"""
Legacy line-oriented schema parser base classes

Used for task inventory and wearables.
"""
from __future__ import annotations

import abc
import calendar
import dataclasses
import datetime as dt
import inspect
import logging
import re
from io import StringIO
from typing import *

import hippolyzer.lib.base.llsd as llsd

from hippolyzer.lib.base.datatypes import UUID

LOG = logging.getLogger(__name__)
_T = TypeVar("_T")


class SchemaFieldSerializer(abc.ABC, Generic[_T]):
    @classmethod
    @abc.abstractmethod
    def deserialize(cls, val: str) -> _T:
        pass

    @classmethod
    @abc.abstractmethod
    def serialize(cls, val: _T) -> str:
        pass

    @classmethod
    def from_llsd(cls, val: Any, flavor: str) -> _T:
        return val

    @classmethod
    def to_llsd(cls, val: _T, flavor: str) -> Any:
        return val


class SchemaDate(SchemaFieldSerializer[dt.datetime]):
    @classmethod
    def deserialize(cls, val: str) -> dt.datetime:
        return dt.datetime.fromtimestamp(int(val), dt.timezone.utc)

    @classmethod
    def serialize(cls, val: dt.datetime) -> str:
        return str(calendar.timegm(val.utctimetuple()))

    @classmethod
    def from_llsd(cls, val: Any, flavor: str) -> dt.datetime:
        return dt.datetime.fromtimestamp(val, dt.timezone.utc)

    @classmethod
    def to_llsd(cls, val: dt.datetime, flavor: str):
        return calendar.timegm(val.utctimetuple())


class SchemaHexInt(SchemaFieldSerializer[int]):
    @classmethod
    def deserialize(cls, val: str) -> int:
        return int(val, 16)

    @classmethod
    def serialize(cls, val: int) -> str:
        return "%08x" % val


class SchemaInt(SchemaFieldSerializer[int]):
    @classmethod
    def deserialize(cls, val: str) -> int:
        return int(val)

    @classmethod
    def serialize(cls, val: int) -> str:
        return str(val)


class SchemaMultilineStr(SchemaFieldSerializer[str]):
    @classmethod
    def deserialize(cls, val: str) -> str:
        # llinventory claims that it will parse multiple lines until it finds
        # an "|" terminator. That's not true. Use llinventory's _actual_ behaviour.
        return val.partition("|")[0]

    @classmethod
    def serialize(cls, val: str) -> str:
        return val + "|"


class SchemaStr(SchemaFieldSerializer[str]):
    @classmethod
    def deserialize(cls, val: str) -> str:
        return val

    @classmethod
    def serialize(cls, val: str) -> str:
        return val


class SchemaUUID(SchemaFieldSerializer[UUID]):
    @classmethod
    def from_llsd(cls, val: Any, flavor: str) -> UUID:
        # FetchInventory2 will return a string, but we want a UUID. It's not an issue
        # for us to return a UUID later there because it'll just cast to string if
        # that's what it wants
        return UUID(val)

    @classmethod
    def deserialize(cls, val: str) -> UUID:
        return UUID(val)

    @classmethod
    def serialize(cls, val: UUID) -> str:
        return str(val)


class SchemaLLSD(SchemaFieldSerializer[_T]):
    """Arbitrary LLSD embedded in a field"""
    @classmethod
    def deserialize(cls, val: str) -> _T:
        return llsd.parse_xml(val.partition("|")[0].encode("utf8"))

    @classmethod
    def serialize(cls, val: _T) -> str:
        # Don't include the XML header
        return llsd.format_xml(val).split(b">", 1)[1].decode("utf8") + "\n|"


