base.py

import copy
import itertools
import re
import warnings
from collections.abc import Mapping

import pymongo
import pymongo.errors
from bson import SON, json_util
from bson.code import Code
from pymongo.collection import ReturnDocument
from pymongo.common import validate_read_preference
from pymongo.read_concern import ReadConcern

from mongoengine import signals
from mongoengine.base import get_document
from mongoengine.common import _import_class
from mongoengine.connection import get_db
from mongoengine.context_managers import (
    set_read_write_concern,
    set_write_concern,
    switch_db,
)
from mongoengine.errors import (
    BulkWriteError,
    InvalidQueryError,
    LookUpError,
    NotUniqueError,
    OperationError,
)
from mongoengine.pymongo_support import count_documents
from mongoengine.queryset import transform
from mongoengine.queryset.field_list import QueryFieldList
from mongoengine.queryset.visitor import Q, QNode

__all__ = ("BaseQuerySet", "DO_NOTHING", "NULLIFY", "CASCADE", "DENY", "PULL")

# Delete rules
DO_NOTHING = 0
NULLIFY = 1
CASCADE = 2
DENY = 3
PULL = 4


class BaseQuerySet:
    """A set of results returned from a query. Wraps a MongoDB cursor,
    providing :class:`~mongoengine.Document` objects as the results.
    """

    __dereference = False
    _auto_dereference = True

    def __init__(self, document, collection):
        self._document = document
        self._collection_obj = collection
        self._mongo_query = None
        self._query_obj = Q()
        self._cls_query = {}
        self._where_clause = None
        self._loaded_fields = QueryFieldList()
        self._ordering = None
        self._snapshot = False
        self._timeout = True
        self._allow_disk_use = False
        self._read_preference = None
        self._read_concern = None
        self._iter = False
        self._scalar = []
        self._none = False
        self._as_pymongo = False
        self._search_text = None

        # If inheritance is allowed, only return instances and instances of
        # subclasses of the class being used
        if document._meta.get("allow_inheritance") is True:
            if len(self._document._subclasses) == 1:
                self._cls_query = {"_cls": self._document._subclasses[0]}
            else:
                self._cls_query = {"_cls": {"$in": self._document._subclasses}}
            self._loaded_fields = QueryFieldList(always_include=["_cls"])

        self._cursor_obj = None
        self._limit = None
        self._skip = None

        self._hint = -1  # Using -1 as None is a valid value for hint
        self._collation = None
        self._batch_size = None
        self._max_time_ms = None
        self._comment = None

        # Hack - As people expect cursor[5:5] to return
        # an empty result set. It's hard to do that right, though, because the
        # server uses limit(0) to mean 'no limit'. So we set _empty
        # in that case and check for it when iterating. We also unset
        # it anytime we change _limit. Inspired by how it is done in pymongo.Cursor
        self._empty = False

    def __call__(self, q_obj=None, **query):
        """Filter the selected documents by calling the
        :class:`~mongoengine.queryset.QuerySet` with a query.

        :param q_obj: a :class:`~mongoengine.queryset.Q` object to be used in
            the query; the :class:`~mongoengine.queryset.QuerySet` is filtered
            multiple times with different :class:`~mongoengine.queryset.Q`
            objects, only the last one will be used.
        :param query: Django-style query keyword arguments.
        """
        query = Q(**query)
        if q_obj:
            # Make sure proper query object is passed.
            if not isinstance(q_obj, QNode):
                msg = (
                    "Not a query object: %s. "
                    "Did you intend to use key=value?" % q_obj
                )
                raise InvalidQueryError(msg)
            query &= q_obj

        queryset = self.clone()
        queryset._query_obj &= query
        queryset._mongo_query = None
        queryset._cursor_obj = None

        return queryset

    def __getstate__(self):
        """
        Need for pickling queryset

        See https://github.com/MongoEngine/mongoengine/issues/442
        """

        obj_dict = self.__dict__.copy()

        # don't picke collection, instead pickle collection params
        obj_dict.pop("_collection_obj")

        # don't pickle cursor
        obj_dict["_cursor_obj"] = None

        return obj_dict

    def __setstate__(self, obj_dict):
        """
        Need for pickling queryset

        See https://github.com/MongoEngine/mongoengine/issues/442
        """

        obj_dict["_collection_obj"] = obj_dict["_document"]._get_collection()

        # update attributes
        self.__dict__.update(obj_dict)

        # forse load cursor
        # self._cursor

    def __getitem__(self, key):
        """Return a document instance corresponding to a given index if
        the key is an integer. If the key is a slice, translate its
        bounds into a skip and a limit, and return a cloned queryset
        with that skip/limit applied. For example:

        >>> User.objects[0]
        <User: User object>
        >>> User.objects[1:3]
        [<User: User object>, <User: User object>]
        """
        queryset = self.clone()
        queryset._empty = False

        # Handle a slice
        if isinstance(key, slice):
            queryset._cursor_obj = queryset._cursor[key]
            queryset._skip, queryset._limit = key.start, key.stop
            if key.start and key.stop:
                queryset._limit = key.stop - key.start
            if queryset._limit == 0:
                queryset._empty = True

            # Allow further QuerySet modifications to be performed
            return queryset

        # Handle an index
        elif isinstance(key, int):
            if queryset._scalar:
                return queryset._get_scalar(
                    queryset._document._from_son(
                        queryset._cursor[key],
                        _auto_dereference=self._auto_dereference,
                    )
                )

            if queryset._as_pymongo:
                return queryset._cursor[key]

            return queryset._document._from_son(
                queryset._cursor[key],
                _auto_dereference=self._auto_dereference,
            )

        raise TypeError("Provide a slice or an integer index")

    def __iter__(self):
        raise NotImplementedError

    def _has_data(self):
        """Return True if cursor has any data."""
        queryset = self.order_by()
        return False if queryset.first() is None else True

    def __bool__(self):
        """Avoid to open all records in an if stmt in Py3."""
        return self._has_data()

    # Core functions

    def all(self):
        """Returns a copy of the current QuerySet."""
        return self.__call__()

    def filter(self, *q_objs, **query):
        """An alias of :meth:`~mongoengine.queryset.QuerySet.__call__`"""
        return self.__call__(*q_objs, **query)

    def search_text(self, text, language=None):
        """
        Start a text search, using text indexes.
        Require: MongoDB server version 2.6+.

        :param language:  The language that determines the list of stop words
            for the search and the rules for the stemmer and tokenizer.
            If not specified, the search uses the default language of the index.
            For supported languages, see
            `Text Search Languages <http://docs.mongodb.org/manual/reference/text-search-languages/#text-search-languages>`.
        """
        queryset = self.clone()
        if queryset._search_text:
            raise OperationError("It is not possible to use search_text two times.")

        query_kwargs = SON({"$search": text})
        if language:
            query_kwargs["$language"] = language

        queryset._query_obj &= Q(__raw__={"$text": query_kwargs})
        queryset._mongo_query = None
        queryset._cursor_obj = None
        queryset._search_text = text

        return queryset

    def get(self, *q_objs, **query):
        """Retrieve the the matching object raising
        :class:`~mongoengine.queryset.MultipleObjectsReturned` or
        `DocumentName.MultipleObjectsReturned` exception if multiple results
        and :class:`~mongoengine.queryset.DoesNotExist` or
        `DocumentName.DoesNotExist` if no results are found.
        """
        queryset = self.clone()
        queryset = queryset.order_by().limit(2)
        queryset = queryset.filter(*q_objs, **query)

        try:
            result = next(queryset)
        except StopIteration:
            msg = "%s matching query does not exist." % queryset._document._class_name
            raise queryset._document.DoesNotExist(msg)

        try:
            # Check if there is another match
            next(queryset)
        except StopIteration:
            return result

