Unable to insert nested object in mongodb using pymongo - python

I am coming today following an issue that doesn't make sense to me using python and mongodb. I am a Go/C# developer so maybe I am missing something but I have the following case:
from datetime import datetime
from bson import ObjectId
class DailyActivity:
user_ids = []
date = None
def __init__(self, user_ids : [ObjectId] = [], date : datetime = None):
self.user_ids = user_ids
self.date = date
class ActivitiesThroughDays:
daily_activies = []
def add_daily_activity(self, daily_activity : DailyActivity = None):
daily_activies.append(daily_activity)
I then have these 2 classes but also another file containing some helper to use mongodb:
from pymongo import MongoClient
def get_client():
return MongoClient('localhost', 27017)
def get_database(database_name: str = None):
if database_name is None:
raise AttributeError("database name is None.")
return get_client().get_database(database_name)
def get_X_database():
return get_database("X")
And here we get to the issue.. I am now building a simple ActivitiesThroughDays object which has only one DailyActivity containing X user ids (as ObjectId array/list).
However, when I try to insert_one, I get the following:
TypeError: document must be an instance of dict, bson.son.SON, bson.raw_bson.RawBSONDocument, or a type that inherits from collections.MutableMapping
this is the piece of code that raise the exception:
def insert_activities_though_days(activities_through_days: ActivitiesThroughDays = None):
if activities_through_days is None:
raise AttributeError("activities_through_days is None.")
col = get_EM_column("activities_through_days")
col.insert_one(activities_through_days)
Based on the above issue, I then tried to convert my ActivitiesThroughDays into dic/json:
col.insert_one(activities_through_days.__dict__)
bson.errors.InvalidDocument: cannot encode object: models. DailyActivity. DailyActivity object at 0x10eea0320, of type: class 'models. DailyActivity. DailyActivity'
col.insert_one(json.dumps(activities_through_days))
TypeError: Object of type ActivitiesThroughDays is not JSON serializable
So based on this, I began to search for different solutions over google and found out solutions such as :
def to_dict(obj):
if not hasattr(obj,"__dict__"):
return obj
result = {}
for key, val in obj.__dict__.items():
if key.startswith("_"):
continue
element = []
if isinstance(val, list):
for item in val:
element.append(to_dict(item))
else:
element = to_dict(val)
result[key] = element
return result
But I got :
bson.errors.InvalidDocument: cannot encode object: property object at 0x10229aa98, of type: class 'property'
For each step I move forward, another issue comes up... To me, all of this doesn't make sense at all because.. there should be a generic serializer/deserializer somewhere that would, from 1 line, convert any nested objects/arrays to be inserted in mongodb..
Also, from one of the solution I tried, I found out that ObjectId were ignored while mapping to json/dict (I don't remember which one)
I am not at all a Python developer so please, feel free to give any tips :)
Thanks

pymongo's interface expects dict and .__dict__ is a very low level attribute.
I'm afraid you'll spend a lot of energy if you try to build an ORM/ODM for mongodb from scratch.
There are existing ORM/ODM libraries that exist for mongodb in python (mongoengine, pymodm which are quite similar) and that could help you to get something working quickly.
Here are a few lines that shows how the models would look with mongoengine and how to save them:
import datetime as dt
from mongoengine import *
connect(host='mongodb://localhost:27017/testdb')
class User(Document):
email = EmailField(required=True)
class DailyActivity(Document):
users = ListField(ReferenceField(User))
date = DateTimeField(default=dt.datetime.utcnow)
user = User(email='test#garbage.com').save()
user2 = User(email='test2#garbage.com').save()
activity = DailyActivity(users=[user, user2]).save()
I hope this helps

Related

convert models to json [duplicate]

