"""Default parameters.

This is the first file that is imported from this software.
All initialization imports should occur here.

Please change depending on your preferences.
"""
import os
import shutil
import warnings
import yaml
from typing import Any, Mapping

from dosma.utils import env

import matplotlib
import nested_lookup

__all__ = ["preferences"]

# Parse preferences file
_resources_dir = env.resources_dir()
_internal_preferences_template_filepath = os.path.join(_resources_dir, "templates/.preferences.yml")
_preferences_cmd_line_filepath = os.path.join(
    _resources_dir, "templates/.preferences_cmd_line_schema.yml"
)

__preferences_filepath__ = os.path.join(_resources_dir, "preferences.yml")

if not os.path.isfile(__preferences_filepath__):
    shutil.copyfile(_internal_preferences_template_filepath, __preferences_filepath__)

# Default rounding for I/O (dicom, nifti, etc) - DO NOT CHANGE
AFFINE_DECIMAL_PRECISION = 4
SCANNER_ORIGIN_DECIMAL_PRECISION = 4

DEFAULT_FONT_SIZE = 16
DEFAULT_TEXT_SPACING = DEFAULT_FONT_SIZE * 0.01


class _Preferences(object):
    """A pseudo-Singleton class implementation to track preferences.

    Do not instantiate this class. To modify/update preferences use the ``preferences``
    module variable defined below.

    However, in the case this class is instantiated, a new object will be created,
    but all config properties will be shared among all instances. Meaning modifying
    the config in one instance will impact the preferences state in another instance.
    """

    _template_filepath = _internal_preferences_template_filepath
    _preferences_filepath = __preferences_filepath__
    __config = {}
    __key_list = []

    def __init__(self):
        # Load config and set information if config has not been initialized.
        if not self.__config:
            with open(self._preferences_filepath, "r") as f:
                self.__config = yaml.safe_load(f)
            with open(self._template_filepath, "r") as f:
                template = yaml.safe_load(f)

            self.__config = self._merge_dicts(self.__config, template)

            matplotlib.rcParams.update(self.__config["visualization"]["matplotlib"]["rcParams"])

            # Store all preference keys.
            self.__key_list = self._unroll_keys(self.config, "")

    def _merge_dicts(self, target, base, prefix=""):
        """Merges dicts from target onto base."""
        target_keys = target.keys()
        base_keys = base.keys()
        all_keys = target_keys | base_keys

        for k in all_keys:
            prefix_k = f"{prefix}/{k}" if prefix else k

            if k in target_keys and k not in base_keys:
                # Allow target config to specify parameters not in the base config.
                # Note this may change in the future.
                pass
            elif k not in target_keys and k in base_keys:
                warnings.warn(
                    "Your preferences file may be outdated. "
                    "Run `preferences.save()` to save your updated preferences."
                )
                target[k] = base[k]
            else:
                target_v = target[k]
                template_v = base[k]

                if isinstance(target_v, Mapping) and isinstance(template_v, Mapping):
                    target[k] = self._merge_dicts(target_v, template_v, prefix=prefix_k)
                elif isinstance(template_v, Mapping):
                    raise ValueError(
                        f"{prefix_k}: Got target type '{type(target_v)}', "
                        f"but base config type is '{type(template_v)}'"
                    )
                else:
                    # If both are not mapping, keep the target value.
                    pass

        return target

    def _unroll_keys(self, subdict: dict, prefix: str) -> list:
        """Recursive method to unroll keys."""
        pref_keys = []
        keys = subdict.keys()
        for k in keys:
            prefix_or_key = "{}/{}".format(prefix, k)
            val = subdict[k]
            if type(val) is dict:
                pref_keys.extend(self._unroll_keys(val, prefix_or_key))
            else:
                pref_keys.append(prefix_or_key)

        return pref_keys

    @staticmethod
    def _get_prefixes(prefix: str = ""):
        if not prefix:
            return ()

        r_prefixes = []
        prefixes = prefix.split("/")
        for p in prefixes:
            if p == "":
                continue
            r_prefixes.append(p)

        return r_prefixes

    def get(self, key: str, prefix: str = ""):
        """Get preference.

        Args:
            key (str): Preference to peek. Can be full path preference.
            prefix (str, optional): Prefix defining which sub-config to search.
                For example, to access the key ``"visualization/rcParams"``, this parameter
                should be ``"visualization"``.

        Returns:
            The preference value.
        """
        return self.__get(self.__config, key, prefix)[0]

    def __get(self, b_dict: dict, key: str, prefix: str = ""):
        """Get preference.