        # If we were able to retrieve the 2nd doc, raise the MultipleObjectsReturned exception.
        raise queryset._document.MultipleObjectsReturned(
            "2 or more items returned, instead of 1"
        )

    def create(self, **kwargs):
        """Create new object. Returns the saved object instance."""
        return self._document(**kwargs).save(force_insert=True)

    def first(self):
        """Retrieve the first object matching the query."""
        queryset = self.clone()
        try:
            result = queryset[0]
        except IndexError:
            result = None
        return result

    def insert(
        self,
        doc_or_docs,
        load_bulk=True,
        write_concern=None,
        signal_kwargs=None,
        ordered=True,
    ):
        """bulk insert documents

        :param doc_or_docs: a document or list of documents to be inserted
        :param load_bulk (optional): If True returns the list of document
            instances
        :param write_concern: Extra keyword arguments are passed down to
                :meth:`~pymongo.collection.Collection.insert`
                which will be used as options for the resultant
                ``getLastError`` command.  For example,
                ``insert(..., {w: 2, fsync: True})`` will wait until at least
                two servers have recorded the write and will force an fsync on
                each server being written to.
        :param signal_kwargs: (optional) kwargs dictionary to be passed to
            the signal calls.
        :param ordered (optional): If True (the default) documents will be
            inserted on the server serially, in the order provided. If an error
            occurs all remaining inserts are aborted. If False, documents will
            be inserted on the server in arbitrary order, possibly in parallel,
            and all document inserts will be attempted.

        By default returns document instances, set ``load_bulk`` to False to
        return just ``ObjectIds``
        """
        Document = _import_class("Document")

        if write_concern is None:
            write_concern = {}

        docs = doc_or_docs
        return_one = False
        if isinstance(docs, Document) or issubclass(docs.__class__, Document):
            return_one = True
            docs = [docs]

        for doc in docs:
            if not isinstance(doc, self._document):
                msg = "Some documents inserted aren't instances of %s" % str(
                    self._document
                )
                raise OperationError(msg)
            if doc.pk and not doc._created:
                msg = "Some documents have ObjectIds, use doc.update() instead"
                raise OperationError(msg)

        signal_kwargs = signal_kwargs or {}
        signals.pre_bulk_insert.send(self._document, documents=docs, **signal_kwargs)

        raw = [doc.to_mongo() for doc in docs]

        with set_write_concern(self._collection, write_concern) as collection:
            insert_func = collection.insert_many
            insert_func_kwargs = {"ordered": ordered}
            if return_one:
                raw = raw[0]
                insert_func = collection.insert_one
                insert_func_kwargs = {}

        try:
            inserted_result = insert_func(raw, **insert_func_kwargs)
            ids = (
                [inserted_result.inserted_id]
                if return_one
                else inserted_result.inserted_ids
            )
        except pymongo.errors.DuplicateKeyError as err:
            message = "Could not save document (%s)"
            raise NotUniqueError(message % err)
        except pymongo.errors.BulkWriteError as err:
            # inserting documents that already have an _id field will
            # give huge performance debt or raise
            if ordered:
                inserted = err.details["nInserted"]
                for doc, raw_doc in zip(docs[:inserted], raw[:inserted]):
                    doc.pk = raw_doc["_id"]
            else:
                not_writed_ids = [
                    error["op"]["_id"] for error in err.details["writeErrors"]
                ]
                for doc, raw_doc in zip(docs, raw):
                    if raw_doc["_id"] not in not_writed_ids:
                        doc.pk = raw_doc["_id"]
            message = "Bulk write error: (%s)"
            raise BulkWriteError(message % err.details)
        except pymongo.errors.OperationFailure as err:
            message = "Could not save document (%s)"
            if re.match("^E1100[01] duplicate key", str(err)):
                # E11000 - duplicate key error index
                # E11001 - duplicate key on update
                message = "Tried to save duplicate unique keys (%s)"
                raise NotUniqueError(message % err)
            raise OperationError(message % err)

        # Apply inserted_ids to documents
        for doc, doc_id in zip(docs, ids):
            doc.pk = doc_id

        if not load_bulk:
            signals.post_bulk_insert.send(
                self._document, documents=docs, loaded=False, **signal_kwargs
            )
            return ids[0] if return_one else ids

        documents = self.in_bulk(ids)
        results = [documents.get(obj_id) for obj_id in ids]
        signals.post_bulk_insert.send(
            self._document, documents=results, loaded=True, **signal_kwargs
        )
        return results[0] if return_one else results

    def count(self, with_limit_and_skip=False):
        """Count the selected elements in the query.

        :param with_limit_and_skip (optional): take any :meth:`limit` or
            :meth:`skip` that has been applied to this cursor into account when
            getting the count
        """
        # mimic the fact that setting .limit(0) in pymongo sets no limit
        # https://docs.mongodb.com/manual/reference/method/cursor.limit/#zero-value
        if (
            self._limit == 0
            and with_limit_and_skip is False
            or self._none
            or self._empty
        ):
            return 0

        kwargs = (
            {"limit": self._limit, "skip": self._skip} if with_limit_and_skip else {}
        )

        if self._limit == 0:
            # mimic the fact that historically .limit(0) sets no limit
            kwargs.pop("limit", None)

        if self._hint not in (-1, None):
            kwargs["hint"] = self._hint

        if self._collation:
            kwargs["collation"] = self._collation

        count = count_documents(
            collection=self._cursor.collection,
            filter=self._query,
            **kwargs,
        )

        self._cursor_obj = None
        return count

    def delete(self, write_concern=None, _from_doc_delete=False, cascade_refs=None):
        """Delete the documents matched by the query.

        :param write_concern: Extra keyword arguments are passed down which
            will be used as options for the resultant
            ``getLastError`` command.  For example,
            ``save(..., write_concern={w: 2, fsync: True}, ...)`` will
            wait until at least two servers have recorded the write and
            will force an fsync on the primary server.
        :param _from_doc_delete: True when called from document delete therefore
            signals will have been triggered so don't loop.

        :returns number of deleted documents
        """
        queryset = self.clone()
        doc = queryset._document

        if write_concern is None:
            write_concern = {}

        # Handle deletes where skips or limits have been applied or
        # there is an untriggered delete signal
        has_delete_signal = signals.signals_available and (
            signals.pre_delete.has_receivers_for(doc)
            or signals.post_delete.has_receivers_for(doc)
        )

        call_document_delete = (
            queryset._skip or queryset._limit or has_delete_signal
        ) and not _from_doc_delete

        if call_document_delete:
            cnt = 0
            for doc in queryset:
                doc.delete(**write_concern)
                cnt += 1
            return cnt

        delete_rules = doc._meta.get("delete_rules") or {}
        delete_rules = list(delete_rules.items())

        # Check for DENY rules before actually deleting/nullifying any other
        # references
        for rule_entry, rule in delete_rules:
            document_cls, field_name = rule_entry
            if document_cls._meta.get("abstract"):
                continue

            if rule == DENY:
                refs = document_cls.objects(**{field_name + "__in": self})
                if refs.limit(1).count() > 0:
                    raise OperationError(
                        "Could not delete document (%s.%s refers to it)"
                        % (document_cls.__name__, field_name)
                    )

