import threading
import warnings

import cupy
from cupy import _util
from cupy._core import _accelerator
from cupy.cuda.device import Handle


_tls = threading.local()

cuquantum = cutensornet = tensornet = None


def _maybe_lazy_load_cutensornet():
    global cuquantum, cutensornet, tensornet

    if cuquantum is not None:
        return

    try:
        import cuquantum
        if hasattr(cuquantum, 'bindings'):
            # cuquantum-python >= 25.03
            from cuquantum.bindings import cutensornet  # binding module
            from cuquantum import tensornet  # module for pythonic APIs
        else:
            # for cuquantum < 25.03, bindings & pythonic APIs
            # all reside under cuquantum.cutensornet
            from cuquantum import cutensornet
            tensornet = cutensornet
    except ImportError:
        pass


@_util.memoize()
def _is_cuqnt_22_11_or_higher():
    ver = [int(i) for i in cuquantum.__version__.split('.')]
    if (ver[0] > 22) or (ver[0] == 22 and ver[1] >= 11):
        return True
    return False


def _is_nonblocking_supported():
    return _is_cuqnt_22_11_or_higher()


def _get_einsum_operands(args):
    """Parse & retrieve einsum operands, assuming ``args`` is in either
    "subscript" or "interleaved" format.
    """

    if len(args) == 0:
        raise ValueError(
            'must specify the einstein sum subscripts string and at least one '
            'operand, or at least one operand and its corresponding '
            'subscripts list')

    if isinstance(args[0], str):
        expr = args[0]
        operands = list(args[1:])
        return expr, operands
    else:
        args = list(args)
        operands = []
        inputs = []
        output = None
        while len(args) >= 2:
            operands.append(args.pop(0))
            inputs.append(args.pop(0))
        if len(args) == 1:
            output = args.pop(0)
        assert not args
        return inputs, operands, output


def _try_use_cutensornet(*args, **kwargs):
    if cupy.cuda.runtime.is_hip:
        return None

    if (_accelerator.ACCELERATOR_CUTENSORNET not in
            _accelerator.get_routine_accelerators()):
        return None

    _maybe_lazy_load_cutensornet()

    if cutensornet is None:
        warnings.warn(
            'using the cuTensorNet backend was requested but it cannot be '
            'imported -- maybe you forgot to install cuQuantum Python? '
            'Please do "pip install cuquantum-python" or "conda install '
            '-c conda-forge cuquantum-python" and retry',
            stacklevel=2)
        return None

    # cannot pop as we might still need kwargs later
    dtype = kwargs.get('dtype', None)
    path = kwargs.get('optimize', False)
    if path is True:
        path = 'greedy'

    # we do very lightweight pre-processing here just to inspect the
    # operands; the actual input verification is deferred to cuTensorNet
    # which can generate far better diagonostic messages
    args = _get_einsum_operands(args)
    operands = [cupy.asarray(op) for op in args[1]]

    if len(operands) == 1:
        # As of cuTENSOR 1.5.0 it still chokes with some common operations
        # like trace ("ii->") so it's easier to just skip all single-operand
        # cases instead of whitelisting what could be done explicitly
        return None

    if (any(op.size == 0 for op in operands) or
            any(len(op.shape) == 0 for op in operands)):
        # To cuTensorNet the shape is invalid
        return None

    # all input dtypes must be identical (to a numerical dtype)
    result_dtype = cupy.result_type(*operands) if dtype is None else dtype
    if result_dtype not in (
            cupy.float32, cupy.float64, cupy.complex64, cupy.complex128):
        return None
    operands = [op.astype(result_dtype, copy=False) for op in operands]

    # prepare cutn inputs
    device = cupy.cuda.runtime.getDevice()
    if not hasattr(_tls, "cutn_handle_cache"):
        cutn_handle_cache = _tls.cutn_handle_cache = {}
    else:
        cutn_handle_cache = _tls.cutn_handle_cache
    handle = cutn_handle_cache.get(device)
    if handle is None:
        handle = cutensornet.create()
        cutn_handle_cache[device] = Handle(handle, cutensornet.destroy)
    else:
        handle = handle.handle
    cutn_options = {'device_id': device, 'handle': handle}
    if _is_nonblocking_supported():
        cutn_options['blocking'] = "auto"

    # TODO(leofang): support all valid combinations:
    # - path from user, contract with cutn (done)
    # - path from cupy, contract with cutn (not yet)
    # - path from cutn, contract with cutn (done)
    # - path from cutn, contract with cupy (not yet)
    raise_warning = False
    if path is False:
        # following the same convention (contracting from the right) as would
        # be produced by _iter_path_pairs(), but converting to a list of pairs
        # due to cuTensorNet's requirement
        path = [(i-1, i-2) for i in range(len(operands), 1, -1)]
    elif len(path) and path[0] == 'einsum_path':
        # let cuTensorNet check if the format is correct
        path = path[1:]
    elif len(path) == 2:
        if isinstance(path[1], (int, float)):
            raise_warning = True
        if path[0] != 'cutensornet':
            raise_warning = True
        path = None
    else:  # path is a string
        if path != 'cutensornet':
            raise_warning = True
        path = None
    if raise_warning:
        warnings.warn(
            'the cuTensorNet backend ignores the "optimize" option '
            'except when an explicit contraction path is provided '
            'or when optimize=False (disable optimization); also, '
            'the maximum intermediate size, if set, is ignored',
            stacklevel=2)
    cutn_optimizer = {'path': path} if path else None

    if len(args) == 2:
        out = tensornet.contract(
            args[0], *operands, options=cutn_options, optimize=cutn_optimizer)
    elif len(args) == 3:
        inputs = [i for pair in zip(operands, args[0]) for i in pair]
        if args[2] is not None:
            inputs.append(args[2])
        out = tensornet.contract(
            *inputs, options=cutn_options, optimize=cutn_optimizer)
    else:
        assert False

    return out