Django has some good automatic serialization of ORM models returned from DB to JSON format.
How to serialize SQLAlchemy query result to JSON format?
I tried jsonpickle.encode but it encodes query object itself.
I tried json.dumps(items) but it returns
TypeError: <Product('3', 'some name', 'some desc')> is not JSON serializable
Is it really so hard to serialize SQLAlchemy ORM objects to JSON /XML? Isn't there any default serializer for it? It's very common task to serialize ORM query results nowadays.
What I need is just to return JSON or XML data representation of SQLAlchemy query result.
SQLAlchemy objects query result in JSON/XML format is needed to be used in javascript datagird (JQGrid http://www.trirand.com/blog/)
You could just output your object as a dictionary:
class User:
def as_dict(self):
return {c.name: getattr(self, c.name) for c in self.__table__.columns}
And then you use User.as_dict() to serialize your object.
As explained in Convert sqlalchemy row object to python dict
A flat implementation
You could use something like this:
from sqlalchemy.ext.declarative import DeclarativeMeta
class AlchemyEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj.__class__, DeclarativeMeta):
# an SQLAlchemy class
fields = {}
for field in [x for x in dir(obj) if not x.startswith('_') and x != 'metadata']:
data = obj.__getattribute__(field)
try:
json.dumps(data) # this will fail on non-encodable values, like other classes
fields[field] = data
except TypeError:
fields[field] = None
# a json-encodable dict
return fields
return json.JSONEncoder.default(self, obj)
and then convert to JSON using:
c = YourAlchemyClass()
print json.dumps(c, cls=AlchemyEncoder)
It will ignore fields that are not encodable (set them to 'None').
It doesn't auto-expand relations (since this could lead to self-references, and loop forever).
A recursive, non-circular implementation
If, however, you'd rather loop forever, you could use:
from sqlalchemy.ext.declarative import DeclarativeMeta
def new_alchemy_encoder():
_visited_objs = []
class AlchemyEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj.__class__, DeclarativeMeta):
# don't re-visit self
if obj in _visited_objs:
return None
_visited_objs.append(obj)
# an SQLAlchemy class
fields = {}
for field in [x for x in dir(obj) if not x.startswith('_') and x != 'metadata']:
fields[field] = obj.__getattribute__(field)
# a json-encodable dict
return fields
return json.JSONEncoder.default(self, obj)
return AlchemyEncoder
And then encode objects using:
print json.dumps(e, cls=new_alchemy_encoder(), check_circular=False)
This would encode all children, and all their children, and all their children... Potentially encode your entire database, basically. When it reaches something its encoded before, it will encode it as 'None'.
A recursive, possibly-circular, selective implementation
Another alternative, probably better, is to be able to specify the fields you want to expand:
def new_alchemy_encoder(revisit_self = False, fields_to_expand = []):
_visited_objs = []
class AlchemyEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj.__class__, DeclarativeMeta):
# don't re-visit self
if revisit_self:
if obj in _visited_objs:
return None
_visited_objs.append(obj)
# go through each field in this SQLalchemy class
fields = {}
for field in [x for x in dir(obj) if not x.startswith('_') and x != 'metadata']:
val = obj.__getattribute__(field)
# is this field another SQLalchemy object, or a list of SQLalchemy objects?
if isinstance(val.__class__, DeclarativeMeta) or (isinstance(val, list) and len(val) > 0 and isinstance(val[0].__class__, DeclarativeMeta)):
# unless we're expanding this field, stop here
if field not in fields_to_expand:
# not expanding this field: set it to None and continue
fields[field] = None
continue
fields[field] = val
# a json-encodable dict
return fields
return json.JSONEncoder.default(self, obj)
return AlchemyEncoder
You can now call it with:
print json.dumps(e, cls=new_alchemy_encoder(False, ['parents']), check_circular=False)
To only expand SQLAlchemy fields called 'parents', for example.
Python 3.7+ and Flask 1.1+ can use the built-in dataclasses package
from dataclasses import dataclass
from datetime import datetime
from flask import Flask, jsonify
from flask_sqlalchemy import SQLAlchemy
app = Flask(__name__)
db = SQLAlchemy(app)
#dataclass
class User(db.Model):
id: int
email: str
id = db.Column(db.Integer, primary_key=True, auto_increment=True)
email = db.Column(db.String(200), unique=True)
#app.route('/users/')
def users():
users = User.query.all()
return jsonify(users)
if __name__ == "__main__":
users = User(email="user1#gmail.com"), User(email="user2#gmail.com")
db.create_all()
db.session.add_all(users)
db.session.commit()
app.run()
The /users/ route will now return a list of users.
[
{"email": "user1#gmail.com", "id": 1},
{"email": "user2#gmail.com", "id": 2}
]
Auto-serialize related models
#dataclass
class Account(db.Model):
id: int
users: User
id = db.Column(db.Integer)
users = db.relationship(User) # User model would need a db.ForeignKey field
The response from jsonify(account) would be this.
{
"id":1,
"users":[
{
"email":"user1#gmail.com",
"id":1
},
{
"email":"user2#gmail.com",
"id":2
}
]
}
Overwrite the default JSON Encoder
from flask.json import JSONEncoder
class CustomJSONEncoder(JSONEncoder):
"Add support for serializing timedeltas"
def default(o):
if type(o) == datetime.timedelta:
return str(o)
if type(o) == datetime.datetime:
return o.isoformat()
return super().default(o)
app.json_encoder = CustomJSONEncoder
You can convert a RowProxy to a dict like this:
d = dict(row.items())
Then serialize that to JSON ( you will have to specify an encoder for things like datetime values )
It's not that hard if you just want one record ( and not a full hierarchy of related records ).
json.dumps([(dict(row.items())) for row in rs])
I recommend using marshmallow. It allows you to create serializers to represent your model instances with support to relations and nested objects.
Here is a truncated example from their docs. Take the ORM model, Author:
class Author(db.Model):
id = db.Column(db.Integer, primary_key=True)
first = db.Column(db.String(80))
last = db.Column(db.String(80))
A marshmallow schema for that class is constructed like this:
class AuthorSchema(Schema):
id = fields.Int(dump_only=True)
first = fields.Str()
last = fields.Str()
formatted_name = fields.Method("format_name", dump_only=True)
def format_name(self, author):
return "{}, {}".format(author.last, author.first)
...and used like this:
author_schema = AuthorSchema()
author_schema.dump(Author.query.first())
...would produce an output like this:
{
"first": "Tim",
"formatted_name": "Peters, Tim",
"id": 1,
"last": "Peters"
}
Have a look at their full Flask-SQLAlchemy Example.
A library called marshmallow-sqlalchemy specifically integrates SQLAlchemy and marshmallow. In that library, the schema for the Author model described above looks like this:
class AuthorSchema(ModelSchema):
class Meta:
model = Author
The integration allows the field types to be inferred from the SQLAlchemy Column types.
marshmallow-sqlalchemy here.
You can use introspection of SqlAlchemy as this :
mysql = SQLAlchemy()
from sqlalchemy import inspect
class Contacts(mysql.Model):
__tablename__ = 'CONTACTS'
id = mysql.Column(mysql.Integer, primary_key=True)
first_name = mysql.Column(mysql.String(128), nullable=False)
last_name = mysql.Column(mysql.String(128), nullable=False)
phone = mysql.Column(mysql.String(128), nullable=False)
email = mysql.Column(mysql.String(128), nullable=False)
street = mysql.Column(mysql.String(128), nullable=False)
zip_code = mysql.Column(mysql.String(128), nullable=False)
city = mysql.Column(mysql.String(128), nullable=False)
def toDict(self):
return { c.key: getattr(self, c.key) for c in inspect(self).mapper.column_attrs }
#app.route('/contacts',methods=['GET'])
def getContacts():
contacts = Contacts.query.all()
contactsArr = []
for contact in contacts:
contactsArr.append(contact.toDict())
return jsonify(contactsArr)
#app.route('/contacts/<int:id>',methods=['GET'])
def getContact(id):
contact = Contacts.query.get(id)
return jsonify(contact.toDict())
Get inspired from an answer here :
Convert sqlalchemy row object to python dict
Flask-JsonTools package has an implementation of JsonSerializableBase Base class for your models.
Usage:
from sqlalchemy.ext.declarative import declarative_base
from flask.ext.jsontools import JsonSerializableBase
Base = declarative_base(cls=(JsonSerializableBase,))
class User(Base):
#...
Now the User model is magically serializable.
If your framework is not Flask, you can just grab the code
For security reasons you should never return all the model's fields. I prefer to selectively choose them.
Flask's json encoding now supports UUID, datetime and relationships (and added query and query_class for flask_sqlalchemy db.Model class). I've updated the encoder as follows:
app/json_encoder.py
from sqlalchemy.ext.declarative import DeclarativeMeta
from flask import json
class AlchemyEncoder(json.JSONEncoder):
def default(self, o):
if isinstance(o.__class__, DeclarativeMeta):
data = {}
fields = o.__json__() if hasattr(o, '__json__') else dir(o)
for field in [f for f in fields if not f.startswith('_') and f not in ['metadata', 'query', 'query_class']]:
value = o.__getattribute__(field)
try:
json.dumps(value)
data[field] = value
except TypeError:
data[field] = None
return data
return json.JSONEncoder.default(self, o)
app/__init__.py
# json encoding
from app.json_encoder import AlchemyEncoder
app.json_encoder = AlchemyEncoder
With this I can optionally add a __json__ property that returns the list of fields I wish to encode:
app/models.py
class Queue(db.Model):
id = db.Column(db.Integer, primary_key=True)
song_id = db.Column(db.Integer, db.ForeignKey('song.id'), unique=True, nullable=False)
song = db.relationship('Song', lazy='joined')
type = db.Column(db.String(20), server_default=u'audio/mpeg')
src = db.Column(db.String(255), nullable=False)
created_at = db.Column(db.DateTime, server_default=db.func.now())
updated_at = db.Column(db.DateTime, server_default=db.func.now(), onupdate=db.func.now())
def __init__(self, song):
self.song = song
self.src = song.full_path
def __json__(self):
return ['song', 'src', 'type', 'created_at']
I add #jsonapi to my view, return the resultlist and then my output is as follows:
[
{
"created_at": "Thu, 23 Jul 2015 11:36:53 GMT",
"song":
{
"full_path": "/static/music/Audioslave/Audioslave [2002]/1 Cochise.mp3",
"id": 2,
"path_name": "Audioslave/Audioslave [2002]/1 Cochise.mp3"
},
"src": "/static/music/Audioslave/Audioslave [2002]/1 Cochise.mp3",
"type": "audio/mpeg"
}
]
A more detailed explanation.
In your model, add:
def as_dict(self):
return {c.name: str(getattr(self, c.name)) for c in self.__table__.columns}
The str() is for python 3 so if using python 2 use unicode(). It should help deserialize dates. You can remove it if not dealing with those.
You can now query the database like this
some_result = User.query.filter_by(id=current_user.id).first().as_dict()
First() is needed to avoid weird errors. as_dict() will now deserialize the result. After deserialization, it is ready to be turned to json
jsonify(some_result)
While the original question goes back awhile, the number of answers here (and my own experiences) suggest it's a non-trivial question with a lot of different approaches of varying complexity with different trade-offs.
That's why I built the SQLAthanor library that extends SQLAlchemy's declarative ORM with configurable serialization/de-serialization support that you might want to take a look at.
The library supports:
Python 2.7, 3.4, 3.5, and 3.6.
SQLAlchemy versions 0.9 and higher
serialization/de-serialization to/from JSON, CSV, YAML, and Python dict
serialization/de-serialization of columns/attributes, relationships, hybrid properties, and association proxies
enabling and disabling of serialization for particular formats and columns/relationships/attributes (e.g. you want to support an inbound password value, but never include an outbound one)
pre-serialization and post-deserialization value processing (for validation or type coercion)
a pretty straightforward syntax that is both Pythonic and seamlessly consistent with SQLAlchemy's own approach
You can check out the (I hope!) comprehensive docs here: https://sqlathanor.readthedocs.io/en/latest
Hope this helps!
Custom serialization and deserialization.
"from_json" (class method) builds a Model object based on json data.
"deserialize" could be called only on instance, and merge all data from json into Model instance.
"serialize" - recursive serialization
__write_only__ property is needed to define write only properties ("password_hash" for example).
class Serializable(object):
__exclude__ = ('id',)
__include__ = ()
__write_only__ = ()
#classmethod
def from_json(cls, json, selfObj=None):
if selfObj is None:
self = cls()
else:
self = selfObj
exclude = (cls.__exclude__ or ()) + Serializable.__exclude__
include = cls.__include__ or ()
if json:
for prop, value in json.iteritems():
# ignore all non user data, e.g. only
if (not (prop in exclude) | (prop in include)) and isinstance(
getattr(cls, prop, None), QueryableAttribute):
setattr(self, prop, value)
return self
def deserialize(self, json):
if not json:
return None
return self.__class__.from_json(json, selfObj=self)
#classmethod
def serialize_list(cls, object_list=[]):
output = []
for li in object_list:
if isinstance(li, Serializable):
output.append(li.serialize())
else:
output.append(li)
return output
def serialize(self, **kwargs):
# init write only props
if len(getattr(self.__class__, '__write_only__', ())) == 0:
self.__class__.__write_only__ = ()
dictionary = {}
expand = kwargs.get('expand', ()) or ()
prop = 'props'
if expand:
# expand all the fields
for key in expand:
getattr(self, key)
iterable = self.__dict__.items()
is_custom_property_set = False
# include only properties passed as parameter
if (prop in kwargs) and (kwargs.get(prop, None) is not None):
is_custom_property_set = True
iterable = kwargs.get(prop, None)
# loop trough all accessible properties
for key in iterable:
accessor = key
if isinstance(key, tuple):
accessor = key[0]
if not (accessor in self.__class__.__write_only__) and not accessor.startswith('_'):
# force select from db to be able get relationships
if is_custom_property_set:
getattr(self, accessor, None)
if isinstance(self.__dict__.get(accessor), list):
dictionary[accessor] = self.__class__.serialize_list(object_list=self.__dict__.get(accessor))
# check if those properties are read only
elif isinstance(self.__dict__.get(accessor), Serializable):
