"""
Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.

NVIDIA CORPORATION and its licensors retain all intellectual property
and proprietary rights in and to this software, related documentation
and any modifications thereto. Any use, reproduction, disclosure or
distribution of this software and related documentation without an express
license agreement from NVIDIA CORPORATION is strictly prohibited.


Gym utilities
"""

from __future__ import print_function, division, absolute_import

from abc import ABC, abstractmethod
import math
import numpy as np
import argparse
from bisect import bisect

from . import gymapi


class LineGeometry(ABC):

    @abstractmethod
    def vertices(self):
        """ Numpy array of Vec3 with shape (num_lines(), 2) """

    @abstractmethod
    def colors(self):
        """ Numpy array of Vec3 with length num_lines() """

    def num_lines(self):
        return self.vertices().shape[0]

    # creates an instance of the vertices with the specified pose
    def instance_verts(self, pose=None):
        if pose is not None:
            return pose.transform_points(self.vertices())
        else:
            return np.copy(self.vertices())


class AxesGeometry(LineGeometry):
    def __init__(self, scale=1.0, pose=None):
        verts = np.empty((3, 2), gymapi.Vec3.dtype)
        verts[0][0] = (0, 0, 0)
        verts[0][1] = (scale, 0, 0)
        verts[1][0] = (0, 0, 0)
        verts[1][1] = (0, scale, 0)
        verts[2][0] = (0, 0, 0)
        verts[2][1] = (0, 0, scale)

        if pose is None:
            self.verts = verts
        else:
            self.verts = pose.transform_points(verts)

        colors = np.empty(3, gymapi.Vec3.dtype)
        colors[0] = (1.0, 0.0, 0.0)
        colors[1] = (0.0, 1.0, 0.0)
        colors[2] = (0.0, 0.0, 1.0)
        self._colors = colors

    def vertices(self):
        return self.verts

    def colors(self):
        return self._colors


class WireframeBoxGeometry(LineGeometry):
    def __init__(self, xdim=1, ydim=1, zdim=1, pose=None, color=None):
        if color is None:
            color = (1, 0, 0)

        num_lines = 12

        x = 0.5 * xdim
        y = 0.5 * ydim
        z = 0.5 * zdim

        verts = np.empty((num_lines, 2), gymapi.Vec3.dtype)
        # top face
        verts[0][0] = (x, y, z)
        verts[0][1] = (x, y, -z)
        verts[1][0] = (-x, y, z)
        verts[1][1] = (-x, y, -z)
        verts[2][0] = (x, y, z)
        verts[2][1] = (-x, y, z)
        verts[3][0] = (x, y, -z)
        verts[3][1] = (-x, y, -z)
        # bottom face
        verts[4][0] = (x, -y, z)
        verts[4][1] = (x, -y, -z)
        verts[5][0] = (-x, -y, z)
        verts[5][1] = (-x, -y, -z)
        verts[6][0] = (x, -y, z)
        verts[6][1] = (-x, -y, z)
        verts[7][0] = (x, -y, -z)
        verts[7][1] = (-x, -y, -z)
        # verticals
        verts[8][0] = (x, y, z)
        verts[8][1] = (x, -y, z)
        verts[9][0] = (x, y, -z)
        verts[9][1] = (x, -y, -z)
        verts[10][0] = (-x, y, z)
        verts[10][1] = (-x, -y, z)
        verts[11][0] = (-x, y, -z)
        verts[11][1] = (-x, -y, -z)

        if pose is None:
            self.verts = verts
        else:
            self.verts = pose.transform_points(verts)

        colors = np.empty(num_lines, gymapi.Vec3.dtype)
        colors.fill(color)
        self._colors = colors

    def vertices(self):
        return self.verts

    def colors(self):
        return self._colors


class WireframeBBoxGeometry(LineGeometry):

    def __init__(self, bbox, pose=None, color=None):
        if bbox.shape != (2, 3):
            raise ValueError('Expected bbox to be a matrix of 2 by 3!')