        # Check all the other rules
        for rule_entry, rule in delete_rules:
            document_cls, field_name = rule_entry
            if document_cls._meta.get("abstract"):
                continue

            if rule == CASCADE:
                cascade_refs = set() if cascade_refs is None else cascade_refs
                # Handle recursive reference
                if doc._collection == document_cls._collection:
                    for ref in queryset:
                        cascade_refs.add(ref.id)
                refs = document_cls.objects(
                    **{field_name + "__in": self, "pk__nin": cascade_refs}
                )
                if refs.count() > 0:
                    refs.delete(write_concern=write_concern, cascade_refs=cascade_refs)
            elif rule == NULLIFY:
                document_cls.objects(**{field_name + "__in": self}).update(
                    write_concern=write_concern, **{"unset__%s" % field_name: 1}
                )
            elif rule == PULL:
                document_cls.objects(**{field_name + "__in": self}).update(
                    write_concern=write_concern, **{"pull_all__%s" % field_name: self}
                )

        with set_write_concern(queryset._collection, write_concern) as collection:
            result = collection.delete_many(queryset._query)

            # If we're using an unack'd write concern, we don't really know how
            # many items have been deleted at this point, hence we only return
            # the count for ack'd ops.
            if result.acknowledged:
                return result.deleted_count

    def update(
        self,
        upsert=False,
        multi=True,
        write_concern=None,
        read_concern=None,
        full_result=False,
        **update,
    ):
        """Perform an atomic update on the fields matched by the query.

        :param upsert: insert if document doesn't exist (default ``False``)
        :param multi: Update multiple documents.
        :param write_concern: Extra keyword arguments are passed down which
            will be used as options for the resultant
            ``getLastError`` command.  For example,
            ``save(..., write_concern={w: 2, fsync: True}, ...)`` will
            wait until at least two servers have recorded the write and
            will force an fsync on the primary server.
        :param read_concern: Override the read concern for the operation
        :param full_result: Return the associated ``pymongo.UpdateResult`` rather than just the number
            updated items
        :param update: Django-style update keyword arguments

        :returns the number of updated documents (unless ``full_result`` is True)
        """
        if not update and not upsert:
            raise OperationError("No update parameters, would remove data")

        if write_concern is None:
            write_concern = {}

        queryset = self.clone()
        query = queryset._query
        update = transform.update(queryset._document, **update)

        # If doing an atomic upsert on an inheritable class
        # then ensure we add _cls to the update operation
        if upsert and "_cls" in query:
            if "$set" in update:
                update["$set"]["_cls"] = queryset._document._class_name
            else:
                update["$set"] = {"_cls": queryset._document._class_name}
        try:
            with set_read_write_concern(
                queryset._collection, write_concern, read_concern
            ) as collection:
                update_func = collection.update_one
                if multi:
                    update_func = collection.update_many
                result = update_func(query, update, upsert=upsert)
            if full_result:
                return result
            elif result.raw_result:
                return result.raw_result["n"]
        except pymongo.errors.DuplicateKeyError as err:
            raise NotUniqueError("Update failed (%s)" % err)
        except pymongo.errors.OperationFailure as err:
            if str(err) == "multi not coded yet":
                message = "update() method requires MongoDB 1.1.3+"
                raise OperationError(message)
            raise OperationError("Update failed (%s)" % err)

    def upsert_one(self, write_concern=None, read_concern=None, **update):
        """Overwrite or add the first document matched by the query.

        :param write_concern: Extra keyword arguments are passed down which
            will be used as options for the resultant
            ``getLastError`` command.  For example,
            ``save(..., write_concern={w: 2, fsync: True}, ...)`` will
            wait until at least two servers have recorded the write and
            will force an fsync on the primary server.
        :param read_concern: Override the read concern for the operation
        :param update: Django-style update keyword arguments

        :returns the new or overwritten document
        """

        atomic_update = self.update(
            multi=False,
            upsert=True,
            write_concern=write_concern,
            read_concern=read_concern,
            full_result=True,
            **update,
        )

        if atomic_update.raw_result["updatedExisting"]:
            document = self.get()
        else:
            document = self._document.objects.with_id(atomic_update.upserted_id)
        return document

    def update_one(self, upsert=False, write_concern=None, full_result=False, **update):
        """Perform an atomic update on the fields of the first document
        matched by the query.

        :param upsert: insert if document doesn't exist (default ``False``)
        :param write_concern: Extra keyword arguments are passed down which
            will be used as options for the resultant
            ``getLastError`` command.  For example,
            ``save(..., write_concern={w: 2, fsync: True}, ...)`` will
            wait until at least two servers have recorded the write and
            will force an fsync on the primary server.
        :param full_result: Return the associated ``pymongo.UpdateResult`` rather than just the number
            updated items
        :param update: Django-style update keyword arguments
            full_result
        :returns the number of updated documents (unless ``full_result`` is True)
        """
        return self.update(
            upsert=upsert,
            multi=False,
            write_concern=write_concern,
            full_result=full_result,
            **update,
        )

    def modify(
        self, upsert=False, full_response=False, remove=False, new=False, **update
    ):
        """Update and return the updated document.

        Returns either the document before or after modification based on `new`
        parameter. If no documents match the query and `upsert` is false,
        returns ``None``. If upserting and `new` is false, returns ``None``.

        If the full_response parameter is ``True``, the return value will be
        the entire response object from the server, including the 'ok' and
        'lastErrorObject' fields, rather than just the modified document.
        This is useful mainly because the 'lastErrorObject' document holds
        information about the command's execution.

        :param upsert: insert if document doesn't exist (default ``False``)
        :param full_response: return the entire response object from the
            server (default ``False``, not available for PyMongo 3+)
        :param remove: remove rather than updating (default ``False``)
        :param new: return updated rather than original document
            (default ``False``)
        :param update: Django-style update keyword arguments
        """

        if remove and new:
            raise OperationError("Conflicting parameters: remove and new")

        if not update and not upsert and not remove:
            raise OperationError("No update parameters, must either update or remove")

        queryset = self.clone()
        query = queryset._query
        if not remove:
            update = transform.update(queryset._document, **update)
        sort = queryset._ordering

        try:
            if full_response:
                msg = "With PyMongo 3+, it is not possible anymore to get the full response."
                warnings.warn(msg, DeprecationWarning)
            if remove:
                result = queryset._collection.find_one_and_delete(
                    query, sort=sort, **self._cursor_args
                )
            else:
                if new:
                    return_doc = ReturnDocument.AFTER
                else:
                    return_doc = ReturnDocument.BEFORE
                result = queryset._collection.find_one_and_update(
                    query,
                    update,
                    upsert=upsert,
                    sort=sort,
                    return_document=return_doc,
                    **self._cursor_args,
                )
        except pymongo.errors.DuplicateKeyError as err:
            raise NotUniqueError("Update failed (%s)" % err)
        except pymongo.errors.OperationFailure as err:
            raise OperationError("Update failed (%s)" % err)

        if full_response:
            if result["value"] is not None:
                result["value"] = self._document._from_son(result["value"])
        else:
            if result is not None:
                result = self._document._from_son(result)

        return result

    def with_id(self, object_id):
        """Retrieve the object matching the id provided.  Uses `object_id` only
        and raises InvalidQueryError if a filter has been applied. Returns
        `None` if no document exists with that id.