dictionary[accessor] = self.__dict__.get(accessor).serialize()
else:
dictionary[accessor] = self.__dict__.get(accessor)
return dictionary
Here is a solution that lets you select the relations you want to include in your output as deep as you would like to go.
NOTE: This is a complete re-write taking a dict/str as an arg rather than a list. fixes some stuff..
def deep_dict(self, relations={}):
"""Output a dict of an SA object recursing as deep as you want.
Takes one argument, relations which is a dictionary of relations we'd
like to pull out. The relations dict items can be a single relation
name or deeper relation names connected by sub dicts
Example:
Say we have a Person object with a family relationship
person.deep_dict(relations={'family':None})
Say the family object has homes as a relation then we can do
person.deep_dict(relations={'family':{'homes':None}})
OR
person.deep_dict(relations={'family':'homes'})
Say homes has a relation like rooms you can do
person.deep_dict(relations={'family':{'homes':'rooms'}})
and so on...
"""
mydict = dict((c, str(a)) for c, a in
self.__dict__.items() if c != '_sa_instance_state')
if not relations:
# just return ourselves
return mydict
# otherwise we need to go deeper
if not isinstance(relations, dict) and not isinstance(relations, str):
raise Exception("relations should be a dict, it is of type {}".format(type(relations)))
# got here so check and handle if we were passed a dict
if isinstance(relations, dict):
# we were passed deeper info
for left, right in relations.items():
myrel = getattr(self, left)
if isinstance(myrel, list):
mydict[left] = [rel.deep_dict(relations=right) for rel in myrel]
else:
mydict[left] = myrel.deep_dict(relations=right)
# if we get here check and handle if we were passed a string
elif isinstance(relations, str):
# passed a single item
myrel = getattr(self, relations)
left = relations
if isinstance(myrel, list):
mydict[left] = [rel.deep_dict(relations=None)
for rel in myrel]
else:
mydict[left] = myrel.deep_dict(relations=None)
return mydict
so for an example using person/family/homes/rooms... turning it into json all you need is
json.dumps(person.deep_dict(relations={'family':{'homes':'rooms'}}))
step1:
class CNAME:
...
def as_dict(self):
return {item.name: getattr(self, item.name) for item in self.__table__.columns}
step2:
list = []
for data in session.query(CNAME).all():
list.append(data.as_dict())
step3:
return jsonify(list)
Even though it's a old post, Maybe I didn't answer the question above, but I want to talk about my serialization, at least it works for me.
I use FastAPI,SqlAlchemy and MySQL, but I don't use orm model;
# from sqlalchemy import create_engine
# from sqlalchemy.orm import sessionmaker
# engine = create_engine(config.SQLALCHEMY_DATABASE_URL, pool_pre_ping=True)
# SessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=engine)
Serialization code
import decimal
import datetime
def alchemy_encoder(obj):
"""JSON encoder function for SQLAlchemy special classes."""
if isinstance(obj, datetime.date):
return obj.strftime("%Y-%m-%d %H:%M:%S")
elif isinstance(obj, decimal.Decimal):
return float(obj)
import json
from sqlalchemy import text
# db is SessionLocal() object
app_sql = 'SELECT * FROM app_info ORDER BY app_id LIMIT :page,:page_size'
# The next two are the parameters passed in
page = 1
page_size = 10
# execute sql and return a <class 'sqlalchemy.engine.result.ResultProxy'> object
app_list = db.execute(text(app_sql), {'page': page, 'page_size': page_size})
# serialize
res = json.loads(json.dumps([dict(r) for r in app_list], default=alchemy_encoder))
If it doesn't work, please ignore my answer. I refer to it here
https://codeandlife.com/2014/12/07/sqlalchemy-results-to-json-the-easy-way/
install simplejson by
pip install simplejson and the create a class
class Serialise(object):
def _asdict(self):
"""
Serialization logic for converting entities using flask's jsonify
:return: An ordered dictionary
:rtype: :class:`collections.OrderedDict`
"""
result = OrderedDict()
# Get the columns
for key in self.__mapper__.c.keys():
if isinstance(getattr(self, key), datetime):
result["x"] = getattr(self, key).timestamp() * 1000
result["timestamp"] = result["x"]
else:
result[key] = getattr(self, key)
return result
and inherit this class to every orm classes so that this _asdict function gets registered to every ORM class and boom.
And use jsonify anywhere
It is not so straighforward. I wrote some code to do this. I'm still working on it, and it uses the MochiKit framework. It basically translates compound objects between Python and Javascript using a proxy and registered JSON converters.
Browser side for database objects is db.js
It needs the basic Python proxy source in proxy.js.
On the Python side there is the base proxy module.
Then finally the SqlAlchemy object encoder in webserver.py.
It also depends on metadata extractors found in the models.py file.
def alc2json(row):
return dict([(col, str(getattr(row,col))) for col in row.__table__.columns.keys()])
I thought I'd play a little code golf with this one.
FYI: I am using automap_base since we have a separately designed schema according to business requirements. I just started using SQLAlchemy today but the documentation states that automap_base is an extension to declarative_base which seems to be the typical paradigm in the SQLAlchemy ORM so I believe this should work.
It does not get fancy with following foreign keys per Tjorriemorrie's solution, but it simply matches columns to values and handles Python types by str()-ing the column values. Our values consist Python datetime.time and decimal.Decimal class type results so it gets the job done.
Hope this helps any passers-by!
I know this is quite an older post. I took solution given by #SashaB and modified as per my need.
I added following things to it:
Field ignore list: A list of fields to be ignored while serializing
Field replace list: A dictionary containing field names to be replaced by values while serializing.
Removed methods and BaseQuery getting serialized
My code is as follows:
def alchemy_json_encoder(revisit_self = False, fields_to_expand = [], fields_to_ignore = [], fields_to_replace = {}):
"""
Serialize SQLAlchemy result into JSon
:param revisit_self: True / False
:param fields_to_expand: Fields which are to be expanded for including their children and all
:param fields_to_ignore: Fields to be ignored while encoding
:param fields_to_replace: Field keys to be replaced by values assigned in dictionary
:return: Json serialized SQLAlchemy object
"""
_visited_objs = []
class AlchemyEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj.__class__, DeclarativeMeta):
# don't re-visit self
if revisit_self:
if obj in _visited_objs:
return None
_visited_objs.append(obj)
# go through each field in this SQLalchemy class
fields = {}
for field in [x for x in dir(obj) if not x.startswith('_') and x != 'metadata' and x not in fields_to_ignore]:
val = obj.__getattribute__(field)
# is this field method defination, or an SQLalchemy object
if not hasattr(val, "__call__") and not isinstance(val, BaseQuery):
field_name = fields_to_replace[field] if field in fields_to_replace else field
# is this field another SQLalchemy object, or a list of SQLalchemy objects?
if isinstance(val.__class__, DeclarativeMeta) or \
(isinstance(val, list) and len(val) > 0 and isinstance(val[0].__class__, DeclarativeMeta)):
# unless we're expanding this field, stop here
if field not in fields_to_expand:
# not expanding this field: set it to None and continue
fields[field_name] = None
continue
fields[field_name] = val
# a json-encodable dict
return fields
return json.JSONEncoder.default(self, obj)
return AlchemyEncoder
Hope it helps someone!
Use the built-in serializer in SQLAlchemy:
from sqlalchemy.ext.serializer import loads, dumps
obj = MyAlchemyObject()
# serialize object
serialized_obj = dumps(obj)
# deserialize object
obj = loads(serialized_obj)
If you're transferring the object between sessions, remember to detach the object from the current session using session.expunge(obj).
To attach it again, just do session.add(obj).
Under Flask, this works and handles datatime fields, transforming a field of type
'time': datetime.datetime(2018, 3, 22, 15, 40) into
"time": "2018-03-22 15:40:00":
obj = {c.name: str(getattr(self, c.name)) for c in self.__table__.columns}
# This to get the JSON body
return json.dumps(obj)
# Or this to get a response object
return jsonify(obj)
following code will serialize sqlalchemy result to json.
import json
from collections import OrderedDict
def asdict(self):
result = OrderedDict()
for key in self.__mapper__.c.keys():
if getattr(self, key) is not None:
result[key] = str(getattr(self, key))
else:
result[key] = getattr(self, key)
return result
def to_array(all_vendors):
v = [ ven.asdict() for ven in all_vendors ]
return json.dumps(v)
Calling fun,
def all_products():
all_products = Products.query.all()
return to_array(all_products)
The AlchemyEncoder is wonderful but sometimes fails with Decimal values. Here is an improved encoder that solves the decimal problem -
class AlchemyEncoder(json.JSONEncoder):
# To serialize SQLalchemy objects
def default(self, obj):
if isinstance(obj.__class__, DeclarativeMeta):
model_fields = {}
for field in [x for x in dir(obj) if not x.startswith('_') and x != 'metadata']:
data = obj.__getattribute__(field)
print data
try:
json.dumps(data) # this will fail on non-encodable values, like other classes
model_fields[field] = data
except TypeError:
model_fields[field] = None
return model_fields
if isinstance(obj, Decimal):
return float(obj)
return json.JSONEncoder.default(self, obj)
When using sqlalchemy to connect to a db I this is a simple solution which is highly configurable. Use pandas.
import pandas as pd
import sqlalchemy
#sqlalchemy engine configuration
engine = sqlalchemy.create_engine....
def my_function():
#read in from sql directly into a pandas dataframe
#check the pandas documentation for additional config options
sql_DF = pd.read_sql_table("table_name", con=engine)
# "orient" is optional here but allows you to specify the json formatting you require
sql_json = sql_DF.to_json(orient="index")
return sql_json
(Tiny tweak on Sasha B's really excellent answer)
This specifically converts datetime objects to strings which in the original answer would be converted to None:
# Standard library imports
from datetime import datetime
import json
# 3rd party imports
from sqlalchemy.ext.declarative import DeclarativeMeta
class JsonEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj.__class__, DeclarativeMeta):
dict = {}
# Remove invalid fields and just get the column attributes
columns = [x for x in dir(obj) if not x.startswith("_") and x != "metadata"]
for column in columns:
value = obj.__getattribute__(column)
try:
json.dumps(value)
dict[column] = value
except TypeError:
if isinstance(value, datetime):
dict[column] = value.__str__()
else:
dict[column] = None
return dict
return json.JSONEncoder.default(self, obj)
class SqlToDict:
def __init__(self, data) -> None:
self.data = data
def to_timestamp(self, date):
if isinstance(date, datetime):
return int(datetime.timestamp(date))
else:
return date
def to_dict(self) -> List:
arr = []
for i in self.data:
keys = [*i.keys()]
values = [*i]
values = [self.to_timestamp(d) for d in values]
arr.append(dict(zip(keys, values)))
return arr
For example:
SqlToDict(data).to_dict()
Very late 2023
My implementation
def obj_to_dict(obj, remove=['_sa_instance_state'], debug=False):
result = {}
if type(obj).__name__ == "Row":
return dict(obj)
obj = obj.__dict__
for key in obj:
if key in remove:
continue
result[key] = obj[key]
if debug:
print(result)
return result
The built in serializer chokes with utf-8 cannot decode invalid start byte for some inputs. Instead, I went with:
def row_to_dict(row):
temp = row.__dict__
temp.pop('_sa_instance_state', None)
return temp
def rows_to_list(rows):
ret_rows = []
for row in rows:
ret_rows.append(row_to_dict(row))
return ret_rows
#website_blueprint.route('/api/v1/some/endpoint', methods=['GET'])
def some_api():
'''
/some_endpoint
'''
rows = rows_to_list(SomeModel.query.all())
response = app.response_class(
response=jsonplus.dumps(rows),
status=200,
mimetype='application/json'
)
return response
Maybe you can use a class like this
from sqlalchemy.ext.declarative import declared_attr
from sqlalchemy import Table
class Custom:
"""Some custom logic here!"""
__table__: Table # def for mypy
#declared_attr
def __tablename__(cls): # pylint: disable=no-self-argument
return cls.__name__ # pylint: disable= no-member
def to_dict(self) -> Dict[str, Any]:
"""Serializes only column data."""
return {c.name: getattr(self, c.name) for c in self.__table__.columns}
Base = declarative_base(cls=Custom)
class MyOwnTable(Base):
#COLUMNS!
With that all objects have the to_dict method
While using some raw sql and undefined objects, using cursor.description appeared to get what I was looking for:
with connection.cursor() as cur:
print(query)
cur.execute(query)
for item in cur.fetchall():
row = {column.name: item[i] for i, column in enumerate(cur.description)}
print(row)
This is a JSONEncoder version that preserves model column order and only keeps recursively defined column and relationship fields. It also formats most JSON unserializable types:
import json
from datetime import datetime
from decimal import Decimal
import arrow
from sqlalchemy.ext.declarative import DeclarativeMeta
class SQLAlchemyJSONEncoder(json.JSONEncoder):
"""
SQLAlchemy ORM JSON Encoder
If you have a "backref" relationship defined in your SQLAlchemy model,
this encoder raises a ValueError to stop an infinite loop.
"""
def default(self, obj):
if isinstance(obj, datetime):
return arrow.get(obj).isoformat()
elif isinstance(obj, Decimal):
return float(obj)
elif isinstance(obj, set):
return sorted(obj)
elif isinstance(obj.__class__, DeclarativeMeta):
for attribute, relationship in obj.__mapper__.relationships.items():
if isinstance(relationship.__getattribute__("backref"), tuple):
raise ValueError(
f'{obj.__class__} object has a "backref" relationship '
"that would cause an infinite loop!"
)
dictionary = {}
column_names = [column.name for column in obj.__table__.columns]
for key in column_names:
value = obj.__getattribute__(key)
if isinstance(value, datetime):
value = arrow.get(value).isoformat()
elif isinstance(value, Decimal):
value = float(value)
elif isinstance(value, set):
value = sorted(value)
dictionary[key] = value
for key in [
attribute
for attribute in dir(obj)
if not attribute.startswith("_")
and attribute != "metadata"
and attribute not in column_names
]:
value = obj.__getattribute__(key)
dictionary[key] = value
return dictionary
return super().default(obj)

