lines = []
        self.font_size = 15
        self.env.render()
        self.text_idx = len(self.env.unwrapped.text_lines)
        self._init_text()
        self.env.unwrapped.viewer.window.on_key_press = self._key_press
        self.env.unwrapped.viewer.window.on_key_release = self._key_release

        self._cycle()

    def _increment_selected_agent_index(self, index: int):
        index += 1
        if index == self.n_agents:
            index = 0
        return index

    def _cycle(self):
        total_rew = [0] * self.n_agents
        while True:
            if self.reset:
                if self.save_render:
                    save_video(
                        self.render_name,
                        self.frame_list,
                        fps=1 / self.env.unwrapped.world.dt,
                    )
                self.env.reset()
                self.reset = False
                total_rew = [0] * self.n_agents

            if self.n_agents > 0:
                action_list = [[0.0] * agent.action_size for agent in self.agents]
                action_list[self.current_agent_index][
                    : self.agents[self.current_agent_index].dynamics.needed_action_size
                ] = self.u[
                    : self.agents[self.current_agent_index].dynamics.needed_action_size
                ]
            else:
                action_list = []

            if self.n_agents > 1 and self.control_two_agents:
                action_list[self.current_agent_index2][
                    : self.agents[self.current_agent_index2].dynamics.needed_action_size
                ] = self.u2[
                    : self.agents[self.current_agent_index2].dynamics.needed_action_size
                ]

            obs, rew, done, info = self.env.step(action_list)

            if self.display_info and self.n_agents > 0:
                # TODO: Determine number of lines of obs_str and render accordingly
                obs_str = str(InteractiveEnv.format_obs(obs[self.current_agent_index]))
                message = f"\t\t{obs_str[len(obs_str) // 2:]}"
                self._write_values(0, message)
                message = f"Obs: {obs_str[:len(obs_str) // 2]}"
                self._write_values(1, message)

                message = f"Rew: {round(rew[self.current_agent_index],3)}"
                self._write_values(2, message)

                total_rew = list(map(add, total_rew, rew))
                message = f"Total rew: {round(total_rew[self.current_agent_index], 3)}"
                self._write_values(3, message)

                message = f"Done: {done}"
                self._write_values(4, message)

                message = f"Selected: {self.env.unwrapped.agents[self.current_agent_index].name}"
                self._write_values(5, message)

            frame = self.env.render(
                mode="rgb_array" if self.save_render else "human",
                visualize_when_rgb=True,
            )
            if self.save_render:
                self.frame_list.append(frame)

            if done:
                self.reset = True

    def _init_text(self):
        from vmas.simulator import rendering

        for i in range(N_TEXT_LINES_INTERACTIVE):
            text_line = rendering.TextLine(
                y=(self.text_idx + i) * 40, font_size=self.font_size
            )
            self.env.unwrapped.viewer.add_geom(text_line)
            self.text_lines.append(text_line)

    def _write_values(self, index: int, message: str):
        self.text_lines[index].set_text(message)

    # keyboard event callbacks
    def _key_press(self, k, mod):
        from pyglet.window import key

        agent_range = self.agents[self.current_agent_index].action.u_range_tensor
        try:
            if k == key.LEFT:
                self.keys[0] = agent_range[0]
            elif k == key.RIGHT:
                self.keys[1] = agent_range[0]
            elif k == key.DOWN:
                self.keys[2] = agent_range[1]
            elif k == key.UP:
                self.keys[3] = agent_range[1]
            elif k == key.M:
                self.keys[4] = agent_range[2]
            elif k == key.N:
                self.keys[5] = agent_range[2]
            elif k == key.TAB:
                self.current_agent_index = self._increment_selected_agent_index(
                    self.current_agent_index
                )
                if self.control_two_agents:
                    while self.current_agent_index == self.current_agent_index2:
                        self.current_agent_index = self._increment_selected_agent_index(
                            self.current_agent_index
                        )

            if self.control_two_agents:
                agent2_range = self.agents[
                    self.current_agent_index2
                ].action.u_range_tensor

                if k == key.A:
                    self.keys2[0] = agent2_range[0]
                elif k == key.D:
                    self.keys2[1] = agent2_range[0]
                elif k == key.S:
                    self.keys2[2] = agent2_range[1]
                elif k == key.W:
                    self.keys2[3] = agent2_range[1]
                elif k == key.E:
                    self.keys2[4] = agent2_range[2]
                elif k == key.Q:
                    self.keys2[5] = agent2_range[2]

                elif k == key.LSHIFT:
                    self.current_agent_index2 = self._increment_selected_agent_index(
                        self.current_agent_index2
                    )
                    while self.current_agent_index == self.current_agent_index2:
                        self.current_agent_index2 = (
                            self._increment_selected_agent_index(
                                self.current_agent_index2
                            )
                        )
        except IndexError:
            print("Action not available")

        if k == key.R:
            self.reset = True

        self.set_u()

