Codificação com Mark: abstraindo junções e subconjuntos em Python

1
{
2
  "_id": { "$oid": "657072b56731c9e580e9dd70" },
3
  "user_id": "4",
4
  "user_name": "@tanya15",
5
  "full_name": "Deborah White",
6
  "birth_date": { "$date": { "$numberLong": "931219200000" } },
7
  "email": "deanjacob@yahoo.com",
8
  "bio": "Music conference able doctor degree debate. Participant usually above relate.",
9
  "follower_count": { "$numberInt": "59" },
10
  "followers": [
11
    {
12
      "_id": { "$oid": "657072b66731c9e580e9dda6" },
13
      "user_id": "58",
14
      "user_name": "@rduncan",
15
      "bio": "Rich beautiful color life. Relationship instead win join enough board successful."
16
    },
17
    {
18
      "_id": { "$oid": "657072b66731c9e580e9dd99" },
19
      "user_id": "45",
20
      "user_name": "@paynericky",
21
      "bio": "Picture day couple democratic morning. Environment manage opportunity option star food she. Occur imagine population single avoid."
22
    },
23
    # ... other followers
24
   ]
25
}

1
class Follower(Document):
2
    _id = Field(transform=str)
3
    user_name = Field()
4
5
class Profile(Document):
6
    _id = Field(transform=str)
7
    followers = SequenceField(type=Follower)

1
profile = Profile(SOME_BSON_DATA)
2
for follower in profile.followers:
3
        assert isinstance(follower, Follower)

1
class SequenceField:
2
    """
3
    Allows an underlying array to have its elements wrapped in
4
    Document instances.
5
    """
6
7
    def __init__(
8
        self,
9
        type,
10
        field_name=None,
11
    ):
12
        self._type = type
13
        self.field_name = field_name
14
15
    def __set_name__(self, owner, name):
16
        """
17
        Called when the enclosing Document subclass (owner) is defined.
18
        """
19
        self.name = name  # Store the attribute name.
20
21
        # If a field-name mapping hasn't been provided,
22
        # the BSON field will have the same name as the attribute name.
23
        if self.field_name is None:
24
            self.field_name = name
25
26
    def __get__(self, ob, cls):
27
        """
28
        Called when the SequenceField attribute is accessed on the enclosed
29
        Document subclass.
30
        """
31
        try:
32
            # Lookup the field in the BSON, and return an array where each item
33
            # is wrapped by the class defined as type in __init__:
34
            return [
35
                self._type(item, ob._db)
36
                for item in ob._doc[self.field_name]
37
                ]
38
        except KeyError as ke:
39
            raise ValueError(
40
                f"Attribute {self.name!r} is mapped to missing document property {self.field_name!r}."
41
            ) from ke

1
return [
2
3
    self._type(item, ob._db)
4
5
    for item in ob._doc[self.field_name]
6
]

1
return [
2
   Follower(item, db=None) for item in profile._doc["followers"]
3
]

1
 {"$match": {"user_id": "4"}}

1
{
2
  "user_name": "@tanya15",
3
  "full_name": "Deborah White",
4
  "birth_date": {
5
    "$date": "1999-07-06T00:00:00.000Z"
6
  },
7
  "email": "deanjacob@yahoo.com",
8
  "bio": "Music conference able doctor degree debate. Participant usually above relate.",
9
  "user_id": "4",
10
  "follower_count": 59,
11
  "followers": [
12
    {
13
      "_id": {
14
        "$oid": "657072b66731c9e580e9dda6"
15
      },
16
      "user_id": "58",
17
      "user_name": "@rduncan",
18
      "bio": "Rich beautiful color life. Relationship instead win join enough board successful."
19
    },
20
    {
21
      "bio": "Picture day couple democratic morning. Environment manage opportunity option star food she. Occur imagine population single avoid.",
22
      "_id": {
23
        "$oid": "657072b66731c9e580e9dd99"
24
      },
25
      "user_id": "45",
26
      "user_name": "@paynericky"
27
    },
28
    {
29
      "_id": {
30
        "$oid": "657072b76731c9e580e9ddba"
31
      },
32
      "user_id": "78",
33
      "user_name": "@tiffanyhicks",
34
      "bio": "Sign writer win. Look television official information laugh. Lay plan effect break expert message during firm."
35
    },
36
   . . .
37
  ],
38
  "_id": {
39
    "$oid": "657072b56731c9e580e9dd70"
40
  }
41
}