        Args:
            b_dict (dict): Dictionary to search.
            key (str): Preference to peek. Can be full path preference.
            prefix (:obj:`str`, optional): Prefix defining which sub-config to search.
                For example, to access the key ``"visualization/rcParams"``,
                this parameter should be ``"visualization"``.

        Returns:
            The preference value, sub-dictionary to search
        """
        p_keys = self._get_prefixes(key)
        key = p_keys[-1]
        k_prefixes = list(p_keys[:-1])

        prefixes = list(self._get_prefixes(prefix))
        prefixes.extend(k_prefixes)

        subdict = b_dict
        for p in prefixes:
            subdict = subdict[p]

        num_keys = nested_lookup.get_occurrence_of_key(subdict, key)

        if num_keys == 0:
            raise KeyError("Key not '%s ' found in prefix '%s'" % (key, prefix))
        if num_keys > 1:
            raise KeyError(
                "Multiple keys '%s 'found in prefix '%s'. Provide a more specific prefix."
                % (key, prefix)
            )

        return nested_lookup.nested_lookup(key, subdict)[0], subdict

    def set(self, key: str, value: Any, prefix: str = ""):
        """Set preference.

        Args:
            key (str): Preference to peek. Can be full path preference.
            value (Any): Value to set preference.
            prefix (str, optional): Prefix defining which sub-config to search.
                For example, to access the key ``visualization/rcParams``,
                this parameter should be `visualization`. Defaults to ''.

        Returns:
            The preference value.
        """
        val, subdict = self.__get(self.__config, key, prefix)

        # type of new value has to be the same type as old value
        if type(value) != type(val):
            try:
                value = type(val)(value)
            except (ValueError, TypeError):
                raise TypeError("Could not convert %s to %s: %s" % (type(value), type(val), value))

        p_keys = self._get_prefixes(key)
        key = p_keys[-1]
        nested_lookup.nested_update(subdict, key, value, in_place=True)

        # if param is an rcParam, update matplotlib
        if key in self.config["visualization"]["matplotlib"]["rcParams"].keys():
            matplotlib.rcParams.update({key: value})

    def save(self):
        """Save preferences to file.

        Args:
            file_path (str, optional): File path to yml file.
                Defaults to local preferences file.
        """
        with open(self._preferences_filepath, "w") as f:
            yaml.dump(self.__config, f, default_flow_style=False)

    @property
    def config(self):
        """Get preferences configuration."""
        return self.__config

    # Make certain preferences easily accessible through this class.
    @property
    def segmentation_batch_size(self) -> int:
        """int: Batch size for segmentation models."""
        return self.get("/segmentation/batch.size")

    @property
    def visualization_use_vmax(self) -> bool:
        return self.get("/visualization/use.vmax")

    @property
    def mask_dilation_rate(self) -> float:
        """float: rate for mask dilation."""
        return self.get("/registration/mask/dilation.rate")

    @property
    def mask_dilation_threshold(self) -> float:
        """float: threshold for mask dilation."""
        return self.get("/registration/mask/dilation.threshold")

    @property
    def fitting_r2_threshold(self) -> float:
        """float: r2 threshold for fitting"""
        return self.get("/fitting/r2.threshold")

    @property
    def image_data_format(self):
        """ImageDataFormat: Format for images (e.g. png, eps, etc.)."""
        from dosma.core.io.format_io import ImageDataFormat

        return ImageDataFormat[self.get("/data/format")]

    @property
    def nipype_logging(self) -> str:
        """str: nipype library logging mode."""
        # TODO: Remove this try/except clause when schema is well defined.
        try:
            return self.get("/logging/nipype")
        except KeyError:
            return "file_stderr"

    def cmd_line_flags(self) -> dict:
        """Provide command line flags for changing preferences via command line.

        Not all preferences are listed here. Only those that should easily be set.

        All default values will be based on the current state of preferences,
        not the static state specified in yaml file.

        Returns:
            Preference keys with corresponding argparse kwarg dict
        """
        with open(_preferences_cmd_line_filepath) as f:
            cmd_line_config = yaml.safe_load(f)

        cmd_line_dict = {}
        for k in self.__key_list:
            try:
                argparse_kwargs, _ = self.__get(cmd_line_config, k)
            except KeyError:
                continue

            argparse_kwargs["default"] = self.get(k)
            argparse_kwargs["type"] = eval(argparse_kwargs["type"])
            cmd_line_dict[k] = argparse_kwargs

        return cmd_line_dict

    def __str__(self):
        return str(self.__config)


preferences = _Preferences()