        :param object_id: the value for the id of the document to look up
        """
        queryset = self.clone()
        if not queryset._query_obj.empty:
            msg = "Cannot use a filter whilst using `with_id`"
            raise InvalidQueryError(msg)
        return queryset.filter(pk=object_id).first()

    def in_bulk(self, object_ids):
        """ "Retrieve a set of documents by their ids.

        :param object_ids: a list or tuple of ObjectId's
        :rtype: dict of ObjectId's as keys and collection-specific
                Document subclasses as values.
        """
        doc_map = {}

        docs = self._collection.find({"_id": {"$in": object_ids}}, **self._cursor_args)
        if self._scalar:
            for doc in docs:
                doc_map[doc["_id"]] = self._get_scalar(self._document._from_son(doc))
        elif self._as_pymongo:
            for doc in docs:
                doc_map[doc["_id"]] = doc
        else:
            for doc in docs:
                doc_map[doc["_id"]] = self._document._from_son(
                    doc,
                    _auto_dereference=self._auto_dereference,
                )

        return doc_map

    def none(self):
        """Returns a queryset that never returns any objects and no query will be executed when accessing the results
        inspired by django none() https://docs.djangoproject.com/en/dev/ref/models/querysets/#none
        """
        queryset = self.clone()
        queryset._none = True
        return queryset

    def no_sub_classes(self):
        """Filter for only the instances of this specific document.

        Do NOT return any inherited documents.
        """
        if self._document._meta.get("allow_inheritance") is True:
            self._cls_query = {"_cls": self._document._class_name}

        return self

    def using(self, alias):
        """This method is for controlling which database the QuerySet will be
        evaluated against if you are using more than one database.

        :param alias: The database alias
        """

        with switch_db(self._document, alias) as cls:
            collection = cls._get_collection()

        return self._clone_into(self.__class__(self._document, collection))

    def clone(self):
        """Create a copy of the current queryset."""
        return self._clone_into(self.__class__(self._document, self._collection_obj))

    def _clone_into(self, new_qs):
        """Copy all of the relevant properties of this queryset to
        a new queryset (which has to be an instance of
        :class:`~mongoengine.queryset.base.BaseQuerySet`).
        """
        if not isinstance(new_qs, BaseQuerySet):
            raise OperationError(
                "%s is not a subclass of BaseQuerySet" % new_qs.__name__
            )

        copy_props = (
            "_mongo_query",
            "_cls_query",
            "_none",
            "_query_obj",
            "_where_clause",
            "_loaded_fields",
            "_ordering",
            "_snapshot",
            "_timeout",
            "_allow_disk_use",
            "_read_preference",
            "_read_concern",
            "_iter",
            "_scalar",
            "_as_pymongo",
            "_limit",
            "_skip",
            "_empty",
            "_hint",
            "_collation",
            "_auto_dereference",
            "_search_text",
            "_max_time_ms",
            "_comment",
            "_batch_size",
        )

        for prop in copy_props:
            val = getattr(self, prop)
            setattr(new_qs, prop, copy.copy(val))

        if self._cursor_obj:
            new_qs._cursor_obj = self._cursor_obj.clone()

        return new_qs

    def select_related(self, max_depth=1):
        """Handles dereferencing of :class:`~bson.dbref.DBRef` objects or
        :class:`~bson.object_id.ObjectId` a maximum depth in order to cut down
        the number queries to mongodb.
        """
        # Make select related work the same for querysets
        max_depth += 1
        queryset = self.clone()
        return queryset._dereference(queryset, max_depth=max_depth)

    def limit(self, n):
        """Limit the number of returned documents to `n`. This may also be
        achieved using array-slicing syntax (e.g. ``User.objects[:5]``).

        :param n: the maximum number of objects to return if n is greater than 0.
        When 0 is passed, returns all the documents in the cursor
        """
        queryset = self.clone()
        queryset._limit = n
        queryset._empty = False  # cancels the effect of empty

        # If a cursor object has already been created, apply the limit to it.
        if queryset._cursor_obj:
            queryset._cursor_obj.limit(queryset._limit)

        return queryset

    def skip(self, n):
        """Skip `n` documents before returning the results. This may also be
        achieved using array-slicing syntax (e.g. ``User.objects[5:]``).

        :param n: the number of objects to skip before returning results
        """
        queryset = self.clone()
        queryset._skip = n

        # If a cursor object has already been created, apply the skip to it.
        if queryset._cursor_obj:
            queryset._cursor_obj.skip(queryset._skip)

        return queryset

    def hint(self, index=None):
        """Added 'hint' support, telling Mongo the proper index to use for the
        query.

        Judicious use of hints can greatly improve query performance. When
        doing a query on multiple fields (at least one of which is indexed)
        pass the indexed field as a hint to the query.

        Hinting will not do anything if the corresponding index does not exist.
        The last hint applied to this cursor takes precedence over all others.
        """
        queryset = self.clone()
        queryset._hint = index

        # If a cursor object has already been created, apply the hint to it.
        if queryset._cursor_obj:
            queryset._cursor_obj.hint(queryset._hint)

        return queryset

    def collation(self, collation=None):
        """
        Collation allows users to specify language-specific rules for string
        comparison, such as rules for lettercase and accent marks.
        :param collation: `~pymongo.collation.Collation` or dict with
        following fields:
            {
                locale: str,
                caseLevel: bool,
                caseFirst: str,
                strength: int,
                numericOrdering: bool,
                alternate: str,
                maxVariable: str,
                backwards: str
            }
        Collation should be added to indexes like in test example
        """
        queryset = self.clone()
        queryset._collation = collation

        if queryset._cursor_obj:
            queryset._cursor_obj.collation(collation)

        return queryset

    def batch_size(self, size):
        """Limit the number of documents returned in a single batch (each
        batch requires a round trip to the server).

        See http://api.mongodb.com/python/current/api/pymongo/cursor.html#pymongo.cursor.Cursor.batch_size
        for details.

        :param size: desired size of each batch.
        """
        queryset = self.clone()
        queryset._batch_size = size

        # If a cursor object has already been created, apply the batch size to it.
        if queryset._cursor_obj:
            queryset._cursor_obj.batch_size(queryset._batch_size)

        return queryset

    def distinct(self, field):
        """Return a list of distinct values for a given field.

        :param field: the field to select distinct values from

        .. note:: This is a command and won't take ordering or limit into
           account.
        """
        queryset = self.clone()

        try:
            field = self._fields_to_dbfields([field]).pop()
        except LookUpError:
            pass

        distinct = self._dereference(
            queryset._cursor.distinct(field), 1, name=field, instance=self._document
        )

        doc_field = self._document._fields.get(field.split(".", 1)[0])
        instance = None

        # We may need to cast to the correct type eg. ListField(EmbeddedDocumentField)
        EmbeddedDocumentField = _import_class("EmbeddedDocumentField")
        ListField = _import_class("ListField")
        GenericEmbeddedDocumentField = _import_class("GenericEmbeddedDocumentField")
        if isinstance(doc_field, ListField):
            doc_field = getattr(doc_field, "field", doc_field)
        if isinstance(doc_field, (EmbeddedDocumentField, GenericEmbeddedDocumentField)):
            instance = getattr(doc_field, "document_type", None)

        # handle distinct on subdocuments
        if "." in field:
            for field_part in field.split(".")[1:]:
                # if looping on embedded document, get the document type instance
                if instance and isinstance(
                    doc_field, (EmbeddedDocumentField, GenericEmbeddedDocumentField)
                ):
                    doc_field = instance
                # now get the subdocument
                doc_field = getattr(doc_field, field_part, doc_field)
                # We may need to cast to the correct type eg. ListField(EmbeddedDocumentField)
                if isinstance(doc_field, ListField):
                    doc_field = getattr(doc_field, "field", doc_field)
                if isinstance(
                    doc_field, (EmbeddedDocumentField, GenericEmbeddedDocumentField)
                ):
                    instance = getattr(doc_field, "document_type", None)

        if instance and isinstance(
            doc_field, (EmbeddedDocumentField, GenericEmbeddedDocumentField)
        ):
            distinct = [instance(**doc) for doc in distinct]

        return distinct

    def only(self, *fields):
        """Load only a subset of this document's fields. ::

            post = BlogPost.objects(...).only('title', 'author.name')

        .. note :: `only()` is chainable and will perform a union ::
            So with the following it will fetch both: `title` and `author.name`::

                post = BlogPost.objects.only('title').only('author.name')

        :func:`~mongoengine.queryset.QuerySet.all_fields` will reset any
        field filters.