1
{"$unwind": "$followers"}

1
# First document:
2
{
3
  "bio": "Music conference able doctor degree debate. Participant usually above relate.",
4
  "follower_count": 59,
5
  "followers": {
6
    "_id": {
7
      "$oid": "657072b66731c9e580e9dda6"
8
    },
9
    "user_id": "58",
10
    "user_name": "@rduncan",
11
    "bio": "Rich beautiful color life. Relationship instead win join enough board successful."
12
  },
13
  "user_id": "4",
14
  "user_name": "@tanya15",
15
  "full_name": "Deborah White",
16
  "birth_date": {
17
    "$date": "1999-07-06T00:00:00.000Z"
18
  },
19
  "email": "deanjacob@yahoo.com",
20
  "_id": {
21
    "$oid": "657072b56731c9e580e9dd70"
22
  }
23
}
24
25
# Second document
26
{
27
  "_id": {
28
    "$oid": "657072b56731c9e580e9dd70"
29
  },
30
  "full_name": "Deborah White",
31
  "email": "deanjacob@yahoo.com",
32
  "bio": "Music conference able doctor degree debate. Participant usually above relate.",
33
  "follower_count": 59,
34
  "user_id": "4",
35
  "user_name": "@tanya15",
36
  "birth_date": {
37
    "$date": "1999-07-06T00:00:00.000Z"
38
  },
39
  "followers": {
40
    "_id": {
41
      "$oid": "657072b66731c9e580e9dd99"
42
    },
43
    "user_id": "45",
44
    "user_name": "@paynericky",
45
    "bio": "Picture day couple democratic morning. Environment manage opportunity option star food she. Occur imagine population single avoid."
46
  }
47
48
# . . . More documents follow

1
{"$replaceRoot": {"newRoot": "$followers"}},

1
# Document 1:
2
{
3
  "_id": {
4
    "$oid": "657072b66731c9e580e9dda6"
5
  },
6
  "user_id": "58",
7
  "user_name": "@rduncan",
8
  "bio": "Rich beautiful color life. Relationship instead win join enough board successful."
9
}
10
11
# Document 2
12
{
13
  "_id": {
14
    "$oid": "657072b66731c9e580e9dd99"
15
  },
16
  "user_id": "45",
17
  "user_name": "@paynericky",
18
  "bio": "Picture day couple democratic morning. Environment manage opportunity option star food she. Occur imagine population single avoid."
19
}
20
} # . . . More documents follow

1
[
2
    {"$match": {"user_id": "4"}},
3
    {"$unwind": "$followers"},
4
    {"$replaceRoot": {"newRoot": "$followers"}},
5
]

1
class Field: 
2
    def __init__(
3
        self,
4
        type,
5
        field_name=None,
6
        superset_collection=None,
7
        superset_query: Callable = None,
8
    ):
9
        self._type = type
10
        self.field_name = field_name
11
        self.superset_collection = superset_collection
12
        self.superset_query = superset_query

1
embedded_followers = self._doc["followers"] # a list
2
cursor = followers.find({"user_id": "4"})   # a lazy database cursor
3
4
# Looping through all_followers will only make a database call if you have 
5
# looped through all of the contents of embedded_followers:
6
all_followers = itertools.chain(embedded_followers, cursor)

1
def __get__(self, ob, cls):
2
    if self.superset_query is None:
3
        # Use an empty sequence if there are no extra items.
4
        # It's still iterable, like a cursor, but immediately exits.
5
        superset = []
6
    else:
7
        # Call the superset_query callable to obtain the generated query:
8
        query = self.superset_query(ob)
9
10
        # If the query is a mapping, it's a find query, otherwise it's an
11
        # aggregation pipeline.
12
        if isinstance(query, Mapping):
13
            superset = ob._db.get_collection(self.superset_collection).find(query)
14
        elif isinstance(query, Iterable):
15
            superset = ob._db.get_collection(self.superset_collection).aggregate(
16
                query
17
            )
18
        else:
19
            raise Exception("Returned was not a mapping or iterable.")
20
21
    try:
22
        # Return an iterable that first yields all the embedded items, and
23
24
        return chain(
25
            [self._type(item, ob._db) for item in ob._doc[self.field_name]],
26
            (self._type(item, ob._db) for item in superset),
27
        )
28
    except KeyError as ke:
29
        raise ValueError(
30
            f"Attribute {self.name!r} is mapped to missing document property {self.field_name!r}."
31
        ) from ke

1
# This is the same as it was originally
2
class Follower(Document):
3
    _id = Field(transform=str)
4
    user_name = Field()
5
6
7
def extra_followers_query(profile):
8
    return [
9
        {
10
            "$match": {"user_id": profile.user_id},
11
        },
12
        {"$unwind": "$followers"},
13
        {"$replaceRoot": {"newRoot": "$followers"}},
14
    ]
15
    
16
class Profile(Document):
17
    _id = Field(transform=str)
18
    followers = SequenceField(
19
        type=Follower,
20
        superset_collection="followers",
21
        superset_query=lambda ob: extra_followers_query,
22
    )

1
for follower in profile.followers:
2
    print(follower.user_name)