Return JSON list from Flask Sqlalchemy query [duplicate]

Django has some good automatic serialization of ORM models returned from DB to JSON format.
How to serialize SQLAlchemy query result to JSON format?
I tried jsonpickle.encode but it encodes query object itself.
I tried json.dumps(items) but it returns
TypeError: <Product('3', 'some name', 'some desc')> is not JSON serializable
Is it really so hard to serialize SQLAlchemy ORM objects to JSON /XML? Isn't there any default serializer for it? It's very common task to serialize ORM query results nowadays.
What I need is just to return JSON or XML data representation of SQLAlchemy query result.
SQLAlchemy objects query result in JSON/XML format is needed to be used in javascript datagird (JQGrid http://www.trirand.com/blog/)
You could just output your object as a dictionary:
class User:
def as_dict(self):
return {c.name: getattr(self, c.name) for c in self.__table__.columns}
And then you use User.as_dict() to serialize your object.
As explained in Convert sqlalchemy row object to python dict
A flat implementation
You could use something like this:
from sqlalchemy.ext.declarative import DeclarativeMeta
class AlchemyEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj.__class__, DeclarativeMeta):
# an SQLAlchemy class
fields = {}
for field in [x for x in dir(obj) if not x.startswith('_') and x != 'metadata']:
data = obj.__getattribute__(field)
try:
json.dumps(data) # this will fail on non-encodable values, like other classes
fields[field] = data
except TypeError:
fields[field] = None
# a json-encodable dict
return fields
return json.JSONEncoder.default(self, obj)
and then convert to JSON using:
c = YourAlchemyClass()
print json.dumps(c, cls=AlchemyEncoder)
It will ignore fields that are not encodable (set them to 'None').
It doesn't auto-expand relations (since this could lead to self-references, and loop forever).
A recursive, non-circular implementation
If, however, you'd rather loop forever, you could use:
from sqlalchemy.ext.declarative import DeclarativeMeta
def new_alchemy_encoder():
_visited_objs = []
class AlchemyEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj.__class__, DeclarativeMeta):
# don't re-visit self
if obj in _visited_objs:
return None
_visited_objs.append(obj)
# an SQLAlchemy class
fields = {}
for field in [x for x in dir(obj) if not x.startswith('_') and x != 'metadata']:
fields[field] = obj.__getattribute__(field)
# a json-encodable dict
return fields
return json.JSONEncoder.default(self, obj)
return AlchemyEncoder
And then encode objects using:
print json.dumps(e, cls=new_alchemy_encoder(), check_circular=False)
This would encode all children, and all their children, and all their children... Potentially encode your entire database, basically. When it reaches something its encoded before, it will encode it as 'None'.
A recursive, possibly-circular, selective implementation
Another alternative, probably better, is to be able to specify the fields you want to expand:
def new_alchemy_encoder(revisit_self = False, fields_to_expand = []):
_visited_objs = []
class AlchemyEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj.__class__, DeclarativeMeta):
# don't re-visit self
if revisit_self:
if obj in _visited_objs:
return None
_visited_objs.append(obj)
# go through each field in this SQLalchemy class
fields = {}
for field in [x for x in dir(obj) if not x.startswith('_') and x != 'metadata']:
val = obj.__getattribute__(field)
# is this field another SQLalchemy object, or a list of SQLalchemy objects?
if isinstance(val.__class__, DeclarativeMeta) or (isinstance(val, list) and len(val) > 0 and isinstance(val[0].__class__, DeclarativeMeta)):
# unless we're expanding this field, stop here
if field not in fields_to_expand:
# not expanding this field: set it to None and continue
fields[field] = None
continue
fields[field] = val
# a json-encodable dict
return fields
return json.JSONEncoder.default(self, obj)
return AlchemyEncoder
You can now call it with:
print json.dumps(e, cls=new_alchemy_encoder(False, ['parents']), check_circular=False)
To only expand SQLAlchemy fields called 'parents', for example.
Python 3.7+ and Flask 1.1+ can use the built-in dataclasses package
from dataclasses import dataclass
from datetime import datetime
from flask import Flask, jsonify
from flask_sqlalchemy import SQLAlchemy
app = Flask(__name__)
db = SQLAlchemy(app)
#dataclass
class User(db.Model):
id: int
email: str
id = db.Column(db.Integer, primary_key=True, auto_increment=True)
email = db.Column(db.String(200), unique=True)
#app.route('/users/')
def users():
users = User.query.all()
return jsonify(users)
if __name__ == "__main__":
users = User(email="user1#gmail.com"), User(email="user2#gmail.com")
db.create_all()
db.session.add_all(users)
db.session.commit()
app.run()
The /users/ route will now return a list of users.
[
{"email": "user1#gmail.com", "id": 1},
{"email": "user2#gmail.com", "id": 2}
]
Auto-serialize related models
#dataclass
class Account(db.Model):
id: int
users: User
id = db.Column(db.Integer)
users = db.relationship(User) # User model would need a db.ForeignKey field
The response from jsonify(account) would be this.
{
"id":1,
"users":[
{
"email":"user1#gmail.com",
"id":1
},
{
"email":"user2#gmail.com",
"id":2
}
]
}
Overwrite the default JSON Encoder
from flask.json import JSONEncoder
class CustomJSONEncoder(JSONEncoder):
"Add support for serializing timedeltas"
def default(o):
if type(o) == datetime.timedelta:
return str(o)
if type(o) == datetime.datetime:
return o.isoformat()
return super().default(o)
app.json_encoder = CustomJSONEncoder
You can convert a RowProxy to a dict like this:
d = dict(row.items())
Then serialize that to JSON ( you will have to specify an encoder for things like datetime values )
It's not that hard if you just want one record ( and not a full hierarchy of related records ).
json.dumps([(dict(row.items())) for row in rs])
I recommend using marshmallow. It allows you to create serializers to represent your model instances with support to relations and nested objects.
Here is a truncated example from their docs. Take the ORM model, Author:
class Author(db.Model):
id = db.Column(db.Integer, primary_key=True)
first = db.Column(db.String(80))
last = db.Column(db.String(80))
A marshmallow schema for that class is constructed like this:
class AuthorSchema(Schema):
id = fields.Int(dump_only=True)
first = fields.Str()
last = fields.Str()
formatted_name = fields.Method("format_name", dump_only=True)
def format_name(self, author):
return "{}, {}".format(author.last, author.first)
...and used like this:
author_schema = AuthorSchema()
author_schema.dump(Author.query.first())
...would produce an output like this:
{
"first": "Tim",
"formatted_name": "Peters, Tim",
"id": 1,
"last": "Peters"
}
Have a look at their full Flask-SQLAlchemy Example.
A library called marshmallow-sqlalchemy specifically integrates SQLAlchemy and marshmallow. In that library, the schema for the Author model described above looks like this:
class AuthorSchema(ModelSchema):
class Meta:
model = Author
The integration allows the field types to be inferred from the SQLAlchemy Column types.
marshmallow-sqlalchemy here.
You can use introspection of SqlAlchemy as this :
mysql = SQLAlchemy()
from sqlalchemy import inspect
class Contacts(mysql.Model):
__tablename__ = 'CONTACTS'
id = mysql.Column(mysql.Integer, primary_key=True)
first_name = mysql.Column(mysql.String(128), nullable=False)
last_name = mysql.Column(mysql.String(128), nullable=False)
phone = mysql.Column(mysql.String(128), nullable=False)
email = mysql.Column(mysql.String(128), nullable=False)
street = mysql.Column(mysql.String(128), nullable=False)
zip_code = mysql.Column(mysql.String(128), nullable=False)
city = mysql.Column(mysql.String(128), nullable=False)
def toDict(self):
return { c.key: getattr(self, c.key) for c in inspect(self).mapper.column_attrs }
#app.route('/contacts',methods=['GET'])
def getContacts():
contacts = Contacts.query.all()
contactsArr = []
for contact in contacts:
contactsArr.append(contact.toDict())
return jsonify(contactsArr)
#app.route('/contacts/<int:id>',methods=['GET'])
def getContact(id):
contact = Contacts.query.get(id)
return jsonify(contact.toDict())
Get inspired from an answer here :
Convert sqlalchemy row object to python dict
Flask-JsonTools package has an implementation of JsonSerializableBase Base class for your models.
Usage:
from sqlalchemy.ext.declarative import declarative_base
from flask.ext.jsontools import JsonSerializableBase
Base = declarative_base(cls=(JsonSerializableBase,))
class User(Base):
#...
Now the User model is magically serializable.
If your framework is not Flask, you can just grab the code
For security reasons you should never return all the model's fields. I prefer to selectively choose them.
Flask's json encoding now supports UUID, datetime and relationships (and added query and query_class for flask_sqlalchemy db.Model class). I've updated the encoder as follows:
app/json_encoder.py
from sqlalchemy.ext.declarative import DeclarativeMeta
from flask import json
class AlchemyEncoder(json.JSONEncoder):
def default(self, o):
if isinstance(o.__class__, DeclarativeMeta):
data = {}
fields = o.__json__() if hasattr(o, '__json__') else dir(o)
for field in [f for f in fields if not f.startswith('_') and f not in ['metadata', 'query', 'query_class']]:
value = o.__getattribute__(field)
try:
json.dumps(value)
data[field] = value
except TypeError:
data[field] = None
return data
return json.JSONEncoder.default(self, o)
app/__init__.py
# json encoding
from app.json_encoder import AlchemyEncoder
app.json_encoder = AlchemyEncoder
With this I can optionally add a __json__ property that returns the list of fields I wish to encode:
app/models.py
class Queue(db.Model):
id = db.Column(db.Integer, primary_key=True)