        if color is None:
            color = (1, 0, 0)

        num_lines = 12

        min_x, min_y, min_z = bbox[0]
        max_x, max_y, max_z = bbox[1]

        verts = np.empty((num_lines, 2), gymapi.Vec3.dtype)
        # top face
        verts[0][0] = (max_x, max_y, max_z)
        verts[0][1] = (max_x, max_y, min_z)
        verts[1][0] = (min_x, max_y, max_z)
        verts[1][1] = (min_x, max_y, min_z)
        verts[2][0] = (max_x, max_y, max_z)
        verts[2][1] = (min_x, max_y, max_z)
        verts[3][0] = (max_x, max_y, min_z)
        verts[3][1] = (min_x, max_y, min_z)

        # bottom face
        verts[4][0] = (max_x, min_y, max_z)
        verts[4][1] = (max_x, min_y, min_z)
        verts[5][0] = (min_x, min_y, max_z)
        verts[5][1] = (min_x, min_y, min_z)
        verts[6][0] = (max_x, min_y, max_z)
        verts[6][1] = (min_x, min_y, max_z)
        verts[7][0] = (max_x, min_y, min_z)
        verts[7][1] = (min_x, min_y, min_z)

        # verticals
        verts[8][0] = (max_x, max_y, max_z)
        verts[8][1] = (max_x, min_y, max_z)
        verts[9][0] = (max_x, max_y, min_z)
        verts[9][1] = (max_x, min_y, min_z)
        verts[10][0] = (min_x, max_y, max_z)
        verts[10][1] = (min_x, min_y, max_z)
        verts[11][0] = (min_x, max_y, min_z)
        verts[11][1] = (min_x, min_y, min_z)

        if pose is None:
            self.verts = verts
        else:
            self.verts = pose.transform_points(verts)

        colors = np.empty(num_lines, gymapi.Vec3.dtype)
        colors.fill(color)
        self._colors = colors

    def vertices(self):
        return self.verts

    def colors(self):
        return self._colors


class WireframeSphereGeometry(LineGeometry):

    def __init__(self, radius=1.0, num_lats=8, num_lons=8, pose=None, color=None, color2=None):
        if color is None:
            color = (1, 0, 0)

        if color2 is None:
            color2 = color

        num_lines = 2 * num_lats * num_lons

        verts = np.empty((num_lines, 2), gymapi.Vec3.dtype)
        colors = np.empty(num_lines, gymapi.Vec3.dtype)
        idx = 0

        ustep = 2 * math.pi / num_lats
        vstep = math.pi / num_lons

        u = 0.0
        for i in range(num_lats):
            v = 0.0
            for j in range(num_lons):
                x1 = radius * math.sin(v) * math.sin(u)
                y1 = radius * math.cos(v)
                z1 = radius * math.sin(v) * math.cos(u)

                x2 = radius * math.sin(v + vstep) * math.sin(u)
                y2 = radius * math.cos(v + vstep)
                z2 = radius * math.sin(v + vstep) * math.cos(u)

                x3 = radius * math.sin(v + vstep) * math.sin(u + ustep)
                y3 = radius * math.cos(v + vstep)
                z3 = radius * math.sin(v + vstep) * math.cos(u + ustep)

                verts[idx][0] = (x1, y1, z1)
                verts[idx][1] = (x2, y2, z2)
                colors[idx] = color

                idx += 1

                verts[idx][0] = (x2, y2, z2)
                verts[idx][1] = (x3, y3, z3)
                colors[idx] = color2

                idx += 1

                v += vstep
            u += ustep

        if pose is None:
            self.verts = verts
        else:
            self.verts = pose.transform_points(verts)

        self._colors = colors

    def vertices(self):
        return self.verts

    def colors(self):
        return self._colors


def draw_lines(geom, gym, viewer, env, pose):
    """
    Add line geometry to viewer
    :param geom: An instance of LineGeometry.
    :param gym: Gym API object.
    :param viewer: GymViewer object.
    :param env: If not None, pose is in that env's coordinate space.  If None, pose is in the global coordinate space.
    :param pose: The pose of the geometry to be drawn.
    """
    verts = geom.instance_verts(pose)
    gym.add_lines(viewer, env, geom.num_lines(), verts, geom.colors())


