WTForms integration

Important

Documentation below is related to project version 2.0.0 or higher, old versions has completely different approach for forms generation.

And despite the fact that the old code is included in version 2.0.0 to keep correct deprecation workflow (where possible), it is not documented (and was not) and not maintained.

If you faced any forms problems, consider migration to new methods and approach.

Flask-Mongoengine and Flask-WTF/WTForms are heavily integrated, to reduce amount of boilerplate code, required to make database model and online form. In the same time a lot of options was created to keep extreme flexibility.

After database model definition user does not require to repeat same code in form definition, instead it is possible to use integrated converter, that will do most of the work.

Flask-Mongoengine will transform some model’s properties to Flask-WTF/WTForms validators, so user does not need to care about standards. For full list of transformations, please review global transforms and specific field documentation below.

In the same time, user is able to adjust database fields definition with specific settings as on stage of Document model definition, as on form generation stage. This allows to create several forms for same model, for different circumstances.

Requirements

For correct integration behavior several requirements should be met:

Global transforms

For all fields, processed by Flask-Mongoengine integration:

  • If model field definition have wtf_validators defined, they will be forwarded to WTForm as validators. This is not protection from validators extension by Flask-Mongoengine.

  • If model field definition have wtf_filters defined, they will be forwarded to WTForm as filters.

  • If model field definition have required, then InputRequired will be added to form validators, otherwise Optional added.

  • If model field definition have verbose_name it will be used as form field label, otherwise pure field name used.

  • If model field definition have help_text it will be used as form field description, otherwise empty string used.

  • Field’s default used as form default, that’s why special WTForms fields implementations was created. Details can be found in flask_mongoengine.wtf.fields module. In new form generator only ‘Mongo’ prefixed classes are used for fields, other classes are deprecated and will be removed in version 3.0.0. If you have own nesting classes, you should check inheritance and make an update.

  • Field’s choices, if exist, used as form choices.

Warning

As at version 2.0.0 there is no wtf_validators duplicates/conflicts check. User should be careful with manual wtf_validators setup. And in case of forms problems this is first place to look on.

wtf_validators and wtf_filters duplication check expected in future versions; PRs are welcome.

Some additional transformations are made by specific field, check exact field documentation below for more info.

BinaryField

Not yet documented. Please help us with new pull request.

BooleanField

Form generation behaviour

BooleanField is very complicated in terms of Mongo database support. In Flask-Mongoengine before version 2.0.0+ database BooleanField used wtforms.fields.BooleanField as form representation, this raised several not clear problems, that was related to how wtforms.fields.BooleanField parse and work with form values. Known problems in version, before 2.0.0+:

  • Default value of field, specified in database definition was ignored, if default is None and nulls allowed, i.e. null=True (Value was always False).

  • Field was always created in database document, even if not checked, as there is impossible to split None and False values, when only checkbox available.

To fix all these issues, and do not create database field by default, Flask-Mongoengine 2.0.0+ uses dropdown field by default.

By default, database BooleanField not allowing None value, meaning that field can be True, False or not created in database at all. If database field configuration allowing None values, i.e. null=True, then, when nothing selected in dropdown, the field will be created with None value.

Important

It is responsobility of developer, to correctly setup database field definition and make proper tests before own application release. BooleanField can create unexpected application behavior in if checks. Developer, should recheck all if checks like:

  • if filed_value: this will match True database value

  • if not filed_value: this will match False or None database value or not existing document key

  • if field_value is None: this will match None database value or not existing document key

  • if field_value is True: this will match True database value

  • if field_value is False: this will match False database value

  • if field_value is not None: this will match True, False database value

  • if field_value is not True: this will match False, None database value or not existing document key

  • if filed_value is not False: this will match True, None database value or not existing document key

Examples

BooleanField with default dropdown

Such definition will not create any field in document, if dropdown not selected.

"""boolean_demo.py"""
from example_app.models import db


class BooleanDemoModel(db.Document):
    """Documentation example model."""

    boolean_field = db.BooleanField()

BooleanField with allowed None value

Such definition will create document field, even if nothing selected. The value will be None. If, during edit, yes or no dropdown values replaced to ---, then saved value in document will be also changed to None.

By default, None value represented as --- text in dropdown.