song_id = db.Column(db.Integer, db.ForeignKey('song.id'), unique=True, nullable=False)
song = db.relationship('Song', lazy='joined')
type = db.Column(db.String(20), server_default=u'audio/mpeg')
src = db.Column(db.String(255), nullable=False)
created_at = db.Column(db.DateTime, server_default=db.func.now())
updated_at = db.Column(db.DateTime, server_default=db.func.now(), onupdate=db.func.now())
def __init__(self, song):
self.song = song
self.src = song.full_path
def __json__(self):
return ['song', 'src', 'type', 'created_at']
I add #jsonapi to my view, return the resultlist and then my output is as follows:
[
{
"created_at": "Thu, 23 Jul 2015 11:36:53 GMT",
"song":
{
"full_path": "/static/music/Audioslave/Audioslave [2002]/1 Cochise.mp3",
"id": 2,
"path_name": "Audioslave/Audioslave [2002]/1 Cochise.mp3"
},
"src": "/static/music/Audioslave/Audioslave [2002]/1 Cochise.mp3",
"type": "audio/mpeg"
}
]
A more detailed explanation.
In your model, add:
def as_dict(self):
return {c.name: str(getattr(self, c.name)) for c in self.__table__.columns}
The str() is for python 3 so if using python 2 use unicode(). It should help deserialize dates. You can remove it if not dealing with those.
You can now query the database like this
some_result = User.query.filter_by(id=current_user.id).first().as_dict()
First() is needed to avoid weird errors. as_dict() will now deserialize the result. After deserialization, it is ready to be turned to json
jsonify(some_result)
While the original question goes back awhile, the number of answers here (and my own experiences) suggest it's a non-trivial question with a lot of different approaches of varying complexity with different trade-offs.
That's why I built the SQLAthanor library that extends SQLAlchemy's declarative ORM with configurable serialization/de-serialization support that you might want to take a look at.
The library supports:
Python 2.7, 3.4, 3.5, and 3.6.
SQLAlchemy versions 0.9 and higher
serialization/de-serialization to/from JSON, CSV, YAML, and Python dict
serialization/de-serialization of columns/attributes, relationships, hybrid properties, and association proxies
enabling and disabling of serialization for particular formats and columns/relationships/attributes (e.g. you want to support an inbound password value, but never include an outbound one)
pre-serialization and post-deserialization value processing (for validation or type coercion)
a pretty straightforward syntax that is both Pythonic and seamlessly consistent with SQLAlchemy's own approach
You can check out the (I hope!) comprehensive docs here: https://sqlathanor.readthedocs.io/en/latest
Hope this helps!
Custom serialization and deserialization.
"from_json" (class method) builds a Model object based on json data.
"deserialize" could be called only on instance, and merge all data from json into Model instance.
"serialize" - recursive serialization
__write_only__ property is needed to define write only properties ("password_hash" for example).
class Serializable(object):
__exclude__ = ('id',)
__include__ = ()
__write_only__ = ()
#classmethod
def from_json(cls, json, selfObj=None):
if selfObj is None:
self = cls()
else:
self = selfObj
exclude = (cls.__exclude__ or ()) + Serializable.__exclude__
include = cls.__include__ or ()
if json:
for prop, value in json.iteritems():
# ignore all non user data, e.g. only
if (not (prop in exclude) | (prop in include)) and isinstance(
getattr(cls, prop, None), QueryableAttribute):
setattr(self, prop, value)
return self
def deserialize(self, json):
if not json:
return None
return self.__class__.from_json(json, selfObj=self)
#classmethod
def serialize_list(cls, object_list=[]):
output = []
for li in object_list:
if isinstance(li, Serializable):
output.append(li.serialize())
else:
output.append(li)
return output
def serialize(self, **kwargs):
# init write only props
if len(getattr(self.__class__, '__write_only__', ())) == 0:
self.__class__.__write_only__ = ()
dictionary = {}
expand = kwargs.get('expand', ()) or ()
prop = 'props'
if expand:
# expand all the fields
for key in expand:
getattr(self, key)
iterable = self.__dict__.items()
is_custom_property_set = False
# include only properties passed as parameter
if (prop in kwargs) and (kwargs.get(prop, None) is not None):
is_custom_property_set = True
iterable = kwargs.get(prop, None)
# loop trough all accessible properties
for key in iterable:
accessor = key
if isinstance(key, tuple):
accessor = key[0]
if not (accessor in self.__class__.__write_only__) and not accessor.startswith('_'):
# force select from db to be able get relationships
if is_custom_property_set:
getattr(self, accessor, None)
if isinstance(self.__dict__.get(accessor), list):
dictionary[accessor] = self.__class__.serialize_list(object_list=self.__dict__.get(accessor))
# check if those properties are read only
elif isinstance(self.__dict__.get(accessor), Serializable):
dictionary[accessor] = self.__dict__.get(accessor).serialize()
else:
dictionary[accessor] = self.__dict__.get(accessor)
return dictionary
Here is a solution that lets you select the relations you want to include in your output as deep as you would like to go.
NOTE: This is a complete re-write taking a dict/str as an arg rather than a list. fixes some stuff..
def deep_dict(self, relations={}):
"""Output a dict of an SA object recursing as deep as you want.
Takes one argument, relations which is a dictionary of relations we'd
like to pull out. The relations dict items can be a single relation
name or deeper relation names connected by sub dicts
Example:
Say we have a Person object with a family relationship
person.deep_dict(relations={'family':None})
Say the family object has homes as a relation then we can do
person.deep_dict(relations={'family':{'homes':None}})
OR
person.deep_dict(relations={'family':'homes'})
Say homes has a relation like rooms you can do
person.deep_dict(relations={'family':{'homes':'rooms'}})
and so on...
"""
mydict = dict((c, str(a)) for c, a in
self.__dict__.items() if c != '_sa_instance_state')
if not relations:
# just return ourselves
return mydict
# otherwise we need to go deeper
if not isinstance(relations, dict) and not isinstance(relations, str):
raise Exception("relations should be a dict, it is of type {}".format(type(relations)))
# got here so check and handle if we were passed a dict
if isinstance(relations, dict):
# we were passed deeper info
for left, right in relations.items():
myrel = getattr(self, left)
if isinstance(myrel, list):
mydict[left] = [rel.deep_dict(relations=right) for rel in myrel]
else:
mydict[left] = myrel.deep_dict(relations=right)
# if we get here check and handle if we were passed a string
elif isinstance(relations, str):
# passed a single item
myrel = getattr(self, relations)
left = relations
if isinstance(myrel, list):
mydict[left] = [rel.deep_dict(relations=None)
for rel in myrel]
else:
mydict[left] = myrel.deep_dict(relations=None)
return mydict
so for an example using person/family/homes/rooms... turning it into json all you need is
json.dumps(person.deep_dict(relations={'family':{'homes':'rooms'}}))
step1:
class CNAME:
...
def as_dict(self):
return {item.name: getattr(self, item.name) for item in self.__table__.columns}
step2:
list = []
for data in session.query(CNAME).all():
list.append(data.as_dict())
step3:
return jsonify(list)
Even though it's a old post, Maybe I didn't answer the question above, but I want to talk about my serialization, at least it works for me.
I use FastAPI,SqlAlchemy and MySQL, but I don't use orm model;
# from sqlalchemy import create_engine
# from sqlalchemy.orm import sessionmaker
# engine = create_engine(config.SQLALCHEMY_DATABASE_URL, pool_pre_ping=True)
# SessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=engine)
Serialization code
import decimal
import datetime
def alchemy_encoder(obj):
"""JSON encoder function for SQLAlchemy special classes."""
if isinstance(obj, datetime.date):
return obj.strftime("%Y-%m-%d %H:%M:%S")
elif isinstance(obj, decimal.Decimal):
return float(obj)
import json
from sqlalchemy import text
# db is SessionLocal() object
app_sql = 'SELECT * FROM app_info ORDER BY app_id LIMIT :page,:page_size'
# The next two are the parameters passed in
page = 1
page_size = 10
# execute sql and return a <class 'sqlalchemy.engine.result.ResultProxy'> object
app_list = db.execute(text(app_sql), {'page': page, 'page_size': page_size})
# serialize
res = json.loads(json.dumps([dict(r) for r in app_list], default=alchemy_encoder))
If it doesn't work, please ignore my answer. I refer to it here
https://codeandlife.com/2014/12/07/sqlalchemy-results-to-json-the-easy-way/
install simplejson by
pip install simplejson and the create a class
class Serialise(object):
def _asdict(self):
"""
Serialization logic for converting entities using flask's jsonify
:return: An ordered dictionary
:rtype: :class:`collections.OrderedDict`
"""
result = OrderedDict()
# Get the columns
for key in self.__mapper__.c.keys():
if isinstance(getattr(self, key), datetime):
result["x"] = getattr(self, key).timestamp() * 1000
result["timestamp"] = result["x"]
else:
result[key] = getattr(self, key)
return result
and inherit this class to every orm classes so that this _asdict function gets registered to every ORM class and boom.
And use jsonify anywhere
It is not so straighforward. I wrote some code to do this. I'm still working on it, and it uses the MochiKit framework. It basically translates compound objects between Python and Javascript using a proxy and registered JSON converters.
Browser side for database objects is db.js
It needs the basic Python proxy source in proxy.js.
On the Python side there is the base proxy module.
Then finally the SqlAlchemy object encoder in webserver.py.
It also depends on metadata extractors found in the models.py file.