        :param fields: fields to include
        """
        fields = {f: QueryFieldList.ONLY for f in fields}
        return self.fields(True, **fields)

    def exclude(self, *fields):
        """Opposite to .only(), exclude some document's fields. ::

            post = BlogPost.objects(...).exclude('comments')

        .. note :: `exclude()` is chainable and will perform a union ::
            So with the following it will exclude both: `title` and `author.name`::

                post = BlogPost.objects.exclude('title').exclude('author.name')

        :func:`~mongoengine.queryset.QuerySet.all_fields` will reset any
        field filters.

        :param fields: fields to exclude
        """
        fields = {f: QueryFieldList.EXCLUDE for f in fields}
        return self.fields(**fields)

    def fields(self, _only_called=False, **kwargs):
        """Manipulate how you load this document's fields. Used by `.only()`
        and `.exclude()` to manipulate which fields to retrieve. If called
        directly, use a set of kwargs similar to the MongoDB projection
        document. For example:

        Include only a subset of fields:

            posts = BlogPost.objects(...).fields(author=1, title=1)

        Exclude a specific field:

            posts = BlogPost.objects(...).fields(comments=0)

        To retrieve a subrange or sublist of array elements,
        support exist for both the `slice` and `elemMatch` projection operator:

            posts = BlogPost.objects(...).fields(slice__comments=5)
            posts = BlogPost.objects(...).fields(elemMatch__comments="test")

        :param kwargs: A set of keyword arguments identifying what to
            include, exclude, or slice.
        """

        # Check for an operator and transform to mongo-style if there is
        operators = ["slice", "elemMatch"]
        cleaned_fields = []
        for key, value in kwargs.items():
            parts = key.split("__")
            if parts[0] in operators:
                op = parts.pop(0)
                value = {"$" + op: value}
            key = ".".join(parts)
            cleaned_fields.append((key, value))

        # Sort fields by their values, explicitly excluded fields first, then
        # explicitly included, and then more complicated operators such as
        # $slice.
        def _sort_key(field_tuple):
            _, value = field_tuple
            if isinstance(value, int):
                return value  # 0 for exclusion, 1 for inclusion
            return 2  # so that complex values appear last

        fields = sorted(cleaned_fields, key=_sort_key)

        # Clone the queryset, group all fields by their value, convert
        # each of them to db_fields, and set the queryset's _loaded_fields
        queryset = self.clone()
        for value, group in itertools.groupby(fields, lambda x: x[1]):
            fields = [field for field, value in group]
            fields = queryset._fields_to_dbfields(fields)
            queryset._loaded_fields += QueryFieldList(
                fields, value=value, _only_called=_only_called
            )

        return queryset

    def all_fields(self):
        """Include all fields. Reset all previously calls of .only() or
        .exclude(). ::

            post = BlogPost.objects.exclude('comments').all_fields()
        """
        queryset = self.clone()
        queryset._loaded_fields = QueryFieldList(
            always_include=queryset._loaded_fields.always_include
        )
        return queryset

    def order_by(self, *keys):
        """Order the :class:`~mongoengine.queryset.QuerySet` by the given keys.

        The order may be specified by prepending each of the keys by a "+" or
        a "-". Ascending order is assumed if there's no prefix.

        If no keys are passed, existing ordering is cleared instead.

        :param keys: fields to order the query results by; keys may be
            prefixed with "+" or a "-" to determine the ordering direction.
        """
        queryset = self.clone()

        old_ordering = queryset._ordering
        new_ordering = queryset._get_order_by(keys)

        if queryset._cursor_obj:

            # If a cursor object has already been created, apply the sort to it
            if new_ordering:
                queryset._cursor_obj.sort(new_ordering)

            # If we're trying to clear a previous explicit ordering, we need
            # to clear the cursor entirely (because PyMongo doesn't allow
            # clearing an existing sort on a cursor).
            elif old_ordering:
                queryset._cursor_obj = None

        queryset._ordering = new_ordering

        return queryset

    def clear_cls_query(self):
        """Clear the default "_cls" query.

        By default, all queries generated for documents that allow inheritance
        include an extra "_cls" clause. In most cases this is desirable, but
        sometimes you might achieve better performance if you clear that
        default query.

        Scan the code for `_cls_query` to get more details.
        """
        queryset = self.clone()
        queryset._cls_query = {}
        return queryset

    def comment(self, text):
        """Add a comment to the query.

        See https://docs.mongodb.com/manual/reference/method/cursor.comment/#cursor.comment
        for details.
        """
        return self._chainable_method("comment", text)

    def explain(self):
        """Return an explain plan record for the
        :class:`~mongoengine.queryset.QuerySet` cursor.
        """
        return self._cursor.explain()

    # DEPRECATED. Has no more impact on PyMongo 3+
    def snapshot(self, enabled):
        """Enable or disable snapshot mode when querying.

        :param enabled: whether or not snapshot mode is enabled
        """
        msg = "snapshot is deprecated as it has no impact when using PyMongo 3+."
        warnings.warn(msg, DeprecationWarning)
        queryset = self.clone()
        queryset._snapshot = enabled
        return queryset

    def allow_disk_use(self, enabled):
        """Enable or disable the use of temporary files on disk while processing a blocking sort operation.
         (To store data exceeding the 100 megabyte system memory limit)

        :param enabled: whether or not temporary files on disk are used
        """
        queryset = self.clone()
        queryset._allow_disk_use = enabled
        return queryset

    def timeout(self, enabled):
        """Enable or disable the default mongod timeout when querying. (no_cursor_timeout option)

        :param enabled: whether or not the timeout is used
        """
        queryset = self.clone()
        queryset._timeout = enabled
        return queryset

    def read_preference(self, read_preference):
        """Change the read_preference when querying.

        :param read_preference: override ReplicaSetConnection-level
            preference.
        """
        validate_read_preference("read_preference", read_preference)
        queryset = self.clone()
        queryset._read_preference = read_preference
        queryset._cursor_obj = None  # we need to re-create the cursor object whenever we apply read_preference
        return queryset

    def read_concern(self, read_concern):
        """Change the read_concern when querying.

        :param read_concern: override ReplicaSetConnection-level
            preference.
        """
        if read_concern is not None and not isinstance(read_concern, Mapping):
            raise TypeError(f"{read_concern!r} is not a valid read concern.")

        queryset = self.clone()
        queryset._read_concern = (
            ReadConcern(**read_concern) if read_concern is not None else None
        )
        queryset._cursor_obj = None  # we need to re-create the cursor object whenever we apply read_concern
        return queryset

    def scalar(self, *fields):
        """Instead of returning Document instances, return either a specific
        value or a tuple of values in order.