"""boolean_demo.py"""
from example_app.models import db


class BooleanDemoModel(db.Document):
    """Documentation example model."""

    boolean_field_with_null = db.BooleanField(null=True)

BooleanField with replaced dropdown text

Dropdown text can be easily replaced, there is only one requirement: New choices, should be correctly coerced by coerce_boolean(), or function should be replaced too.

"""boolean_demo.py"""
from example_app.models import db


class BooleanDemoModel(db.Document):
    """Documentation example model."""

    boolean_field_with_as_choices_replace = db.BooleanField(
        wtf_options={
            "choices": [("", "Not selected"), ("yes", "Positive"), ("no", "Negative")]
        }
    )

BooleanField with default True value, but with allowed nulls

"""boolean_demo.py"""
from example_app.models import db


class BooleanDemoModel(db.Document):
    """Documentation example model."""

    true_boolean_field_with_allowed_null = db.BooleanField(default=True, null=True)

ComplexDateTimeField

Form generation behaviour

ComplexDateTimeField stores date and time information in database string format. This format allow precision up to microseconds dimension.

Unfortunately, there is no HTML5 field, that allow so high precision. That’s why, by default the generated field will use HTML5 <input type="datetime-local"> with precision set to milliseconds.

If you require concrete microseconds for edit purposes, please use wtforms.fields.DateTimeField with correct format (see examples below).

Field is easy adjustable, to use any other precision. Check examples and example app for more details.

Examples

dates_demo.py in example app contain basic non-requirement example. You can adjust it to any provided example for test purposes.

ComplexDateTimeField with milliseconds precision

"""dates_demo.py"""
from example_app.models import db


class DateTimeModel(db.Document):
    """Documentation example model."""

    complex_datetime = db.ComplexDateTimeField()

ComplexDateTimeField with seconds precision

"""dates_demo.py"""
from example_app.models import db


class DateTimeModel(db.Document):
    """Documentation example model."""

    complex_datetime_sec = db.ComplexDateTimeField(
        wtf_options={"render_kw": {"step": "1"}}
    )

ComplexDateTimeField with microseconds precision (text)

"""dates_demo.py"""
from wtforms.fields import DateTimeField

from example_app.models import db


class DateTimeModel(db.Document):
    """Documentation example model."""

    complex_datetime_microseconds = db.ComplexDateTimeField(
        wtf_field_class=DateTimeField, wtf_options={"format": "%Y-%m-%d %H:%M:%S.%f"}
    )

DateField

Form generation behaviour

DateField is one of the simplest fields in the forms generation process. By default, the field use wtforms.fields.DateField WTForms class, representing a form input with standard HTML5 <input type="date">. No custom additional transformation done, during field generation. Field is fully controllable by global transforms.

Examples

dates_demo.py in example app contain basic non-requirement example. You can adjust it to any provided example for test purposes.

Not limited DateField

"""dates_demo.py"""
from example_app.models import db


class DateTimeModel(db.Document):
    """Documentation example model."""

    date = db.DateField()

DateTimeField

Form generation behaviour

DateTimeField stores date and time information in database date format. This format allow precision up to milliseconds dimension. By default, generated form will use HTML5 <input type="datetime-local"> with precision set to seconds.

Field is easy adjustable, to use any other precision. Check examples and example app for more details.

It is possible to use wtforms.fields.DateTimeField for text input behaviour.

Examples

dates_demo.py in example app contain basic non-requirement example. You can adjust it to any provided example for test purposes.

DateTimeField with seconds precision

"""dates_demo.py"""
from example_app.models import db


class DateTimeModel(db.Document):
    """Documentation example model."""

    datetime = db.DateTimeField()

DateTimeField without seconds

"""dates_demo.py"""
from example_app.models import db


class DateTimeModel(db.Document):
    """Documentation example model."""

    datetime_no_sec = db.DateTimeField(wtf_options={"render_kw": {"step": "60"}})

DateTimeField with milliseconds precision

"""dates_demo.py"""
from example_app.models import db


class DateTimeModel(db.Document):
    """Documentation example model."""

    datetime_ms = db.DateTimeField(wtf_options={"render_kw": {"step": "0.001"}})

DecimalField

Form generation behaviour

From form generation side this field is pretty standard and do not use any form generation adjustments.

If database field definition has any of min_value or max_value, then NumberRange validator will be added to form field.