def alc2json(row):
return dict([(col, str(getattr(row,col))) for col in row.__table__.columns.keys()])
I thought I'd play a little code golf with this one.
FYI: I am using automap_base since we have a separately designed schema according to business requirements. I just started using SQLAlchemy today but the documentation states that automap_base is an extension to declarative_base which seems to be the typical paradigm in the SQLAlchemy ORM so I believe this should work.
It does not get fancy with following foreign keys per Tjorriemorrie's solution, but it simply matches columns to values and handles Python types by str()-ing the column values. Our values consist Python datetime.time and decimal.Decimal class type results so it gets the job done.
Hope this helps any passers-by!
I know this is quite an older post. I took solution given by #SashaB and modified as per my need.
I added following things to it:
Field ignore list: A list of fields to be ignored while serializing
Field replace list: A dictionary containing field names to be replaced by values while serializing.
Removed methods and BaseQuery getting serialized
My code is as follows:
def alchemy_json_encoder(revisit_self = False, fields_to_expand = [], fields_to_ignore = [], fields_to_replace = {}):
"""
Serialize SQLAlchemy result into JSon
:param revisit_self: True / False
:param fields_to_expand: Fields which are to be expanded for including their children and all
:param fields_to_ignore: Fields to be ignored while encoding
:param fields_to_replace: Field keys to be replaced by values assigned in dictionary
:return: Json serialized SQLAlchemy object
"""
_visited_objs = []
class AlchemyEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj.__class__, DeclarativeMeta):
# don't re-visit self
if revisit_self:
if obj in _visited_objs:
return None
_visited_objs.append(obj)
# go through each field in this SQLalchemy class
fields = {}
for field in [x for x in dir(obj) if not x.startswith('_') and x != 'metadata' and x not in fields_to_ignore]:
val = obj.__getattribute__(field)
# is this field method defination, or an SQLalchemy object
if not hasattr(val, "__call__") and not isinstance(val, BaseQuery):
field_name = fields_to_replace[field] if field in fields_to_replace else field
# is this field another SQLalchemy object, or a list of SQLalchemy objects?
if isinstance(val.__class__, DeclarativeMeta) or \
(isinstance(val, list) and len(val) > 0 and isinstance(val[0].__class__, DeclarativeMeta)):
# unless we're expanding this field, stop here
if field not in fields_to_expand:
# not expanding this field: set it to None and continue
fields[field_name] = None
continue
fields[field_name] = val
# a json-encodable dict
return fields
return json.JSONEncoder.default(self, obj)
return AlchemyEncoder
Hope it helps someone!
Use the built-in serializer in SQLAlchemy:
from sqlalchemy.ext.serializer import loads, dumps
obj = MyAlchemyObject()
# serialize object
serialized_obj = dumps(obj)
# deserialize object
obj = loads(serialized_obj)
If you're transferring the object between sessions, remember to detach the object from the current session using session.expunge(obj).
To attach it again, just do session.add(obj).
Under Flask, this works and handles datatime fields, transforming a field of type
'time': datetime.datetime(2018, 3, 22, 15, 40) into
"time": "2018-03-22 15:40:00":
obj = {c.name: str(getattr(self, c.name)) for c in self.__table__.columns}
# This to get the JSON body
return json.dumps(obj)
# Or this to get a response object
return jsonify(obj)
following code will serialize sqlalchemy result to json.
import json
from collections import OrderedDict
def asdict(self):
result = OrderedDict()
for key in self.__mapper__.c.keys():
if getattr(self, key) is not None:
result[key] = str(getattr(self, key))
else:
result[key] = getattr(self, key)
return result
def to_array(all_vendors):
v = [ ven.asdict() for ven in all_vendors ]
return json.dumps(v)
Calling fun,
def all_products():
all_products = Products.query.all()
return to_array(all_products)
The AlchemyEncoder is wonderful but sometimes fails with Decimal values. Here is an improved encoder that solves the decimal problem -
class AlchemyEncoder(json.JSONEncoder):
# To serialize SQLalchemy objects
def default(self, obj):
if isinstance(obj.__class__, DeclarativeMeta):
model_fields = {}
for field in [x for x in dir(obj) if not x.startswith('_') and x != 'metadata']:
data = obj.__getattribute__(field)
print data
try:
json.dumps(data) # this will fail on non-encodable values, like other classes
model_fields[field] = data
except TypeError:
model_fields[field] = None
return model_fields
if isinstance(obj, Decimal):
return float(obj)
return json.JSONEncoder.default(self, obj)
When using sqlalchemy to connect to a db I this is a simple solution which is highly configurable. Use pandas.
import pandas as pd
import sqlalchemy
#sqlalchemy engine configuration
engine = sqlalchemy.create_engine....
def my_function():
#read in from sql directly into a pandas dataframe
#check the pandas documentation for additional config options
sql_DF = pd.read_sql_table("table_name", con=engine)
# "orient" is optional here but allows you to specify the json formatting you require
sql_json = sql_DF.to_json(orient="index")
return sql_json
(Tiny tweak on Sasha B's really excellent answer)
This specifically converts datetime objects to strings which in the original answer would be converted to None:
# Standard library imports
from datetime import datetime
import json
# 3rd party imports
from sqlalchemy.ext.declarative import DeclarativeMeta
class JsonEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj.__class__, DeclarativeMeta):
dict = {}
# Remove invalid fields and just get the column attributes
columns = [x for x in dir(obj) if not x.startswith("_") and x != "metadata"]
for column in columns:
value = obj.__getattribute__(column)
try:
json.dumps(value)
dict[column] = value
except TypeError:
if isinstance(value, datetime):
dict[column] = value.__str__()
else:
dict[column] = None
return dict
return json.JSONEncoder.default(self, obj)
class SqlToDict:
def __init__(self, data) -> None:
self.data = data
def to_timestamp(self, date):
if isinstance(date, datetime):
return int(datetime.timestamp(date))
else:
return date
def to_dict(self) -> List:
arr = []
for i in self.data:
keys = [*i.keys()]
values = [*i]
values = [self.to_timestamp(d) for d in values]
arr.append(dict(zip(keys, values)))
return arr
For example:
SqlToDict(data).to_dict()
Very late 2023
My implementation
def obj_to_dict(obj, remove=['_sa_instance_state'], debug=False):
result = {}
if type(obj).__name__ == "Row":
return dict(obj)
obj = obj.__dict__
for key in obj:
if key in remove:
continue
result[key] = obj[key]
if debug:
print(result)
return result
The built in serializer chokes with utf-8 cannot decode invalid start byte for some inputs. Instead, I went with:
def row_to_dict(row):
temp = row.__dict__
temp.pop('_sa_instance_state', None)
return temp
def rows_to_list(rows):
ret_rows = []
for row in rows:
ret_rows.append(row_to_dict(row))
return ret_rows
#website_blueprint.route('/api/v1/some/endpoint', methods=['GET'])
def some_api():
'''
/some_endpoint
'''
rows = rows_to_list(SomeModel.query.all())
response = app.response_class(
response=jsonplus.dumps(rows),
status=200,
mimetype='application/json'
)
return response
Maybe you can use a class like this
from sqlalchemy.ext.declarative import declared_attr
from sqlalchemy import Table
class Custom:
"""Some custom logic here!"""
__table__: Table # def for mypy
#declared_attr
def __tablename__(cls): # pylint: disable=no-self-argument
return cls.__name__ # pylint: disable= no-member
def to_dict(self) -> Dict[str, Any]:
"""Serializes only column data."""
return {c.name: getattr(self, c.name) for c in self.__table__.columns}
Base = declarative_base(cls=Custom)
class MyOwnTable(Base):
#COLUMNS!
With that all objects have the to_dict method
While using some raw sql and undefined objects, using cursor.description appeared to get what I was looking for:
with connection.cursor() as cur:
print(query)
cur.execute(query)
for item in cur.fetchall():
row = {column.name: item[i] for i, column in enumerate(cur.description)}
print(row)
This is a JSONEncoder version that preserves model column order and only keeps recursively defined column and relationship fields. It also formats most JSON unserializable types:
import json
from datetime import datetime
from decimal import Decimal
import arrow
from sqlalchemy.ext.declarative import DeclarativeMeta
class SQLAlchemyJSONEncoder(json.JSONEncoder):
"""
SQLAlchemy ORM JSON Encoder
If you have a "backref" relationship defined in your SQLAlchemy model,
this encoder raises a ValueError to stop an infinite loop.
"""
def default(self, obj):
if isinstance(obj, datetime):
return arrow.get(obj).isoformat()
elif isinstance(obj, Decimal):
return float(obj)
elif isinstance(obj, set):
return sorted(obj)
elif isinstance(obj.__class__, DeclarativeMeta):
for attribute, relationship in obj.__mapper__.relationships.items():
if isinstance(relationship.__getattribute__("backref"), tuple):
raise ValueError(
f'{obj.__class__} object has a "backref" relationship '
"that would cause an infinite loop!"
)
dictionary = {}
column_names = [column.name for column in obj.__table__.columns]
for key in column_names:
value = obj.__getattribute__(key)
if isinstance(value, datetime):
value = arrow.get(value).isoformat()
elif isinstance(value, Decimal):
value = float(value)
elif isinstance(value, set):
value = sorted(value)
dictionary[key] = value
for key in [
attribute
for attribute in dir(obj)
if not attribute.startswith("_")
and attribute != "metadata"
and attribute not in column_names
]:
value = obj.__getattribute__(key)
dictionary[key] = value
return dictionary
return super().default(obj)

