# Copyright (c) 2015 Hewlett-Packard Development Company, L.P.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#    http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import contextlib
import fnmatch
import functools
import inspect
import jmespath
import munch
import netifaces
import re
import six
import sre_constants
import sys
import time
import uuid

from decorator import decorator

from shade import _log
from shade import exc
from shade import meta

_decorated_methods = []


def _exc_clear():
    """Because sys.exc_clear is gone in py3 and is not in six."""
    if sys.version_info[0] == 2:
        sys.exc_clear()


def _iterate_timeout(timeout, message, wait=2):
    """Iterate and raise an exception on timeout.

    This is a generator that will continually yield and sleep for
    wait seconds, and if the timeout is reached, will raise an exception
    with <message>.

    """
    log = _log.setup_logging('shade.iterate_timeout')

    try:
        # None as a wait winds up flowing well in the per-resource cache
        # flow. We could spread this logic around to all of the calling
        # points, but just having this treat None as "I don't have a value"
        # seems friendlier
        if wait is None:
            wait = 2
        elif wait == 0:
            # wait should be < timeout, unless timeout is None
            wait = 0.1 if timeout is None else min(0.1, timeout)
        wait = float(wait)
    except ValueError:
        raise exc.OpenStackCloudException(
            "Wait value must be an int or float value. {wait} given"
            " instead".format(wait=wait))

    start = time.time()
    count = 0
    while (timeout is None) or (time.time() < start + timeout):
        count += 1
        yield count
        log.debug('Waiting %s seconds', wait)
        time.sleep(wait)
    raise exc.OpenStackCloudTimeout(message)


def _make_unicode(input):
    """Turn an input into unicode unconditionally

    :param input:
       A unicode, string or other object
    """
    try:
        if isinstance(input, unicode):
            return input
        if isinstance(input, str):
            return input.decode('utf-8')
        else:
            # int, for example
            return unicode(input)
    except NameError:
        # python3!
        return str(input)


def _dictify_resource(resource):
    if isinstance(resource, list):
        return [_dictify_resource(r) for r in resource]
    else:
        if hasattr(resource, 'toDict'):
            return resource.toDict()
        else:
            return resource


def _filter_list(data, name_or_id, filters):
    """Filter a list by name/ID and arbitrary meta data.

    :param list data:
        The list of dictionary data to filter. It is expected that
        each dictionary contains an 'id' and 'name'
        key if a value for name_or_id is given.
    :param string name_or_id:
        The name or ID of the entity being filtered. Can be a glob pattern,
        such as 'nb01*'.
    :param filters:
        A dictionary of meta data to use for further filtering. Elements
        of this dictionary may, themselves, be dictionaries. Example::

            {
              'last_name': 'Smith',
              'other': {
                  'gender': 'Female'
              }
            }
        OR
        A string containing a jmespath expression for further filtering.
    """
    # The logger is shade.fmmatch to allow a user/operator to configure logging
    # not to communicate about fnmatch misses (they shouldn't be too spammy,
    # but one never knows)
    log = _log.setup_logging('shade.fnmatch')
    if name_or_id:
        # name_or_id might already be unicode
        name_or_id = _make_unicode(name_or_id)
        identifier_matches = []
        bad_pattern = False
        try:
            fn_reg = re.compile(fnmatch.translate(name_or_id))
        except sre_constants.error:
            # If the fnmatch re doesn't compile, then we don't care,
            # but log it in case the user DID pass a pattern but did
            # it poorly and wants to know what went wrong with their
            # search
            fn_reg = None
        for e in data:
            e_id = _make_unicode(e.get('id', None))
            e_name = _make_unicode(e.get('name', None))

            if ((e_id and e_id == name_or_id) or
                    (e_name and e_name == name_or_id)):
                identifier_matches.append(e)
            else:
                # Only try fnmatch if we don't match exactly
                if not fn_reg:
                    # If we don't have a pattern, skip this, but set the flag
                    # so that we log the bad pattern
                    bad_pattern = True
                    continue
                if ((e_id and fn_reg.match(e_id)) or
                        (e_name and fn_reg.match(e_name))):
                    identifier_matches.append(e)
        if not identifier_matches and bad_pattern:
            log.debug("Bad pattern passed to fnmatch", exc_info=True)
        data = identifier_matches