        Can be used along with
        :func:`~mongoengine.queryset.QuerySet.no_dereference` to turn off
        dereferencing.

        .. note:: This effects all results and can be unset by calling
                  ``scalar`` without arguments. Calls ``only`` automatically.

        :param fields: One or more fields to return instead of a Document.
        """
        queryset = self.clone()
        queryset._scalar = list(fields)

        if fields:
            queryset = queryset.only(*fields)
        else:
            queryset = queryset.all_fields()

        return queryset

    def values_list(self, *fields):
        """An alias for scalar"""
        return self.scalar(*fields)

    def as_pymongo(self):
        """Instead of returning Document instances, return raw values from
        pymongo.

        This method is particularly useful if you don't need dereferencing
        and care primarily about the speed of data retrieval.
        """
        queryset = self.clone()
        queryset._as_pymongo = True
        return queryset

    def max_time_ms(self, ms):
        """Wait `ms` milliseconds before killing the query on the server

        :param ms: the number of milliseconds before killing the query on the server
        """
        return self._chainable_method("max_time_ms", ms)

    # JSON Helpers

    def to_json(self, *args, **kwargs):
        """Converts a queryset to JSON"""
        return json_util.dumps(self.as_pymongo(), *args, **kwargs)

    def from_json(self, json_data):
        """Converts json data to unsaved objects"""
        son_data = json_util.loads(json_data)
        return [self._document._from_son(data) for data in son_data]

    def aggregate(self, pipeline, *suppl_pipeline, **kwargs):
        """Perform a aggregate function based in your queryset params

        :param pipeline: list of aggregation commands,\
            see: http://docs.mongodb.org/manual/core/aggregation-pipeline/
        :param suppl_pipeline: unpacked list of pipeline (added to support deprecation of the old interface)
            parameter will be removed shortly
        :param kwargs: (optional) kwargs dictionary to be passed to pymongo's aggregate call
            See https://api.mongodb.com/python/current/api/pymongo/collection.html#pymongo.collection.Collection.aggregate
        """
        using_deprecated_interface = isinstance(pipeline, dict) or bool(suppl_pipeline)
        user_pipeline = [pipeline] if isinstance(pipeline, dict) else list(pipeline)

        if using_deprecated_interface:
            msg = "Calling .aggregate() with un unpacked list (*pipeline) is deprecated, it will soon change and will expect a list (similar to pymongo.Collection.aggregate interface), see documentation"
            warnings.warn(msg, DeprecationWarning)

        user_pipeline += suppl_pipeline

        initial_pipeline = []
        if self._query:
            initial_pipeline.append({"$match": self._query})

        if self._ordering:
            initial_pipeline.append({"$sort": dict(self._ordering)})

        if self._limit is not None:
            # As per MongoDB Documentation (https://docs.mongodb.com/manual/reference/operator/aggregation/limit/),
            # keeping limit stage right after sort stage is more efficient. But this leads to wrong set of documents
            # for a skip stage that might succeed these. So we need to maintain more documents in memory in such a
            # case (https://stackoverflow.com/a/24161461).
            initial_pipeline.append({"$limit": self._limit + (self._skip or 0)})

        if self._skip is not None:
            initial_pipeline.append({"$skip": self._skip})

        final_pipeline = initial_pipeline + user_pipeline

        collection = self._collection
        if self._read_preference is not None or self._read_concern is not None:
            collection = self._collection.with_options(
                read_preference=self._read_preference, read_concern=self._read_concern
            )

        return collection.aggregate(final_pipeline, cursor={}, **kwargs)

    # JS functionality
    def map_reduce(
        self, map_f, reduce_f, output, finalize_f=None, limit=None, scope=None
    ):
        """Perform a map/reduce query using the current query spec
        and ordering. While ``map_reduce`` respects ``QuerySet`` chaining,
        it must be the last call made, as it does not return a maleable
        ``QuerySet``.

        See the :meth:`~mongoengine.tests.QuerySetTest.test_map_reduce`
        and :meth:`~mongoengine.tests.QuerySetTest.test_map_advanced`
        tests in ``tests.queryset.QuerySetTest`` for usage examples.

        :param map_f: map function, as :class:`~bson.code.Code` or string
        :param reduce_f: reduce function, as
                         :class:`~bson.code.Code` or string
        :param output: output collection name, if set to 'inline' will try to
           use :class:`~pymongo.collection.Collection.inline_map_reduce`
           This can also be a dictionary containing output options
           see: http://docs.mongodb.org/manual/reference/command/mapReduce/#dbcmd.mapReduce
        :param finalize_f: finalize function, an optional function that
                           performs any post-reduction processing.
        :param scope: values to insert into map/reduce global scope. Optional.
        :param limit: number of objects from current query to provide
                      to map/reduce method

        Returns an iterator yielding
        :class:`~mongoengine.document.MapReduceDocument`.

        .. note::

            Map/Reduce changed in server version **>= 1.7.4**. The PyMongo
            :meth:`~pymongo.collection.Collection.map_reduce` helper requires
            PyMongo version **>= 1.11**.
        """
        queryset = self.clone()

        MapReduceDocument = _import_class("MapReduceDocument")

        map_f_scope = {}
        if isinstance(map_f, Code):
            map_f_scope = map_f.scope
            map_f = str(map_f)
        map_f = Code(queryset._sub_js_fields(map_f), map_f_scope or None)

        reduce_f_scope = {}
        if isinstance(reduce_f, Code):
            reduce_f_scope = reduce_f.scope
            reduce_f = str(reduce_f)
        reduce_f_code = queryset._sub_js_fields(reduce_f)
        reduce_f = Code(reduce_f_code, reduce_f_scope or None)

        mr_args = {"query": queryset._query}

        if finalize_f:
            finalize_f_scope = {}
            if isinstance(finalize_f, Code):
                finalize_f_scope = finalize_f.scope
                finalize_f = str(finalize_f)
            finalize_f_code = queryset._sub_js_fields(finalize_f)
            finalize_f = Code(finalize_f_code, finalize_f_scope or None)
            mr_args["finalize"] = finalize_f

        if scope:
            mr_args["scope"] = scope

        if limit:
            mr_args["limit"] = limit

        if output == "inline" and not queryset._ordering:
            map_reduce_function = "inline_map_reduce"
        else:
            map_reduce_function = "map_reduce"

            if isinstance(output, str):
                mr_args["out"] = output

            elif isinstance(output, dict):
                ordered_output = []

                for part in ("replace", "merge", "reduce"):
                    value = output.get(part)
                    if value:
                        ordered_output.append((part, value))
                        break

                else:
                    raise OperationError("actionData not specified for output")

                db_alias = output.get("db_alias")
                remaing_args = ["db", "sharded", "nonAtomic"]

                if db_alias:
                    ordered_output.append(("db", get_db(db_alias).name))
                    del remaing_args[0]

                for part in remaing_args:
                    value = output.get(part)
                    if value:
                        ordered_output.append((part, value))

                mr_args["out"] = SON(ordered_output)

        results = getattr(queryset._collection, map_reduce_function)(
            map_f, reduce_f, **mr_args
        )

        if map_reduce_function == "map_reduce":
            results = results.find()

        if queryset._ordering:
            results = results.sort(queryset._ordering)

        for doc in results:
            yield MapReduceDocument(
                queryset._document, queryset._collection, doc["_id"], doc["value"]
            )

    def exec_js(self, code, *fields, **options):
        """Execute a Javascript function on the server. A list of fields may be
        provided, which will be translated to their correct names and supplied
        as the arguments to the function. A few extra variables are added to
        the function's scope: ``collection``, which is the name of the
        collection in use; ``query``, which is an object representing the
        current query; and ``options``, which is an object containing any
        options specified as keyword arguments.