    def _key_release(self, k, mod):
        from pyglet.window import key

        if k == key.LEFT:
            self.keys[0] = 0
        elif k == key.RIGHT:
            self.keys[1] = 0
        elif k == key.DOWN:
            self.keys[2] = 0
        elif k == key.UP:
            self.keys[3] = 0
        elif k == key.M:
            self.keys[4] = 0
        elif k == key.N:
            self.keys[5] = 0

        if self.control_two_agents:
            if k == key.A:
                self.keys2[0] = 0
            elif k == key.D:
                self.keys2[1] = 0
            elif k == key.S:
                self.keys2[2] = 0
            elif k == key.W:
                self.keys2[3] = 0
            elif k == key.E:
                self.keys2[4] = 0
            elif k == key.Q:
                self.keys2[5] = 0

        self.set_u()

    def set_u(self):
        if self.continuous:
            self.u = [
                self.keys[1] - self.keys[0],
                self.keys[3] - self.keys[2],
                self.keys[4] - self.keys[5],
            ]
            self.u2 = [
                self.keys2[1] - self.keys2[0],
                self.keys2[3] - self.keys2[2],
                self.keys2[4] - self.keys2[5],
            ]
        else:
            if np.sum(self.keys[:4]) >= 1:
                self.u[0] = np.argmax(self.keys[:4]) + 1
            else:
                self.u[0] = 0
            if np.sum(self.keys[4:]) >= 1:
                self.u[1] = np.argmax(self.keys[4:]) + 1
            else:
                self.u[1] = 0

            if np.sum(self.keys2[:4]) >= 1:
                self.u2[0] = np.argmax(self.keys2[:4]) + 1
            else:
                self.u2[0] = 0
            if np.sum(self.keys2[4:]) >= 1:
                self.u2[1] = np.argmax(self.keys2[4:]) + 1
            else:
                self.u2[1] = 0

    @staticmethod
    def format_obs(obs):
        if isinstance(obs, (Tensor, np.ndarray)):
            return list(np.around(obs.tolist(), decimals=2))
        elif isinstance(obs, Dict):
            return {key: InteractiveEnv.format_obs(value) for key, value in obs.items()}
        else:
            raise NotImplementedError(f"Invalid type of observation {obs}")


[docs] def render_interactively( scenario: Union[str, BaseScenario], control_two_agents: bool = False, display_info: bool = True, save_render: bool = False, **kwargs, ): """Executes a scenario and renders it so that you can debug and control agents interactively. You can change the agent to control by pressing TAB. You can reset the environment by pressing R. You can control agent actions with the arrow keys and M/N (left/right control the first action, up/down control the second, M/N controls the third) If you have more than 1 agent, you can control another one with W,A,S,D and Q,E in the same way. and switch the agent using LSHIFT. Args: scenario (Union[str, BaseScenario]): Scenario to load. Can be the name of a file in `vmas.scenarios` folder or a :class:`~vmas.simulator.scenario.BaseScenario` class control_two_agents (bool, optional): Whether to control two agents or just one. Defaults to ``False``. display_info (bool, optional): Whether to display on the screen the following info from the first controlled agent: name, reward, total reward, done, and observation. Defaults to ``True``. save_render (bool, optional): Whether to save a video of the render up to the first reset. The video will be saved in the directory of this file with the name ``{scenario}_interactive``. Defaults to ``False``. Examples: >>> from vmas import render_interactively >>> render_interactively( ... "waterfall", ... control_two_agents=True, ... save_render=False, ... display_info=True, ... ) """ InteractiveEnv( make_env( scenario=scenario, num_envs=1, device="cpu", continuous_actions=True, wrapper="gym", seed=0, wrapper_kwargs={"return_numpy": False}, # Environment specific variables **kwargs, ), control_two_agents=control_two_agents, display_info=display_info, save_render=save_render, render_name=f"{scenario}_interactive" if isinstance(scenario, str) else "interactive", )
def parse_args(): parser = ArgumentParser(description="Interactive rendering") parser.add_argument( "--scenario", type=str, default="waterfall", help="Scenario to load. Can be the name of a file in `vmas.scenarios` folder or a :class:`~vmas.simulator.scenario.BaseScenario` class", ) parser.add_argument( "--control_two_agents", action=BooleanOptionalAction, default=True, help="Whether to control two agents or just one", ) parser.add_argument( "--display_info", action=BooleanOptionalAction, default=True, help="Whether to display on the screen the following info from the first controlled agent: name, reward, total reward, done, and observation", ) parser.add_argument( "--save_render", action="store_true", help="Whether to save a video of the render up to the first reset", ) return parser.parse_args() if __name__ == "__main__": # Use this script to interactively play with scenarios # # You can change agent by pressing TAB # You can reset the environment by pressing R # You can control agent actions with the arrow keys and M/N (left/right control the first action, up/down control the second, M/N controls the third) # If you have more than 1 agent, you can control another one with W,A,S,D and Q,E in the same way. # and switch the agent with these controls using LSHIFT args = parse_args() render_interactively( scenario=args.scenario, control_two_agents=args.control_two_agents, save_render=args.save_render, display_info=args.display_info, )