Unable to call dictionary dataclass attribute with cls in function but can if I use class name

I'm trying to run the below code and getting some strange errors... Wondering if anyone can help!
import feedparser
from dataclasses import dataclass, field
from typing import Dict
#dataclass
class NewsParser:
shipping_news: str = "shippingurlgoeshere"
all_shipping_news: Dict = field(default_factory=lambda: {
'shipping news': "shippingurlgoeshere"
})
#classmethod
def parse_feed(cls):
"""
Parses a news feed for relevant information.
:return: relevant_items_as_dictionary, dictionary (unsurprisingly) which contains lists of the relevant items.
"""
# Get articles from feed
content = feedparser.parse(cls.all_shipping_news["shipping news"])
articles = content.entries
# Parse feed
titles = [x.title for x in articles]
urls = [x.link for x in articles]
pub_date_string = [x.published for x in articles]
pub_date_tuple = [x.published_parsed for x in articles]
html_summaries = [x.summary for x in articles]
relevant_items_as_dictionary = {"titles": titles, "urls": urls, "pub_date_string": pub_date_string,
"pub_date_tuple": pub_date_tuple, "html_summaries": html_summaries}
return relevant_items_as_dictionary
NewsParser.parse_feed()
This gives me the error AttributeError: type object 'NewsParser' has no attribute 'all_shipping_news'
But if I comment out NewsParser.parse_feed() and just print NewsParser after instantiating it, I get the correct result:
NewsParser(shipping_news='shippingurlgoeshere', all_shipping_news={'shipping news': 'shippingurlgoeshere'})
But if instead of writing cls.all_shipping_news I write NewsParser.all_shipping_news it works fine.
Last piece of the puzzle... If in the parse_feed method I write cls.shipping_news then it can access the shipping_news string perfectly, but as above this doesn't happen when I write cls.all_shipping_news