        As fields in MongoEngine may use different names in the database (set
        using the :attr:`db_field` keyword argument to a :class:`Field`
        constructor), a mechanism exists for replacing MongoEngine field names
        with the database field names in Javascript code. When accessing a
        field, use square-bracket notation, and prefix the MongoEngine field
        name with a tilde (~).

        :param code: a string of Javascript code to execute
        :param fields: fields that you will be using in your function, which
            will be passed in to your function as arguments
        :param options: options that you want available to the function
            (accessed in Javascript through the ``options`` object)
        """
        queryset = self.clone()

        code = queryset._sub_js_fields(code)

        fields = [queryset._document._translate_field_name(f) for f in fields]
        collection = queryset._document._get_collection_name()

        scope = {"collection": collection, "options": options or {}}

        query = queryset._query
        if queryset._where_clause:
            query["$where"] = queryset._where_clause

        scope["query"] = query
        code = Code(code, scope=scope)

        db = queryset._document._get_db()
        return db.eval(code, *fields)

    def where(self, where_clause):
        """Filter ``QuerySet`` results with a ``$where`` clause (a Javascript
        expression). Performs automatic field name substitution like
        :meth:`mongoengine.queryset.Queryset.exec_js`.

        .. note:: When using this mode of query, the database will call your
                  function, or evaluate your predicate clause, for each object
                  in the collection.
        """
        queryset = self.clone()
        where_clause = queryset._sub_js_fields(where_clause)
        queryset._where_clause = where_clause
        return queryset

    def sum(self, field):
        """Sum over the values of the specified field.

        :param field: the field to sum over; use dot notation to refer to
            embedded document fields
        """
        db_field = self._fields_to_dbfields([field]).pop()
        pipeline = [
            {"$match": self._query},
            {"$group": {"_id": "sum", "total": {"$sum": "$" + db_field}}},
        ]

        # if we're performing a sum over a list field, we sum up all the
        # elements in the list, hence we need to $unwind the arrays first
        ListField = _import_class("ListField")
        field_parts = field.split(".")
        field_instances = self._document._lookup_field(field_parts)
        if isinstance(field_instances[-1], ListField):
            pipeline.insert(1, {"$unwind": "$" + field})

        result = tuple(self._document._get_collection().aggregate(pipeline))

        if result:
            return result[0]["total"]
        return 0

    def average(self, field):
        """Average over the values of the specified field.

        :param field: the field to average over; use dot notation to refer to
            embedded document fields
        """
        db_field = self._fields_to_dbfields([field]).pop()
        pipeline = [
            {"$match": self._query},
            {"$group": {"_id": "avg", "total": {"$avg": "$" + db_field}}},
        ]

        # if we're performing an average over a list field, we average out
        # all the elements in the list, hence we need to $unwind the arrays
        # first
        ListField = _import_class("ListField")
        field_parts = field.split(".")
        field_instances = self._document._lookup_field(field_parts)
        if isinstance(field_instances[-1], ListField):
            pipeline.insert(1, {"$unwind": "$" + field})

        result = tuple(self._document._get_collection().aggregate(pipeline))
        if result:
            return result[0]["total"]
        return 0

    def item_frequencies(self, field, normalize=False, map_reduce=True):
        """Returns a dictionary of all items present in a field across
        the whole queried set of documents, and their corresponding frequency.
        This is useful for generating tag clouds, or searching documents.

        .. note::

            Can only do direct simple mappings and cannot map across
            :class:`~mongoengine.fields.ReferenceField` or
            :class:`~mongoengine.fields.GenericReferenceField` for more complex
            counting a manual map reduce call is required.

        If the field is a :class:`~mongoengine.fields.ListField`, the items within
        each list will be counted individually.

        :param field: the field to use
        :param normalize: normalize the results so they add to 1.0
        :param map_reduce: Use map_reduce over exec_js
        """
        if map_reduce:
            return self._item_frequencies_map_reduce(field, normalize=normalize)
        return self._item_frequencies_exec_js(field, normalize=normalize)

    # Iterator helpers

    def __next__(self):
        """Wrap the result in a :class:`~mongoengine.Document` object."""
        if self._none or self._empty:
            raise StopIteration

        raw_doc = next(self._cursor)

        if self._as_pymongo:
            return raw_doc

        doc = self._document._from_son(
            raw_doc,
            _auto_dereference=self._auto_dereference,
        )

        if self._scalar:
            return self._get_scalar(doc)

        return doc

    def rewind(self):
        """Rewind the cursor to its unevaluated state."""
        self._iter = False
        self._cursor.rewind()

    # Properties

    @property
    def _collection(self):
        """Property that returns the collection object. This allows us to
        perform operations only if the collection is accessed.
        """
        return self._collection_obj

    @property
    def _cursor_args(self):
        fields_name = "projection"
        # snapshot is not handled at all by PyMongo 3+
        # TODO: evaluate similar possibilities using modifiers
        if self._snapshot:
            msg = "The snapshot option is not anymore available with PyMongo 3+"
            warnings.warn(msg, DeprecationWarning)

        cursor_args = {}
        if not self._timeout:
            cursor_args["no_cursor_timeout"] = True

        if self._allow_disk_use:
            cursor_args["allow_disk_use"] = True

        if self._loaded_fields:
            cursor_args[fields_name] = self._loaded_fields.as_dict()

        if self._search_text:
            if fields_name not in cursor_args:
                cursor_args[fields_name] = {}

            cursor_args[fields_name]["_text_score"] = {"$meta": "textScore"}

        return cursor_args

    @property
    def _cursor(self):
        """Return a PyMongo cursor object corresponding to this queryset."""

        # If _cursor_obj already exists, return it immediately.
        if self._cursor_obj is not None:
            return self._cursor_obj

        # Create a new PyMongo cursor.
        # XXX In PyMongo 3+, we define the read preference on a collection
        # level, not a cursor level. Thus, we need to get a cloned collection
        # object using `with_options` first.
        if self._read_preference is not None or self._read_concern is not None:
            self._cursor_obj = self._collection.with_options(
                read_preference=self._read_preference, read_concern=self._read_concern
            ).find(self._query, **self._cursor_args)
        else:
            self._cursor_obj = self._collection.find(self._query, **self._cursor_args)

        # Apply "where" clauses to cursor
        if self._where_clause:
            where_clause = self._sub_js_fields(self._where_clause)
            self._cursor_obj.where(where_clause)

        # Apply ordering to the cursor.
        # XXX self._ordering can be equal to:
        # * None if we didn't explicitly call order_by on this queryset.
        # * A list of PyMongo-style sorting tuples.
        # * An empty list if we explicitly called order_by() without any
        #   arguments. This indicates that we want to clear the default
        #   ordering.
        if self._ordering:
            # explicit ordering
            self._cursor_obj.sort(self._ordering)
        elif self._ordering is None and self._document._meta["ordering"]:
            # default ordering
            order = self._get_order_by(self._document._meta["ordering"])
            self._cursor_obj.sort(order)

        if self._limit is not None:
            self._cursor_obj.limit(self._limit)

        if self._skip is not None:
            self._cursor_obj.skip(self._skip)

        if self._hint != -1:
            self._cursor_obj.hint(self._hint)

        if self._collation is not None:
            self._cursor_obj.collation(self._collation)

        if self._batch_size is not None:
            self._cursor_obj.batch_size(self._batch_size)

        if self._comment is not None:
            self._cursor_obj.comment(self._comment)

        return self._cursor_obj

    def __deepcopy__(self, memo):
        """Essential for chained queries with ReferenceFields involved"""
        return self.clone()