Examples

numbers_demo.py in example app contain basic non-requirement example. You can adjust it to any provided example for test purposes.

Not limited DecimalField

"""numbers_demo.py"""
from example_app.models import db


class NumbersDemoModel(db.Document):
    """Documentation example model."""

    decimal_field_unlimited = db.DecimalField()

Limited DecimalField

"""numbers_demo.py"""
from decimal import Decimal

from example_app.models import db


class NumbersDemoModel(db.Document):
    """Documentation example model."""

    decimal_field_limited = db.DecimalField(
        min_value=Decimal("1"), max_value=Decimal("200.455")
    )

DictField

DictField has Object type in terms of Mongo database itself, so basically it defines document inside document, but without pre-defined structure. That’s why this is one of fields, that has default value specified inside Mongoengine itself, and that’s why is always (almost) created.

The developer should understand that database keyword argument default is forwarded to form by default, but can be separately overwritten in form. This brings a lot of options for form field configuration.

Also, should be additionally noted that database Null value in form is represented as empty string. Non-existing field is represented with form default for new forms (without instance inside) or with empty string for non-empty forms.

Complicated? Probably. That’s why this field was completely rewritten in version 2.0.0. Check examples, and everything will be clear.

Form generation behaviour

Our default form generation follow Mongoengine internals and will use database field default (empty dict) to populate to new form or to not filled field in existing form.

In the same time, we are allowing extending of this behaviour, and not creating field in database, if default value provided as None. In this case, generated field for new form will be empty, without any pre-filled value.

Same empty field will be displayed in case, when both default=None and null=True selected, during database form initialization. In this case form field will be empty, without any placeholder, but on save null object will be created in document.

Also, we even support separated defaults for form field and database field, allowing any form+database behaviour.

Examples

DictField with default empty dict value

Will place {} to form for existing/new fields. This value is hardcodded in parent MongoEngine project.

"""dict_demo.py"""
from example_app.models import db


class DictDemoModel(db.Document):
    """Documentation example model."""

    dict_field = db.DictField()

DictField with default None value, ignored by database

Reminder: Such field is empty in form, and will not create anything in database if not filled.

"""dict_demo.py"""
from example_app.models import db


class DictDemoModel(db.Document):
    """Documentation example model."""

    no_dict_field = db.DictField(default=None)

DictField with default None value, saved to database

Reminder: Such field is empty in form, and will create null object in database if not filled.

"""dict_demo.py"""
from example_app.models import db


class DictDemoModel(db.Document):
    """Documentation example model."""

    null_dict_field = db.DictField(default=None, null=True)

DictField with pre-defined default dict

This value is pre-defined on database level. So behaviour of form and in-code creation of such objects will be the same - default dict will be saved to database, if nothing provided to form/instance. Form will be pre-filled with default dict.

"""dict_demo.py"""
from example_app.models import db


def get_default_dict():
    """Example of default dict specification."""
    return {"alpha": 1, "text": "text", "float": 1.2}


class DictDemoModel(db.Document):
    """Documentation example model."""

    dict_default = db.DictField(default=get_default_dict)

DictField with pre-defined value and no document object creation on Null

This is a case when you do not want to create any record in database document, if user completely delete pre-filled value in new document form. Here we use different null and default values in form field generation and during database object generation.

"""dict_demo.py"""
from example_app.models import db


def get_default_dict():
    """Example of default dict specification."""
    return {"alpha": 1, "text": "text", "float": 1.2}


class DictDemoModel(db.Document):
    """Documentation example model."""

    no_dict_prefilled = db.DictField(
        default=None,
        null=False,
        wtf_options={"default": get_default_dict, "null": True},
    )

DictField with pre-defined default dict with Null fallback

This is very rare case, when some default value is given, meaning that this value will be populated to the field, but if completely deleted, than Null will be saved in database.

"""dict_demo.py"""
from example_app.models import db


def get_default_dict():
    """Example of default dict specification."""
    return {"alpha": 1, "text": "text", "float": 1.2}


class DictDemoModel(db.Document):
    """Documentation example model."""

    null_dict_default = db.DictField(default=get_default_dict, null=True)

EmailField

Form generation behaviour

Unlike StringField WTForm class of the field is not adjusted by normal form generation sequence and always match MongoEmailField. All other adjustments, related to validators insert are work with EmailField in the same way, as in StringField.