Peewee model to JSON

I'm creating an API using peewee as the ORM and I need the ability to convert a peewee model object into a JSON object to send to the user. Does anyone know of a good way to do this?
Peewee has a model_to_dict and dict_to_model helpers in the playhouse.shortcuts extension module.
http://docs.peewee-orm.com/en/latest/peewee/playhouse.html#model_to_dict
http://docs.peewee-orm.com/en/latest/peewee/playhouse.html#dict_to_model
You could use these as follows:
from playhouse.shortcuts import model_to_dict, dict_to_model
user_obj = User.select().where(User.username == 'charlie').get()
json_data = json.dumps(model_to_dict(user_obj))
Also note that model_to_dict() can recurse through related models, include back-referenced models, and exclude certain fields from being serialized.
when single fetch
user = User.select().where(User.id == 1).get()
model_to_dict(user) #to Dict
when Multiple fetch
users = list(User.select().where(User.name ** 'a%').dicts())
also, you can get model as a dict, and then convert to json with correct field types (bool, int, float, etc.):
import peewee
import json
from bson import json_util
from datetime import datetime
class User(peewee.Model):
email = CharField()
status = BooleanField(default=True)
firstname = CharField()
lastname = CharField()
age = IntegerField()
created = DateTimeField(default=datetime.now())
class Meta:
database = db
user = User.select().dicts().get()
print json.dumps(user, default=json_util.default)
For anybody having issues like TypeError: Object of type date is not JSON serializable, this works for me (tested on Python 3.8.2).
from playhouse.shortcuts import model_to_dict
import json
def print_model(model):
print(json.dumps(model_to_dict(model), indent=4, sort_keys=True, default=str))
def print_models(models):
print(json.dumps(list(models.dicts()), indent=4, sort_keys=True, default=str))
Usage 1 - Single model
for person in Person.select():
print_model(person)
Usage 2 - Many models
print_models(Person.select())
I had this very same problem and ended up defining my own parser extension for JSON types that could not be automatically serialized. I'm fine for now in using strings as data represented (although you could possibly use different datatypes, but beware of approximation using floating points!
In the following example, I put this in a file called json_serialize.py inside a utils folder:
from decimal import Decimal
import datetime
try:
import uuid
_use_uuid = True
except ImportError:
_use_uuid = False
datetime_format = "%Y/%m/%d %H:%M:%S"
date_format = "%Y/%m/%d"
time_format = "%H:%M:%S"
def set_datetime_format(fmt_string):
datetime_format = fmt_string
def set_date_format(fmt_string):
date_format = fmt_string
def set_time_format(fmt_string):
time_format = fmt_string
def more(obj):
if isinstance(obj, Decimal):
return str(obj)
if isinstance(obj, datetime.datetime):
return obj.strftime(datetime_format)
if isinstance(obj, datetime.date):
return obj.strftime(date_format)
if isinstance(obj, datetime.time):
return obj.strftime(time_format)
if _use_uuid and isinstance(obj, uuid.UUID):
return str(obj.db_value())
raise TypeError("%r is not JSON serializable" % obj)
Then, in my app:
import json
from utils import json_serialize
...
json.dumps(model_to_dict(User.get()), default=json_serialize.more)
edit just to add: this is very largely inspired by json_utils.default module found in mongodb but mainly relies on the json module and needs no import of mongodb own bson/json_utils module.
Usually I update it to support new types as soon as my app raises the TypeError for it found a type not able to serialize
I usually implement the model to dict and dict to model functions, for maximum security and understanding of the inner workings of the code. Peewee does a lot of magic and you want to be in control over it.
The most obvious argument for why you should not iterate on the fields but rather explicitly specify them is because of security considerations. Not all fields can be exposed to the user, and I assume you need this functionality to implement some sort of REST API.
So - you should do something like this:
class UserData(db.Model):
user = db.ForeignKeyField(User)
data = db.CharField()
def serialize():
# front end does not need user ID here
return {
'data': self.data
}
#classmethod
def from_json(cls, json_data):
UserData.create(
# we enforce user to be the current user
user=current_user,
data=json_data['data']
)
you can do something like that:
class MyModel(peewee.Model):
def __str__(self):
r = {}
for k in self._data.keys():
try:
r[k] = str(getattr(self, k))
except:
r[k] = json.dumps(getattr(self, k))
return str(r)
class User(MyModel):
email = CharField()
status = CharField(default="enabled")
firstname = CharField()
lastname = CharField()
class Meta:
database = db

How would you inherit from and override the django model classes to create a listOfStringsField?

I want to create a new type of field for django models that is basically a ListOfStrings. So in your model code you would have the following:
models.py:
from django.db import models
class ListOfStringsField(???):
???
class myDjangoModelClass():
myName = models.CharField(max_length=64)
myFriends = ListOfStringsField() #
other.py:
myclass = myDjangoModelClass()
myclass.myName = "bob"
myclass.myFriends = ["me", "myself", "and I"]
myclass.save()
id = myclass.id
loadedmyclass = myDjangoModelClass.objects.filter(id__exact=id)
myFriendsList = loadedclass.myFriends
# myFriendsList is a list and should equal ["me", "myself", "and I"]
How would you go about writing this field type, with the following stipulations?
We don't want to do create a field which just crams all the strings together and separates them with a token in one field like this. It is a good solution in some cases, but we want to keep the string data normalized so tools other than django can query the data.
The field should automatically create any secondary tables needed to store the string data.
The secondary table should ideally have only one copy of each unique string. This is optional, but would be nice to have.
Looking in the Django code it looks like I would want to do something similar to what ForeignKey is doing, but the documentation is sparse.
This leads to the following questions:
Can this be done?
Has it been done (and if so where)?
Is there any documentation on Django about how to extend and override their model classes, specifically their relationship classes? I have not seen a lot of documentation on that aspect of their code, but there is this.
This is comes from this question.
There's some very good documentation on creating custom fields here.
However, I think you're overthinking this. It sounds like you actually just want a standard foreign key, but with the additional ability to retrieve all the elements as a single list. So the easiest thing would be to just use a ForeignKey, and define a get_myfield_as_list method on the model:
class Friends(model.Model):
name = models.CharField(max_length=100)
my_items = models.ForeignKey(MyModel)
class MyModel(models.Model):
...
def get_my_friends_as_list(self):
return ', '.join(self.friends_set.values_list('name', flat=True))
Now calling get_my_friends_as_list() on an instance of MyModel will return you a list of strings, as required.
What you have described sounds to me really similar to the tags.
So, why not using django tagging?
It works like a charm, you can install it independently from your application and its API is quite easy to use.
I also think you're going about this the wrong way. Trying to make a Django field create an ancillary database table is almost certainly the wrong approach. It would be very difficult to do, and would likely confuse third party developers if you are trying to make your solution generally useful.
If you're trying to store a denormalized blob of data in a single column, I'd take an approach similar to the one you linked to, serializing the Python data structure and storing it in a TextField. If you want tools other than Django to be able to operate on the data then you can serialize to JSON (or some other format that has wide language support):
from django.db import models
from django.utils import simplejson
class JSONDataField(models.TextField):
__metaclass__ = models.SubfieldBase
def to_python(self, value):
if value is None:
return None
if not isinstance(value, basestring):
return value
return simplejson.loads(value)
def get_db_prep_save(self, value):
if value is None:
return None
return simplejson.dumps(value)
If you just want a django Manager-like descriptor that lets you operate on a list of strings associated with a model then you can manually create a join table and use a descriptor to manage the relationship. It's not exactly what you need, but this code should get you started.
Thanks for all those that answered. Even if I didn't use your answer directly the examples and links got me going in the right direction.
I am not sure if this is production ready, but it appears to be working in all my tests so far.
class ListValueDescriptor(object):
def __init__(self, lvd_parent, lvd_model_name, lvd_value_type, lvd_unique, **kwargs):
"""
This descriptor object acts like a django field, but it will accept
a list of values, instead a single value.
For example:
# define our model
class Person(models.Model):
name = models.CharField(max_length=120)
friends = ListValueDescriptor("Person", "Friend", "CharField", True, max_length=120)
# Later in the code we can do this
p = Person("John")
p.save() # we have to have an id
p.friends = ["Jerry", "Jimmy", "Jamail"]
...
p = Person.objects.get(name="John")
friends = p.friends
# and now friends is a list.
lvd_parent - The name of our parent class
lvd_model_name - The name of our new model
lvd_value_type - The value type of the value in our new model
This has to be the name of one of the valid django
model field types such as 'CharField', 'FloatField',
or a valid custom field name.
lvd_unique - Set this to true if you want the values in the list to
be unique in the table they are stored in. For
example if you are storing a list of strings and
the strings are always "foo", "bar", and "baz", your
data table would only have those three strings listed in
it in the database.
kwargs - These are passed to the value field.
"""
self.related_set_name = lvd_model_name.lower() + "_set"
self.model_name = lvd_model_name
self.parent = lvd_parent
self.unique = lvd_unique
# only set this to true if they have not already set it.
# this helps speed up the searchs when unique is true.
kwargs['db_index'] = kwargs.get('db_index', True)
filter = ["lvd_parent", "lvd_model_name", "lvd_value_type", "lvd_unique"]
evalStr = """class %s (models.Model):\n""" % (self.model_name)
evalStr += """ value = models.%s(""" % (lvd_value_type)
evalStr += self._params_from_kwargs(filter, **kwargs)
evalStr += ")\n"
if self.unique:
evalStr += """ parent = models.ManyToManyField('%s')\n""" % (self.parent)
else:
evalStr += """ parent = models.ForeignKey('%s')\n""" % (self.parent)
evalStr += "\n"
evalStr += """self.innerClass = %s\n""" % (self.model_name)
print evalStr
exec (evalStr) # build the inner class
def __get__(self, instance, owner):
value_set = instance.__getattribute__(self.related_set_name)
l = []
for x in value_set.all():
l.append(x.value)
return l
def __set__(self, instance, values):
value_set = instance.__getattribute__(self.related_set_name)
for x in values:
value_set.add(self._get_or_create_value(x))
def __delete__(self, instance):
pass # I should probably try and do something here.
def _get_or_create_value(self, x):
if self.unique:
# Try and find an existing value
try:
return self.innerClass.objects.get(value=x)
except django.core.exceptions.ObjectDoesNotExist:
pass
v = self.innerClass(value=x)
v.save() # we have to save to create the id.
return v
def _params_from_kwargs(self, filter, **kwargs):
"""Given a dictionary of arguments, build a string which
represents it as a parameter list, and filter out any
keywords in filter."""
params = ""
for key in kwargs:
if key not in filter:
value = kwargs[key]
params += "%s=%s, " % (key, value.__repr__())
return params[:-2] # chop off the last ', '
class Person(models.Model):
name = models.CharField(max_length=120)
friends = ListValueDescriptor("Person", "Friend", "CharField", True, max_length=120)
Ultimately I think this would still be better if it were pushed deeper into the django code and worked more like the ManyToManyField or the ForeignKey.
I think what you want is a custom model field.

Categories

Resources