    if not filters:
        return data

    if isinstance(filters, six.string_types):
        return jmespath.search(filters, data)

    def _dict_filter(f, d):
        if not d:
            return False
        for key in f.keys():
            if isinstance(f[key], dict):
                if not _dict_filter(f[key], d.get(key, None)):
                    return False
            elif d.get(key, None) != f[key]:
                return False
        return True

    filtered = []
    for e in data:
        filtered.append(e)
        for key in filters.keys():
            if isinstance(filters[key], dict):
                if not _dict_filter(filters[key], e.get(key, None)):
                    filtered.pop()
                    break
            elif e.get(key, None) != filters[key]:
                filtered.pop()
                break
    return filtered


def _get_entity(cloud, resource, name_or_id, filters, **kwargs):
    """Return a single entity from the list returned by a given method.

    :param object cloud:
        The controller class (Example: the main OpenStackCloud object) .
    :param string or callable resource:
        The string that identifies the resource to use to lookup the
        get_<>_by_id or search_<resource>s methods(Example: network)
        or a callable to invoke.
    :param string name_or_id:
        The name or ID of the entity being filtered or an object or dict.
        If this is an object/dict with an 'id' attr/key, we return it and
        bypass resource lookup.
    :param filters:
        A dictionary of meta data to use for further filtering.
        OR
        A string containing a jmespath expression for further filtering.
        Example:: "[?last_name==`Smith`] | [?other.gender]==`Female`]"
    """

    # Sometimes in the control flow of shade, we already have an object
    # fetched. Rather than then needing to pull the name or id out of that
    # object, pass it in here and rely on caching to prevent us from making
    # an additional call, it's simple enough to test to see if we got an
    # object and just short-circuit return it.

    if (hasattr(name_or_id, 'id') or
       (isinstance(name_or_id, dict) and 'id' in name_or_id)):
        return name_or_id

    # If a uuid is passed short-circuit it calling the
    # get_<resorce_name>_by_id method
    if getattr(cloud, 'use_direct_get', False) and _is_uuid_like(name_or_id):
        get_resource = getattr(cloud, 'get_%s_by_id' % resource, None)
        if get_resource:
            return get_resource(name_or_id)

    search = resource if callable(resource) else getattr(
        cloud, 'search_%ss' % resource, None)
    if search:
        entities = search(name_or_id, filters, **kwargs)
        if entities:
            if len(entities) > 1:
                raise exc.OpenStackCloudException(
                    "Multiple matches found for %s" % name_or_id)
            return entities[0]
    return None


def normalize_keystone_services(services):
    """Normalize the structure of keystone services

    In keystone v2, there is a field called "service_type". In v3, it's
    "type". Just make the returned dict have both.

    :param list services: A list of keystone service dicts

    :returns: A list of normalized dicts.
    """
    ret = []
    for service in services:
        service_type = service.get('type', service.get('service_type'))
        new_service = {
            'id': service['id'],
            'name': service['name'],
            'description': service.get('description', None),
            'type': service_type,
            'service_type': service_type,
            'enabled': service['enabled']
        }
        ret.append(new_service)
    return meta.obj_list_to_munch(ret)


def localhost_supports_ipv6():
    """Determine whether the local host supports IPv6

    We look for a default route that supports the IPv6 address family,
    and assume that if it is present, this host has globally routable
    IPv6 connectivity.
    """

    try:
        return netifaces.AF_INET6 in netifaces.gateways()['default']
    except AttributeError:
        return False


def normalize_users(users):
    ret = [
        dict(
            id=user.get('id'),
            email=user.get('email'),
            name=user.get('name'),
            username=user.get('username'),
            default_project_id=user.get('default_project_id',
                                        user.get('tenantId')),
            domain_id=user.get('domain_id'),
            enabled=user.get('enabled'),
            description=user.get('description')
        ) for user in users
    ]
    return meta.obj_list_to_munch(ret)


def normalize_domains(domains):
    ret = [
        dict(
            id=domain.get('id'),
            name=domain.get('name'),
            description=domain.get('description'),
            enabled=domain.get('enabled'),
        ) for domain in domains
    ]
    return meta.obj_list_to_munch(ret)


def normalize_groups(domains):
    """Normalize Identity groups."""
    ret = [
        dict(
            id=domain.get('id'),
            name=domain.get('name'),
            description=domain.get('description'),
            domain_id=domain.get('domain_id'),
        ) for domain in domains
    ]
    return meta.obj_list_to_munch(ret)


def normalize_role_assignments(assignments):
    """Put role_assignments into a form that works with search/get interface.