Additional Email validator is also inserted to form field, to exclude unnecessary database request, if form data incorrect.

Field respect user’s adjustments in wtf_field_class option of db_fields.EmailField. This will change form field display, but will not change inserted validators.

Examples

strings_demo.py in example app contain basic non-requirement example. You can adjust it to any provided example for test purposes.

Not required EmailField

"""strings_demo.py"""
from example_app.models import db


class StringsDemoModel(db.Document):
    """Documentation example model."""

    url_field = db.EmailField()

Required EmailField

"""strings_demo.py"""
from example_app.models import db


class StringsDemoModel(db.Document):
    """Documentation example model."""

    required_url_field = db.EmailField(required=True)

EmbeddedDocumentField

Not yet documented. Please help us with new pull request.

FileField

Not yet documented. Please help us with new pull request.

FloatField

Changed in version 2.0.0: Default form field class changed from: wtforms.fields.FloatField to MongoFloatField.

Form generation behaviour

For Mongo database FloatField special WTForm field was created. This field’s behaviour is the same, as for wtforms.fields.FloatField, but the widget is replaced to NumberInput, this should make a look of generated form better. It is possible, that in some cases usage of base, wtforms.fields.FloatField can be required by form design. Both fields are completely compatible, and replace can be done with wtf_field_class db form parameter.

If database field definition has any of min_value or max_value, then NumberRange validator will be added to form field.

Examples

numbers_demo.py in example app contain basic non-requirement example. You can adjust it to any provided example for test purposes.

Not limited FloatField

"""numbers_demo.py"""
from example_app.models import db


class NumbersDemoModel(db.Document):
    """Documentation example model."""

    float_field_unlimited = db.FloatField()

Limited FloatField

"""numbers_demo.py"""
from example_app.models import db


class NumbersDemoModel(db.Document):
    """Documentation example model."""

    float_field_limited = db.FloatField(min_value=float(1), max_value=200.455)

IntField

Form generation behaviour

From form generation side this field is pretty standard and do not use any form generation adjustments.

If database field definition has any of min_value or max_value, then NumberRange validator will be added to form field.

Examples

numbers_demo.py in example app contain basic non-requirement example. You can adjust it to any provided example for test purposes.

Not limited IntField

"""numbers_demo.py"""
from example_app.models import db


class NumbersDemoModel(db.Document):
    """Documentation example model."""

    integer_field_unlimited = db.IntField()

Limited IntField

"""numbers_demo.py"""
from example_app.models import db


class NumbersDemoModel(db.Document):
    """Documentation example model."""

    integer_field_limited = db.IntField(min_value=1, max_value=200)

ListField

Not yet documented. Please help us with new pull request.

ReferenceField

Not yet documented. Please help us with new pull request.

SortedListField (partly?)

Not yet documented. Please help us with new pull request.

StringField

Form generation behaviour

By default, during WTForm generation for fields without specified size ( min_length or max_length) class MongoTextAreaField is used, in case when min_length or max_length set, then MongoStringField used and Length will be added to form field validators. This allows to keep documents of any size in mongodb.

In some cases class MongoStringField is not the best choice for field, even with limited size. In this case user can easily overwrite generated field class by providing wtf_field_class on db_fields.StringField field declaration, as on document, as well as on form generation steps.

If database field definition has regex parameter set, then Regexp validator will be added to the form field.

Features deprecated

Field declaration step keyword arguments password and textarea are deprecated in Flask-Mongoengine version 2.0.0 and exist only to make migration steps easy.

To implement same behaviour, user should use wtf_field_class setting on db_fields.StringField init.

Examples

strings_demo.py in example app contain basic non-requirement example. You can adjust it to any provided example for test purposes.

Not limited StringField as MongoTextAreaField

"""strings_demo.py"""
from example_app.models import db


class StringsDemoModel(db.Document):
    """Documentation example model."""

    string_field = db.StringField()

Not limited StringField as MongoTelField

"""strings_demo.py"""
from example_app.models import db
from flask_mongoengine.wtf import fields as mongo_fields


class StringsDemoModel(db.Document):
    """Documentation example model."""

    tel_field = db.StringField(wtf_field_class=mongo_fields.MongoTelField)

Not limited StringField as MongoTextAreaField with https regex

mongoengine and wtforms projects are not consistent in how they work with regex. You will be safe, if you use re.compile() each time, when you work with regex settings, before parent projects itself.