def draw_line(p1, p2, color, gym, viewer, env):
    verts = np.empty((1, 2), dtype=gymapi.Vec3.dtype)
    verts[0][0] = (p1.x, p1.y, p1.z)
    verts[0][1] = (p2.x, p2.y, p2.z)
    colors = np.empty(1, dtype=gymapi.Vec3.dtype)
    colors[0] = (color.x, color.y, color.z)
    gym.add_lines(viewer, env, 1, verts, colors)


def parse_device_str(device_str):
    # defaults
    device = 'cpu'
    device_id = 0

    if device_str == 'cpu' or device_str == 'cuda':
        device = device_str
        device_id = 0
    else:
        device_args = device_str.split(':')
        assert len(device_args) == 2 and device_args[0] == 'cuda', f'Invalid device string "{device_str}"'
        device, device_id_s = device_args
        try:
            device_id = int(device_id_s)
        except ValueError:
            raise ValueError(f'Invalid device string "{device_str}". Cannot parse "{device_id}"" as a valid device id')
    return device, device_id

# parses all of the common Gym example arguments and returns them to the caller
# note that args.physics_engine stores the gymapi value for the desired physics engine


def parse_arguments(description="Isaac Gym Example", headless=False, no_graphics=False, custom_parameters=[]):
    parser = argparse.ArgumentParser(description=description)
    if headless:
        parser.add_argument('--headless', action='store_true', help='Run headless without creating a viewer window')
    if no_graphics:
        parser.add_argument('--nographics', action='store_true',
                            help='Disable graphics context creation, no viewer window is created, and no headless rendering is available')
    parser.add_argument('--sim_device', type=str, default="cuda:0", help='Physics Device in PyTorch-like syntax')
    parser.add_argument('--pipeline', type=str, default="gpu", help='Tensor API pipeline (cpu/gpu)')
    parser.add_argument('--graphics_device_id', type=int, default=0, help='Graphics Device ID')

    physics_group = parser.add_mutually_exclusive_group()
    physics_group.add_argument('--flex', action='store_true', help='Use FleX for physics')
    physics_group.add_argument('--physx', action='store_true', help='Use PhysX for physics')

    parser.add_argument('--num_threads', type=int, default=0, help='Number of cores used by PhysX')
    parser.add_argument('--subscenes', type=int, default=0, help='Number of PhysX subscenes to simulate in parallel')
    parser.add_argument('--slices', type=int, help='Number of client threads that process env slices')

    for argument in custom_parameters:
        if ("name" in argument) and ("type" in argument or "action" in argument):
            help_str = ""
            if "help" in argument:
                help_str = argument["help"]

            if "type" in argument:
                if "default" in argument:
                    parser.add_argument(argument["name"], type=argument["type"], default=argument["default"], help=help_str)
                else:
                    parser.add_argument(argument["name"], type=argument["type"], help=help_str)
            elif "action" in argument:
                parser.add_argument(argument["name"], action=argument["action"], help=help_str)

        else:
            print()
            print("ERROR: command line argument name, type/action must be defined, argument not added to parser")
            print("supported keys: name, type, default, action, help")
            print()

    args = parser.parse_args()

    args.sim_device_type, args.compute_device_id = parse_device_str(args.sim_device)
    pipeline = args.pipeline.lower()

    assert (pipeline == 'cpu' or pipeline in ('gpu', 'cuda')), f"Invalid pipeline '{args.pipeline}'. Should be either cpu or gpu."
    args.use_gpu_pipeline = (pipeline in ('gpu', 'cuda'))

    if args.sim_device_type != 'cuda' and args.flex:
        print("Can't use Flex with CPU. Changing sim device to 'cuda:0'")
        args.sim_device = 'cuda:0'
        args.sim_device_type, args.compute_device_id = parse_device_str(args.sim_device)

    if (args.sim_device_type != 'cuda' and pipeline == 'gpu'):
        print("Can't use GPU pipeline with CPU Physics. Changing pipeline to 'CPU'.")
        args.pipeline = 'CPU'
        args.use_gpu_pipeline = False