    @property
    def _query(self):
        if self._mongo_query is None:
            self._mongo_query = self._query_obj.to_query(self._document)
            if self._cls_query:
                if "_cls" in self._mongo_query:
                    self._mongo_query = {"$and": [self._cls_query, self._mongo_query]}
                else:
                    self._mongo_query.update(self._cls_query)
        return self._mongo_query

    @property
    def _dereference(self):
        if not self.__dereference:
            self.__dereference = _import_class("DeReference")()
        return self.__dereference

    def no_dereference(self):
        """Turn off any dereferencing for the results of this queryset."""
        queryset = self.clone()
        queryset._auto_dereference = False
        return queryset

    # Helper Functions

    def _item_frequencies_map_reduce(self, field, normalize=False):
        map_func = """
            function() {{
                var path = '{{{{~{field}}}}}'.split('.');
                var field = this;

                for (p in path) {{
                    if (typeof field != 'undefined')
                       field = field[path[p]];
                    else
                       break;
                }}
                if (field && field.constructor == Array) {{
                    field.forEach(function(item) {{
                        emit(item, 1);
                    }});
                }} else if (typeof field != 'undefined') {{
                    emit(field, 1);
                }} else {{
                    emit(null, 1);
                }}
            }}
        """.format(
            field=field
        )
        reduce_func = """
            function(key, values) {
                var total = 0;
                var valuesSize = values.length;
                for (var i=0; i < valuesSize; i++) {
                    total += parseInt(values[i], 10);
                }
                return total;
            }
        """
        values = self.map_reduce(map_func, reduce_func, "inline")
        frequencies = {}
        for f in values:
            key = f.key
            if isinstance(key, float):
                if int(key) == key:
                    key = int(key)
            frequencies[key] = int(f.value)

        if normalize:
            count = sum(frequencies.values())
            frequencies = {k: float(v) / count for k, v in frequencies.items()}

        return frequencies

    def _item_frequencies_exec_js(self, field, normalize=False):
        """Uses exec_js to execute"""
        freq_func = """
            function(path) {
                var path = path.split('.');

                var total = 0.0;
                db[collection].find(query).forEach(function(doc) {
                    var field = doc;
                    for (p in path) {
                        if (field)
                            field = field[path[p]];
                         else
                            break;
                    }
                    if (field && field.constructor == Array) {
                       total += field.length;
                    } else {
                       total++;
                    }
                });

                var frequencies = {};
                var types = {};
                var inc = 1.0;

                db[collection].find(query).forEach(function(doc) {
                    field = doc;
                    for (p in path) {
                        if (field)
                            field = field[path[p]];
                        else
                            break;
                    }
                    if (field && field.constructor == Array) {
                        field.forEach(function(item) {
                            frequencies[item] = inc + (isNaN(frequencies[item]) ? 0: frequencies[item]);
                        });
                    } else {
                        var item = field;
                        types[item] = item;
                        frequencies[item] = inc + (isNaN(frequencies[item]) ? 0: frequencies[item]);
                    }
                });
                return [total, frequencies, types];
            }
        """
        total, data, types = self.exec_js(freq_func, field)
        values = {types.get(k): int(v) for k, v in data.items()}

        if normalize:
            values = {k: float(v) / total for k, v in values.items()}

        frequencies = {}
        for k, v in values.items():
            if isinstance(k, float):
                if int(k) == k:
                    k = int(k)

            frequencies[k] = v

        return frequencies

    def _fields_to_dbfields(self, fields):
        """Translate fields' paths to their db equivalents."""
        subclasses = []
        if self._document._meta["allow_inheritance"]:
            subclasses = [get_document(x) for x in self._document._subclasses][1:]

        db_field_paths = []
        for field in fields:
            field_parts = field.split(".")
            try:
                field = ".".join(
                    f if isinstance(f, str) else f.db_field
                    for f in self._document._lookup_field(field_parts)
                )
                db_field_paths.append(field)
            except LookUpError as err:
                found = False

                # If a field path wasn't found on the main document, go
                # through its subclasses and see if it exists on any of them.
                for subdoc in subclasses:
                    try:
                        subfield = ".".join(
                            f if isinstance(f, str) else f.db_field
                            for f in subdoc._lookup_field(field_parts)
                        )
                        db_field_paths.append(subfield)
                        found = True
                        break
                    except LookUpError:
                        pass

                if not found:
                    raise err

        return db_field_paths

    def _get_order_by(self, keys):
        """Given a list of MongoEngine-style sort keys, return a list
        of sorting tuples that can be applied to a PyMongo cursor. For
        example:

        >>> qs._get_order_by(['-last_name', 'first_name'])
        [('last_name', -1), ('first_name', 1)]
        """
        key_list = []
        for key in keys:
            if not key:
                continue

            if key == "$text_score":
                key_list.append(("_text_score", {"$meta": "textScore"}))
                continue

            direction = pymongo.ASCENDING
            if key[0] == "-":
                direction = pymongo.DESCENDING

            if key[0] in ("-", "+"):
                key = key[1:]

            key = key.replace("__", ".")
            try:
                key = self._document._translate_field_name(key)
            except Exception:
                # TODO this exception should be more specific
                pass

            key_list.append((key, direction))

        return key_list

    def _get_scalar(self, doc):
        def lookup(obj, name):
            chunks = name.split("__")
            for chunk in chunks:
                obj = getattr(obj, chunk)
            return obj

        data = [lookup(doc, n) for n in self._scalar]
        if len(data) == 1:
            return data[0]

        return tuple(data)

    def _sub_js_fields(self, code):
        """When fields are specified with [~fieldname] syntax, where
        *fieldname* is the Python name of a field, *fieldname* will be
        substituted for the MongoDB name of the field (specified using the
        :attr:`name` keyword argument in a field's constructor).
        """

        def field_sub(match):
            # Extract just the field name, and look up the field objects
            field_name = match.group(1).split(".")
            fields = self._document._lookup_field(field_name)
            # Substitute the correct name for the field into the javascript
            return '["%s"]' % fields[-1].db_field

        def field_path_sub(match):
            # Extract just the field name, and look up the field objects
            field_name = match.group(1).split(".")
            fields = self._document._lookup_field(field_name)
            # Substitute the correct name for the field into the javascript
            return ".".join([f.db_field for f in fields])

        code = re.sub(r"\[\s*~([A-z_][A-z_0-9.]+?)\s*\]", field_sub, code)
        code = re.sub(r"\{\{\s*~([A-z_][A-z_0-9.]+?)\s*\}\}", field_path_sub, code)
        return code

    def _chainable_method(self, method_name, val):
        """Call a particular method on the PyMongo cursor call a particular chainable method
        with the provided value.
        """
        queryset = self.clone()

        # Get an existing cursor object or create a new one
        cursor = queryset._cursor

        # Find the requested method on the cursor and call it with the
        # provided value
        getattr(cursor, method_name)(val)

        # Cache the value on the queryset._{method_name}
        setattr(queryset, "_" + method_name, val)

        return queryset