You are an AI in robot simulation code and task design. I will provide you some example tasks, code implementation, and some guidelines for how to generate tasks and then you will help me generate a new task. My goal is to design diverse and feasible tasks for tabletop manipulation. I will first ask you to describe the task in natural languages and then will let you write the code for it. 

=========
Here are all the assets. Use only these assets in the task and code design. 
"""
insertion/:
ell.urdf  fixture.urdf

bowl/:
bowl.urdf

box/:
box-template.urdf

stacking/:
block.urdf  stand.urdf

zone/:
zone.obj  zone.urdf

pallet/:
pallet.obj  pallet.urdf

ball/:
ball-template.urdf

cylinder/:
cylinder-template.urdf

bowl/:
bowl.urdf

# assets not for picking
corner/:
corner-template.urdf

line/:
single-green-line-template.urdf

container/:
container-template.urdf
"""

"""
import numpy as np
from cliport.tasks.task import Task
from cliport.utils import utils
import pybullet as p


class PlaceRedInGreen(Task):
    """pick up the red blocks and place them into the green bowls amidst other objects."""

    def __init__(self):
        super().__init__()
        self.max_steps = 10
        self.lang_template = "put the red blocks in a green bowl"
        self.task_completed_desc = "done placing blocks in bowls."
        self.additional_reset()

    def reset(self, env):
        super().reset(env)
        n_bowls = np.random.randint(1, 4)
        n_blocks = np.random.randint(1, n_bowls + 1)

        # Add bowls.
        # x, y, z dimensions for the asset size
        bowl_size = (0.12, 0.12, 0)
        bowl_urdf = 'bowl/bowl.urdf'
        bowl_poses = []
        for _ in range(n_bowls):
            bowl_pose = self.get_random_pose(env, obj_size=bowl_size)
            env.add_object(urdf=bowl_urdf, pose=bowl_pose, category='fixed')
            bowl_poses.append(bowl_pose)

        # Add blocks.
        # x, y, z dimensions for the asset size
        blocks = []
        block_size = (0.04, 0.04, 0.04)
        block_urdf = 'stacking/block.urdf'
        for _ in range(n_blocks):
            block_pose = self.get_random_pose(env, obj_size=block_size)
            block_id = env.add_object(block_urdf, block_pose)
            blocks.append(block_id)

        # Goal: each red block is in a different green bowl.
        self.add_goal(objs=blocks, matches=np.ones((len(blocks), len(bowl_poses))), targ_poses=bowl_poses, replace=False,
                rotations=True, metric='pose', params=None, step_max_reward=1)
        self.lang_goals.append(self.lang_template)

        # Colors of distractor objects.
        # IMPORTANT: RETRIEVE THE ACTUAL COLOR VALUES
        bowl_colors = [utils.COLORS[c] for c in utils.COLORS if c != 'green']
        block_colors = [utils.COLORS[c] for c in utils.COLORS if c != 'red']

        # Add distractors.
        n_distractors = 0
        while n_distractors < 6:
            is_block = np.random.rand() > 0.5
            urdf = block_urdf if is_block else bowl_urdf
            size = block_size if is_block else bowl_size
            colors = block_colors if is_block else bowl_colors
            pose = self.get_random_pose(env, obj_size=size)
            color = colors[n_distractors % len(colors)]

            obj_id = env.add_object(urdf, pose, color=color)
            n_distractors += 1
"""

=========
Please describe the task "TASK_NAME_TEMPLATE" in natural languages and format the answer in a python dictionary with keys "task-name" and value type string, "task-description" (one specific sentence) and value type string, and "assets-used" and value type list of strings.

=========
Now write the pybullet simulation code for the task "TASK_NAME_TEMPLATE" in python code block starting with ```python.