    # Default to PhysX
    args.physics_engine = gymapi.SIM_PHYSX
    args.use_gpu = (args.sim_device_type == 'cuda')

    if args.flex:
        args.physics_engine = gymapi.SIM_FLEX

    # Using --nographics implies --headless
    if no_graphics and args.nographics:
        args.headless = True

    if args.slices is None:
        args.slices = args.subscenes

    return args

# parse sim options provided in sim_cfg to gym api sim_options

# NOTE: This function is deprecated and will not be supported in future releases.


def parse_sim_config(sim_cfg: dict, sim_options: gymapi.SimParams):
    # Todo:: add param value checks
    opt = "dt"
    if opt in sim_cfg:
        val = float(sim_cfg[opt])
        sim_options.dt = val

    opt = "substeps"
    if opt in sim_cfg:
        val = int(sim_cfg[opt])
        sim_options.substeps = val

    opt = "up_axis"
    if opt in sim_cfg:
        val = gymapi.UpAxis(sim_cfg[opt])
        sim_options.up_axis = val

    opt = "gravity"
    if opt in sim_cfg:
        val = tuple(sim_cfg[opt])
        sim_options.gravity = gymapi.Vec3(val[0], val[1], val[2])

    opt = "use_gpu_pipeline"
    if opt in sim_cfg:
        val = bool(sim_cfg[opt])
        sim_options.use_gpu_pipeline = val

    if "flex" in sim_cfg:
        parse_flex_config(sim_cfg["flex"], sim_options)

    if "physx" in sim_cfg:
        parse_physx_config(sim_cfg["physx"], sim_options)

# parse flex sim options

# NOTE: This function is deprecated and will not be supported in future releases.


def parse_flex_config(flex_cfg: dict, sim_options: gymapi.SimParams):
    ints = ["solver_type", "num_outer_iterations", "num_inner_iterations", "friction_mode"]
    floats = ["relaxation", "warm_start", "contact_regularization",
              "geometric_stiffness", "shape_collision_distance", "shape_collision_margin",
              "dynamic_friction", "static_friction", "particle_friction"]
    bools = ["deterministic_mode"]

    params = {"bool": bools, "int": ints, "float": floats}

    parse_float_int_bool(flex_cfg, sim_options.flex, params)

# parse physx sim options

# NOTE: This function is deprecated and will not be supported in future releases.


def parse_physx_config(physx_cfg: dict, sim_options: gymapi.SimParams):
    ints = ["num_threads", "solver_type", "num_position_iterations", "num_velocity_iterations", "max_gpu_contact_pairs", "num_subscenes", "contact_collection"]
    floats = ["contact_offset", "rest_offset", "bounce_threshold_velocity", "max_depenetration_velocity",
              "default_buffer_size_multiplier", "friction_correlation_distance", "friction_offset_threshold"]
    bools = ["use_gpu", "always_use_articulations"]

    params = {"bool": bools, "int": ints, "float": floats}

    parse_float_int_bool(physx_cfg, sim_options.physx, params)

# parses float, int, and bools listed in float_int_bool from cfg and stores them in opts

# NOTE: This function is deprecated and will not be supported in future releases.


def parse_float_int_bool(cfg: dict, opts: object, float_int_bool: dict):
    if "float" in float_int_bool:
        for opt in float_int_bool["float"]:
            if opt in cfg:
                val = float(cfg[opt])
                setattr(opts, opt, val)

    if "int" in float_int_bool:
        for opt in float_int_bool["int"]:
            if opt in cfg:
                if opt == "contact_collection":
                    val = gymapi.ContactCollection(cfg[opt])
                else:
                    val = int(cfg[opt])
                setattr(opts, opt, val)

    if "bool" in float_int_bool:
        for opt in float_int_bool["bool"]:
            if opt in cfg:
                val = bool(cfg[opt])
                setattr(opts, opt, val)


def parse_bool(v):
    if isinstance(v, bool):
        return v
    if isinstance(v, int):
        if v == 1:
            return True
        elif v == 0:
            return False
    if isinstance(v, str):
        if v.lower() in ("true", "yes", "t", "y", "1"):
            return True
        elif v.lower() in ("false", "no", "f", "n", "0"):
            return False
    else:
        raise argparse.ArgumentTypeError("Boolean value expected.")