    Role assignments have the structure::

        [
            {
                "role": {
                    "id": "--role-id--"
                },
                "scope": {
                    "domain": {
                        "id": "--domain-id--"
                    }
                },
                "user": {
                    "id": "--user-id--"
                }
            },
        ]

    Which is hard to work with in the rest of our interface. Map this to be::

        [
            {
                "id": "--role-id--",
                "domain": "--domain-id--",
                "user": "--user-id--",
            }
        ]

    Scope can be "domain" or "project" and "user" can also be "group".

    :param list assignments: A list of dictionaries of role assignments.

    :returns: A list of flattened/normalized role assignment dicts.
    """
    new_assignments = []
    for assignment in assignments:
        new_val = munch.Munch({'id': assignment['role']['id']})
        for scope in ('project', 'domain'):
            if scope in assignment['scope']:
                new_val[scope] = assignment['scope'][scope]['id']
        for assignee in ('user', 'group'):
            if assignee in assignment:
                new_val[assignee] = assignment[assignee]['id']
        new_assignments.append(new_val)
    return new_assignments


def normalize_flavor_accesses(flavor_accesses):
    """Normalize Flavor access list."""
    return [munch.Munch(
        dict(
            flavor_id=acl.get('flavor_id'),
            project_id=acl.get('project_id') or acl.get('tenant_id'),
        )
    ) for acl in flavor_accesses
    ]


def valid_kwargs(*valid_args):
    # This decorator checks if argument passed as **kwargs to a function are
    # present in valid_args.
    #
    # Typically, valid_kwargs is used when we want to distinguish between
    # None and omitted arguments and we still want to validate the argument
    # list.
    #
    # Example usage:
    #
    # @valid_kwargs('opt_arg1', 'opt_arg2')
    # def my_func(self, mandatory_arg1, mandatory_arg2, **kwargs):
    #   ...
    #
    @decorator
    def func_wrapper(func, *args, **kwargs):
        argspec = inspect.getargspec(func)
        for k in kwargs:
            if k not in argspec.args[1:] and k not in valid_args:
                raise TypeError(
                    "{f}() got an unexpected keyword argument "
                    "'{arg}'".format(f=inspect.stack()[1][3], arg=k))
        return func(*args, **kwargs)
    return func_wrapper


def _func_wrap(f):
    # NOTE(morgan): This extra wrapper is intended to eliminate ever
    # passing a bound method to dogpile.cache's cache_on_arguments. In
    # 0.7.0 and later it is impossible to pass bound methods to the
    # decorator. This was introduced when utilizing the decorate module in
    # lieu of a direct wrap implementation.
    @functools.wraps(f)
    def inner(*args, **kwargs):
        return f(*args, **kwargs)
    return inner


def cache_on_arguments(*cache_on_args, **cache_on_kwargs):
    _cache_name = cache_on_kwargs.pop('resource', None)

    def _inner_cache_on_arguments(func):
        def _cache_decorator(obj, *args, **kwargs):
            the_method = obj._get_cache(_cache_name).cache_on_arguments(
                *cache_on_args, **cache_on_kwargs)(
                    _func_wrap(func.__get__(obj, type(obj))))
            return the_method(*args, **kwargs)

        def invalidate(obj, *args, **kwargs):
            return obj._get_cache(
                _cache_name).cache_on_arguments()(func).invalidate(
                    *args, **kwargs)

        _cache_decorator.invalidate = invalidate
        _cache_decorator.func = func
        _decorated_methods.append(func.__name__)

        return _cache_decorator
    return _inner_cache_on_arguments


@contextlib.contextmanager
def shade_exceptions(error_message=None):
    """Context manager for dealing with shade exceptions.

    :param string error_message: String to use for the exception message
        content on non-OpenStackCloudExceptions.