_SCHEMA_SPEC = Union[Type[Union["SchemaBase", SchemaFieldSerializer]], SchemaFieldSerializer]


def schema_field(spec: _SCHEMA_SPEC, *, default=dataclasses.MISSING, init=True,
                 repr=True, hash=None, compare=True, llsd_name=None, llsd_only=False,
                 include_none=False) -> dataclasses.Field:  # noqa
    """Describe a field in the inventory schema and the shape of its value"""
    return dataclasses.field(  # noqa
        metadata={"spec": spec, "llsd_name": llsd_name, "llsd_only": llsd_only, "include_none": include_none},
        default=default, init=init, repr=repr, hash=hash, compare=compare,
    )


class SchemaParsingError(Exception):
    pass


# The schema is meant to allow multi-line strings, but in practice
# it does not due to scanf() shenanigans. This is fine.
_SCHEMA_LINE_TOKENS_RE = re.compile(r'\A\s*([^\s]+)(\s+([^\t\r\n]+))?$')


def parse_schema_line(line: str):
    g = _SCHEMA_LINE_TOKENS_RE.search(line)
    if not g:
        raise SchemaParsingError(f"{line!r} doesn't match the token regex")
    return g.group(1), g.group(3)


@dataclasses.dataclass
class SchemaBase(abc.ABC):
    @classmethod
    def _get_fields_dict(cls, llsd_flavor: Optional[str] = None):
        fields_dict = {}
        for field in dataclasses.fields(cls):
            field_name = field.name
            if llsd_flavor:
                field_name = field.metadata.get("llsd_name") or field_name
            fields_dict[field_name] = field
        return fields_dict

    @classmethod
    def from_str(cls, text: str):
        return cls.from_reader(StringIO(text))

    @classmethod
    @abc.abstractmethod
    def from_reader(cls: Type[_T], reader: StringIO) -> _T:
        pass

    @classmethod
    def from_bytes(cls, data: bytes):
        return cls.from_str(data.decode("utf8"))

    @classmethod
    def from_llsd(cls, inv_dict: Dict, flavor: str = "legacy"):
        fields = cls._get_fields_dict(llsd_flavor=flavor)
        obj_dict = {}
        try:
            for key, val in inv_dict.items():
                if key in fields:
                    field: dataclasses.Field = fields[key]
                    key = field.name
                    spec = field.metadata.get("spec")
                    # Not a real key, an internal var on our dataclass
                    if not spec:
                        LOG.warning(f"Internal key {key!r}")
                        continue

                    spec_cls = spec
                    if not inspect.isclass(spec_cls):
                        spec_cls = spec_cls.__class__

                    # some kind of nested structure like sale_info
                    if issubclass(spec_cls, SchemaBase):
                        obj_dict[key] = spec.from_llsd(val, flavor)
                    elif issubclass(spec_cls, SchemaFieldSerializer):
                        obj_dict[key] = spec.from_llsd(val, flavor)
                    else:
                        raise ValueError(f"Unsupported spec for {key!r}, {spec!r}")
                else:
                    if flavor != "ais":
                        # AIS has a number of different fields that are irrelevant depending on
                        # what exactly sent the payload
                        LOG.warning(f"Unknown key {key!r}")
        except:
            LOG.error(f"Failed to parse inventory schema: {inv_dict!r}")
            raise
        return cls._obj_from_dict(obj_dict)

    def to_bytes(self) -> bytes:
        return self.to_str().encode("utf8")

    def to_str(self) -> str:
        writer = StringIO()
        self.to_writer(writer)
        writer.seek(0)
        return writer.read()

    def to_llsd(self, flavor: str = "legacy"):
        obj_dict = {}
        for field_name, field in self._get_fields_dict(llsd_flavor=flavor).items():
            spec = field.metadata.get("spec")
            # Not meant to be serialized
            if not spec:
                continue

            val = getattr(self, field.name)
            if val is None:
                continue

            spec_cls = spec
            if not inspect.isclass(spec_cls):
                spec_cls = spec_cls.__class__

            # Some kind of nested structure like sale_info
            if isinstance(val, SchemaBase):
                val = val.to_llsd(flavor)
            elif issubclass(spec_cls, SchemaFieldSerializer):
                val = spec.to_llsd(val, flavor)
            else:
                raise ValueError(f"Bad inventory spec {spec!r}")
            obj_dict[field_name] = val
        return obj_dict

    @abc.abstractmethod
    def to_writer(self, writer: StringIO):
        pass

    @classmethod
    def _obj_from_dict(cls, obj_dict: Dict):
        return cls(**obj_dict)  # type: ignore