def get_property_setter_map(gym):
    property_to_setters = {
        "dof_properties": gym.set_actor_dof_properties,
        "tendon_properties": gym.set_actor_tendon_properties,
        "rigid_body_properties": gym.set_actor_rigid_body_properties,
        "rigid_shape_properties": gym.set_actor_rigid_shape_properties,
        "sim_params": gym.set_sim_params,
    }

    return property_to_setters


def get_property_getter_map(gym):
    property_to_getters = {
        "dof_properties": gym.get_actor_dof_properties,
        "tendon_properties": gym.get_actor_tendon_properties,
        "rigid_body_properties": gym.get_actor_rigid_body_properties,
        "rigid_shape_properties": gym.get_actor_rigid_shape_properties,
        "sim_params": gym.get_sim_params,
    }

    return property_to_getters


def get_default_setter_args(gym):
    property_to_setter_args = {
        "dof_properties": [],
        "tendon_properties": [],
        "rigid_body_properties": [True],
        "rigid_shape_properties": [],
        "sim_params": [],
    }

    return property_to_setter_args


def generate_random_samples(attr_randomization_params, shape, randomization_ct,
                            extern_sample=None):
    rand_range = attr_randomization_params['range']
    distribution = attr_randomization_params['distribution']
    sched_type = attr_randomization_params['schedule'] if 'schedule' in attr_randomization_params else None
    sched_step = attr_randomization_params['schedule_steps'] if 'schedule' in attr_randomization_params else None
    operation = attr_randomization_params['operation']
    if sched_type == 'linear':
        sched_scaling = 1 / sched_step * min(randomization_ct, sched_step)
    elif sched_type == 'constant':
        sched_scaling = 0 if randomization_ct < sched_step else 1
    else:
        sched_scaling = 1

    if extern_sample is not None:
        sample = extern_sample
        if operation == 'additive':
            sample *= sched_scaling
        elif operation == 'scaling':
            sample = sample * sched_scaling + 1 * (1 - sched_scaling)
    elif distribution == "gaussian":
        mu, var = rand_range
        if operation == 'additive':
            mu *= sched_scaling
            var *= sched_scaling
        elif operation == 'scaling':
            var = var * sched_scaling  # scale up var over time
            mu = mu * sched_scaling + 1 * (1 - sched_scaling)  # linearly interpolate
        sample = np.random.normal(mu, var, shape)
    elif distribution == "loguniform":
        lo, hi = rand_range
        if operation == 'additive':
            lo *= sched_scaling
            hi *= sched_scaling
        elif operation == 'scaling':
            lo = lo * sched_scaling + 1 * (1 - sched_scaling)
            hi = hi * sched_scaling + 1 * (1 - sched_scaling)
        sample = np.exp(np.random.uniform(np.log(lo), np.log(hi), shape))
    elif distribution == "uniform":
        lo, hi = rand_range
        if operation == 'additive':
            lo *= sched_scaling
            hi *= sched_scaling
        elif operation == 'scaling':
            lo = lo * sched_scaling + 1 * (1 - sched_scaling)
            hi = hi * sched_scaling + 1 * (1 - sched_scaling)
        sample = np.random.uniform(lo, hi, shape)
    return sample


def get_bucketed_val(new_prop_val, attr_randomization_params):
    if attr_randomization_params['distribution'] == 'uniform':
        # range of buckets defined by uniform distribution
        lo, hi = attr_randomization_params['range'][0], attr_randomization_params['range'][1]
    else:
        # for gaussian, set range of buckets to be 2 stddev away from mean
        lo = attr_randomization_params['range'][0] - 2 * np.sqrt(attr_randomization_params['range'][1])
        hi = attr_randomization_params['range'][0] + 2 * np.sqrt(attr_randomization_params['range'][1])
    num_buckets = attr_randomization_params['num_buckets']
    buckets = [(hi - lo) * i / num_buckets + lo for i in range(num_buckets)]
    return buckets[bisect(buckets, new_prop_val) - 1]