    Useful for avoiding wrapping shade OpenStackCloudException exceptions
    within themselves. Code called from within the context may throw such
    exceptions without having to catch and reraise them.

    Non-OpenStackCloudException exceptions thrown within the context will
    be wrapped and the exception message will be appended to the given error
    message.
    """
    try:
        yield
    except exc.OpenStackCloudException:
        raise
    except Exception as e:
        if error_message is None:
            error_message = str(e)
        raise exc.OpenStackCloudException(error_message)


def safe_dict_min(key, data):
    """Safely find the minimum for a given key in a list of dict objects.

    This will find the minimum integer value for specific dictionary key
    across a list of dictionaries. The values for the given key MUST be
    integers, or string representations of an integer.

    The dictionary key does not have to be present in all (or any)
    of the elements/dicts within the data set.

    :param string key: The dictionary key to search for the minimum value.
    :param list data: List of dicts to use for the data set.

    :returns: None if the field was not not found in any elements, or
        the minimum value for the field otherwise.
    """
    min_value = None
    for d in data:
        if (key in d) and (d[key] is not None):
            try:
                val = int(d[key])
            except ValueError:
                raise exc.OpenStackCloudException(
                    "Search for minimum value failed. "
                    "Value for {key} is not an integer: {value}".format(
                        key=key, value=d[key])
                )
            if (min_value is None) or (val < min_value):
                min_value = val
    return min_value


def safe_dict_max(key, data):
    """Safely find the maximum for a given key in a list of dict objects.

    This will find the maximum integer value for specific dictionary key
    across a list of dictionaries. The values for the given key MUST be
    integers, or string representations of an integer.

    The dictionary key does not have to be present in all (or any)
    of the elements/dicts within the data set.

    :param string key: The dictionary key to search for the maximum value.
    :param list data: List of dicts to use for the data set.

    :returns: None if the field was not not found in any elements, or
        the maximum value for the field otherwise.
    """
    max_value = None
    for d in data:
        if (key in d) and (d[key] is not None):
            try:
                val = int(d[key])
            except ValueError:
                raise exc.OpenStackCloudException(
                    "Search for maximum value failed. "
                    "Value for {key} is not an integer: {value}".format(
                        key=key, value=d[key])
                )
            if (max_value is None) or (val > max_value):
                max_value = val
    return max_value


def _call_client_and_retry(client, url, retry_on=None,
                           call_retries=3, retry_wait=2,
                           **kwargs):
    """Method to provide retry operations.

    Some APIs utilize HTTP errors on certian operations to indicate that
    the resource is presently locked, and as such this mechanism provides
    the ability to retry upon known error codes.

    :param object client: The client method, such as:
                          ``self.baremetal_client.post``
    :param string url: The URL to perform the operation upon.
    :param integer retry_on: A list of error codes that can be retried on.
                             The method also supports a single integer to be
                             defined.
    :param integer call_retries: The number of times to retry the call upon
                                 the error code defined by the 'retry_on'
                                 parameter. Default: 3
    :param integer retry_wait: The time in seconds to wait between retry
                               attempts. Default: 2

    :returns: The object returned by the client call.
    """

    # NOTE(TheJulia): This method, as of this note, does not have direct
    # unit tests, although is fairly well tested by the tests checking
    # retry logic in test_baremetal_node.py.
    log = _log.setup_logging('shade.http')

    if isinstance(retry_on, int):
        retry_on = [retry_on]

    count = 0
    while (count < call_retries):
        count += 1
        try:
            ret_val = client(url, **kwargs)
        except exc.OpenStackCloudHTTPError as e:
            if (retry_on is not None and
                    e.response.status_code in retry_on):
                log.debug('Received retryable error {err}, waiting '
                          '{wait} seconds to retry', {
                              'err': e.response.status_code,
                              'wait': retry_wait
                          })
                time.sleep(retry_wait)
                continue
            else:
                raise
        # Break out of the loop, since the loop should only continue
        # when we encounter a known connection error.
        return ret_val


def parse_range(value):
    """Parse a numerical range string.