"""strings_demo.py"""
import re

from example_app.models import db


class StringsDemoModel(db.Document):
    """Documentation example model."""

    regexp_string_field = db.StringField(regex=re.compile(
        r"^(https:\/\/)[\w.-]+(?:\.[\w\.-]+)+[\w\-\._~:/?#[\]@!\$&'\(\)\*\+,;=]+$"
    ))

Size limited StringField as MongoStringField

"""strings_demo.py"""
from example_app.models import db


class StringsDemoModel(db.Document):
    """Documentation example model."""

    sized_string_field = db.StringField(min_length=5)

Required password field with minimum size

"""strings_demo.py"""
from example_app.models import db
from flask_mongoengine.wtf import fields as mongo_fields


class StringsDemoModel(db.Document):
    """Documentation example model."""

    password_field = db.StringField(
        wtf_field_class=mongo_fields.MongoPasswordField,
        required=True,
        min_length=5,
    )

URLField

Form generation behaviour

Unlike StringField WTForm class of the field is not adjusted by normal form generation sequence and always match MongoURLField. All other adjustments, related to validators insert are work with EmailField in the same way, as in StringField.

Additional Regexp validator is also inserted to form field, to exclude unnecessary database request, if form data incorrect. This validator use regexp, provided in url_regex of db_fields.URLField, or default URL regexp from mongoengine project. This is different from Flask-Mongoengine version 1.0.0 or earlier, where URL was inserted. This was changed, to exclude validators conflicts.

Important

model_form() is still use URL for compatibility with old setups.

Field respect user’s adjustments in wtf_field_class option of db_fields.URLField. This will change form field display, but will not change inserted validators.

Examples

strings_demo.py in example app contain basic non-requirement example. You can adjust it to any provided example for test purposes.

Not required URLField

"""strings_demo.py"""
from example_app.models import db


class StringsDemoModel(db.Document):
    """Documentation example model."""

    url_field = db.URLField()

Required URLField with minimum size

"""strings_demo.py"""
from example_app.models import db


class StringsDemoModel(db.Document):
    """Documentation example model."""

    required_url_field = db.URLField(required=True, min_length=25)

URLField with https only regexp check, if data exist

Regexp for url_regex should be prepared by re.

"""strings_demo.py"""
import re

from example_app.models import db


class StringsDemoModel(db.Document):
    """Documentation example model."""

    https_url_field = db.URLField(
        url_regex=re.compile(
            r"^(https:\/\/)[\w.-]+(?:\.[\w\.-]+)+[\w\-\._~:/?#[\]@!\$&'\(\)\*\+,;=]+$"
        ),
    )

Unsupported fields

CachedReferenceField

Not yet documented. Please help us with new pull request.

DynamicField

Not yet documented. Please help us with new pull request.

EmbeddedDocumentListField

Not yet documented. Please help us with new pull request.

EnumField

Not yet documented. Please help us with new pull request.

GenericEmbeddedDocumentField

Not yet documented. Please help us with new pull request.

GenericLazyReferenceField

Not yet documented. Please help us with new pull request.

GeoJsonBaseField

Not yet documented. Please help us with new pull request.

GeoPointField

Not yet documented. Please help us with new pull request.

ImageField

Not yet documented. Please help us with new pull request.

LazyReferenceField

Not yet documented. Please help us with new pull request.

LineStringField

Not yet documented. Please help us with new pull request.

LongField

Not yet documented. Please help us with new pull request.

MapField

Not yet documented. Please help us with new pull request.

MultiLineStringField

Not yet documented. Please help us with new pull request.

MultiPointField

Not yet documented. Please help us with new pull request.

MultiPolygonField

Not yet documented. Please help us with new pull request.

PointField

Not yet documented. Please help us with new pull request.

PolygonField

Not yet documented. Please help us with new pull request.

SequenceField

Not yet documented. Please help us with new pull request.

UUIDField

Not yet documented. Please help us with new pull request.

Unsure

GenericReferenceField

Not yet documented. Please help us with new pull request.

ObjectIdField

Not yet documented. Please help us with new pull request.