def apply_random_samples(prop, og_prop, attr, attr_randomization_params,
                         randomization_ct, extern_sample=None):
    if isinstance(prop, gymapi.SimParams):
        if attr == 'gravity':
            sample = generate_random_samples(attr_randomization_params, 3, randomization_ct)
            if attr_randomization_params['operation'] == 'scaling':
                prop.gravity.x = og_prop['gravity'].x * sample[0]
                prop.gravity.y = og_prop['gravity'].y * sample[1]
                prop.gravity.z = og_prop['gravity'].z * sample[2]
            elif attr_randomization_params['operation'] == 'additive':
                prop.gravity.x = og_prop['gravity'].x + sample[0]
                prop.gravity.y = og_prop['gravity'].y + sample[1]
                prop.gravity.z = og_prop['gravity'].z + sample[2]
    elif isinstance(prop, np.ndarray):
        sample = generate_random_samples(attr_randomization_params, prop[attr].shape,
                                         randomization_ct, extern_sample)
        if attr_randomization_params['operation'] == 'scaling':
            new_prop_val = og_prop[attr] * sample
        elif attr_randomization_params['operation'] == 'additive':
            new_prop_val = og_prop[attr] + sample

        if 'num_buckets' in attr_randomization_params and attr_randomization_params['num_buckets'] > 0:
            new_prop_val = get_bucketed_val(new_prop_val, attr_randomization_params)
        prop[attr] = new_prop_val
    else:
        sample = generate_random_samples(attr_randomization_params, 1,
                                         randomization_ct, extern_sample)
        cur_attr_val = og_prop[attr]
        if attr_randomization_params['operation'] == 'scaling':
            new_prop_val = cur_attr_val * sample
        elif attr_randomization_params['operation'] == 'additive':
            new_prop_val = cur_attr_val + sample

        if 'num_buckets' in attr_randomization_params and attr_randomization_params['num_buckets'] > 0:
            new_prop_val = get_bucketed_val(new_prop_val, attr_randomization_params)
        setattr(prop, attr, new_prop_val)


def check_buckets(gym, envs, dr_params):
    total_num_buckets = 0
    for actor, actor_properties in dr_params["actor_params"].items():
        cur_num_buckets = 0

        if 'rigid_shape_properties' in actor_properties.keys():
            prop_attrs = actor_properties['rigid_shape_properties']
            if 'restitution' in prop_attrs and 'num_buckets' in prop_attrs['restitution']:
                cur_num_buckets = prop_attrs['restitution']['num_buckets']
            if 'friction' in prop_attrs and 'num_buckets' in prop_attrs['friction']:
                if cur_num_buckets > 0:
                    cur_num_buckets *= prop_attrs['friction']['num_buckets']
                else:
                    cur_num_buckets = prop_attrs['friction']['num_buckets']
            total_num_buckets += cur_num_buckets

    assert total_num_buckets <= 64000, 'Explicit material bucketing has been specified, but the provided total bucket count exceeds 64K: {} specified buckets'.format(
        total_num_buckets)

    shape_ct = 0

    # Separate loop because we should not assume that each actor is present in each env
    for env in envs:
        for i in range(gym.get_actor_count(env)):
            actor_handle = gym.get_actor_handle(env, i)
            actor_name = gym.get_actor_name(env, actor_handle)
            if actor_name in dr_params["actor_params"] and 'rigid_shape_properties' in dr_params["actor_params"][actor_name]:
                shape_ct += gym.get_actor_rigid_shape_count(env, actor_handle)

    assert shape_ct <= 64000 or total_num_buckets > 0, 'Explicit material bucketing is not used but the total number of shapes exceeds material limit. Please specify bucketing to limit material count.'


def _indent_xml(elem, level=0):
    i = "\n" + level * "  "
    if len(elem):
        if not elem.text or not elem.text.strip():
            elem.text = i + "  "
        if not elem.tail or not elem.tail.strip():
            elem.tail = i
        for elem in elem:
            _indent_xml(elem, level + 1)
        if not elem.tail or not elem.tail.strip():
            elem.tail = i
    else:
        if level and (not elem.tail or not elem.tail.strip()):
            elem.tail = i