    Breakdown a range expression into its operater and numerical parts.
    This expression must be a string. Valid values must be an integer string,
    optionally preceeded by one of the following operators::

        - "<"  : Less than
        - ">"  : Greater than
        - "<=" : Less than or equal to
        - ">=" : Greater than or equal to

    Some examples of valid values and function return values::

        - "1024"  : returns (None, 1024)
        - "<5"    : returns ("<", 5)
        - ">=100" : returns (">=", 100)

    :param string value: The range expression to be parsed.

    :returns: A tuple with the operator string (or None if no operator
        was given) and the integer value. None is returned if parsing failed.
    """
    if value is None:
        return None

    range_exp = re.match('(<|>|<=|>=){0,1}(\d+)$', value)
    if range_exp is None:
        return None

    op = range_exp.group(1)
    num = int(range_exp.group(2))
    return (op, num)


def range_filter(data, key, range_exp):
    """Filter a list by a single range expression.

    :param list data: List of dictionaries to be searched.
    :param string key: Key name to search within the data set.
    :param string range_exp: The expression describing the range of values.

    :returns: A list subset of the original data set.
    :raises: OpenStackCloudException on invalid range expressions.
    """
    filtered = []
    range_exp = str(range_exp).upper()

    if range_exp == "MIN":
        key_min = safe_dict_min(key, data)
        if key_min is None:
            return []
        for d in data:
            if int(d[key]) == key_min:
                filtered.append(d)
        return filtered
    elif range_exp == "MAX":
        key_max = safe_dict_max(key, data)
        if key_max is None:
            return []
        for d in data:
            if int(d[key]) == key_max:
                filtered.append(d)
        return filtered

    # Not looking for a min or max, so a range or exact value must
    # have been supplied.
    val_range = parse_range(range_exp)

    # If parsing the range fails, it must be a bad value.
    if val_range is None:
        raise exc.OpenStackCloudException(
            "Invalid range value: {value}".format(value=range_exp))

    op = val_range[0]
    if op:
        # Range matching
        for d in data:
            d_val = int(d[key])
            if op == '<':
                if d_val < val_range[1]:
                    filtered.append(d)
            elif op == '>':
                if d_val > val_range[1]:
                    filtered.append(d)
            elif op == '<=':
                if d_val <= val_range[1]:
                    filtered.append(d)
            elif op == '>=':
                if d_val >= val_range[1]:
                    filtered.append(d)
        return filtered
    else:
        # Exact number match
        for d in data:
            if int(d[key]) == val_range[1]:
                filtered.append(d)
        return filtered


def generate_patches_from_kwargs(operation, **kwargs):
    """Given a set of parameters, returns a list with the
    valid patch values.

    :param string operation: The operation to perform.
    :param list kwargs: Dict of parameters.

    :returns: A list with the right patch values.
    """
    patches = []
    for k, v in kwargs.items():
        patch = {'op': operation,
                 'value': v,
                 'path': '/%s' % k}
        patches.append(patch)
    return sorted(patches)


class FileSegment(object):
    """File-like object to pass to requests."""

    def __init__(self, filename, offset, length):
        self.filename = filename
        self.offset = offset
        self.length = length
        self.pos = 0
        self._file = open(filename, 'rb')
        self.seek(0)

    def tell(self):
        return self._file.tell() - self.offset

    def seek(self, offset, whence=0):
        if whence == 0:
            self._file.seek(self.offset + offset, whence)
        elif whence == 1:
            self._file.seek(offset, whence)
        elif whence == 2:
            self._file.seek(self.offset + self.length - offset, 0)

    def read(self, size=-1):
        remaining = self.length - self.pos
        if remaining <= 0:
            return b''

        to_read = remaining if size < 0 else min(size, remaining)
        chunk = self._file.read(to_read)
        self.pos += len(chunk)

        return chunk

    def reset(self):
        self._file.seek(self.offset, 0)


def _format_uuid_string(string):
    return (string.replace('urn:', '')
                  .replace('uuid:', '')
                  .strip('{}')
                  .replace('-', '')
                  .lower())


def _is_uuid_like(val):
    """Returns validation of a value as a UUID.

    :param val: Value to verify
    :type val: string
    :returns: bool

    .. versionchanged:: 1.1.1
       Support non-lowercase UUIDs.
    """
    try:
        return str(uuid.UUID(val)).replace('-', '') == _format_uuid_string(val)
    except (TypeError, ValueError, AttributeError):
        return False