django-custom-table

Overview

Django Custom Table is a framework for building the back end to a no-code platform in Django or for adding no-code customizible tables to a Django application.

Installation

Install using pip

pip install django-custom-table

or clone the project from github

git clone https://github.com/warrenwestfall/django-custom-table

Add custom_table to your INSTALLED_APPS setting.

INSTALLED_APPS = [
    ...
    'custom_table',
]

Example Usage

First create a metadata model. This model will store all of the customizations users make to customizable tables. The model stores the app label and model name of the custom table model being customized, so you can have only one metadata model for mutlple customizable table models.

Edit or create an app's models.py module
```
from custom_table.models import Metadata
from custom_table.formats import DefaultFormat


class MyMetadata(Metadata):
    created = models.DateTimeField('Created', auto_now_add=True, db_index=True)
    modified = models.DateTimeField('Modified', auto_now=True)
    
    class Meta:
        ct_format_class = DefaultFormat()
```
Your metadata model will need to inherit from custom_table.models.Metadata and provide a format class (as the ct_format_class meta attribute) that tells Django Custom Table how customazaton metadata will be stored and how to format the metadata for use in form and list renderng. How to build your own format class is decribribed later. This example adds some additional audit trail fields.

Next create one or more customzable models.

Create a new customizable model or edit and existing one to make it customizable

from django.db import models
from custom_table.models import CustomizableMeta, CustomizableMixin


STORAGE_FIELDS = {
    "indexed_char": {
        "num_to_create": 10,
        "field_class": models.CharField,
        "field_class_params": {
            "max_length": 128,
            "db_index": True,
        },
    },
    "char": {
        "num_to_create": 30,
        "field_class": models.CharField,
        "field_class_params": {
            "max_length": 128,
        },
    },
    "indexed_integer": {
        "num_to_create": 20,
        "field_class": models.IntegerField,
        "field_class_params": {
            "db_index": True,
        },
    },
    "integer": {
        "num_to_create": 20,
        "field_class": models.IntegerField,
        "field_class_params": {},
    },
    "text": {
        "num_to_create": 30,
        "field_class": models.TextField,
        "field_class_params": {},
    },
    "float": {
        "num_to_create": 10,
        "field_class": models.FloatField,
        "field_class_params": {},
    },
    "boolean": {
        "num_to_create": 10,
        "field_class": models.BooleanField,
        "field_class_params": {},
    },
    "datetime": {
        "num_to_create": 10,
        "field_class": models.DateTimeField,
        "field_class_params": {},
    },
    "decimal-1000-2": {
        "num_to_create": 10,
        "field_class": models.DecimalField,
        "field_class_params": {
            "max_digits": 1000,
            "decimal_places": 2,
        },
    },
}


class MyCustomTable(models.Model, CustomizableMixin, metaclass=CustomizableMeta):
    created = models.DateTimeField('Created', auto_now_add=True, db_index=True)
    modified = models.DateTimeField('Modified', auto_now=True)
    example_static_field = models.CharField('Example Static Field', max_length=32)

    class Meta:
        ct_metadata_model = MyMetadata
        ct_storage_fields = STORAGE_FIELDS
        ct_db_field_prefix = 'ctf_'

Set custom_table.models.CustomizableMeta as the metaclass and provide a ct_storage_fields meta attribute, a dictionary describing how to generate storage fields and how to map your users custom fields to storage fields. Also it is best to provide ct_db_field_prefix with a string to prefix to all storage field names and prevent conflicts with ant static fields in your model.

If you require your own form and/or list metadata formats you can create your own metadata format class

from custom_table.formats import BaseFormat


class RestSpaFormat(BaseFormat):
    """ Format metadata for use in a Single Page App with a REST back end
        and a Javascript front end.
        Form metadate follows react-jsonschema-form
    """

    def get_custom_fields(self, metadata):
    """ Must return a list of dictionaries containing field_name and field_type
        [
            {
                "name": "example_field_name",
                "type": "char"
            },
        ]
        Overide this in order to store the metadata in the custom_data field
        in a different format than what is expected by Metadata.save().
        This is used by the metadata model to calculate storage mappings

        In this example we will assume that our apllication stores the metadata
        in the expected format above.
    """
    return metadata.custom_data


    def get_list_metadata(self, metadata):
        """ Should return the metadata required for a Django or front end view
        to render a grid or list.
        Overide this to produce output required by your appication
        If it is desired that the list view be entirely dynamic that method
        should combine both static Django fields and custom fields.

        In this example we include static Django fields and convert them to
        dictionary/json objects. We also add some simplication of the type
        to combine types that are stored differently but rendered the same. 
    """
        all_fields = self.get_all_fields(metadata)
        # print(all_fields)
        list_metadata = []
        for field in all_fields:
            type = field['type']
            if type.startswith('indexed_'):
                type = type[8:]
            list_field = {
                'name': field['name'],
                'type': type,
            }
            if 'list' in field:
                for list_property, value in field['list'].items():
                    list_field[list_property] = value
            list_metadata.append(list_field)
        return list_metadata


    def get_form_metadata(self, metadata):
        """ Should return the metadata required for a Django or front end view 
            to render a form.
            Overide this to produce output required by your appication
            If it is desired that the form view be entirely dynamic that method
            should combine both static Django fields and custom fields.

            In this example we include static Django fields and
            output from metadata in a format that should work with 
            react-jsonschema-form
        """
        all_fields = self.get_all_fields(metadata)
        form_metadata = {
            'title': metadata.title,
            'type': 'object',
            'properties': {},
        }
        for field in all_fields:
            properties = {
                'type': field['type'],
            }
            if field['type'] in ('indexed_char', 'char', 'text',):
                properties['type'] = 'string'
            if field['typethon'] in ('indexed_integer', 'integer',):
                properties['type'] = 'integer'
            if field['type'] == 'float':
                properties['type'] = 'number'
            if field['type'] == 'datetime':
                properties['type'] = 'string'
                properties['format'] = 'date-time'
            if field['type'].startswith('decimal'):
                # _, max_digits, decimal_places = field['type'].split('_')
                properties['type'] = 'string'
                properties['format'] = field['type']
            if 'form' in field:
                for form_property, value in field['form'].items():
                    properties[snake_to_camel(form_property)] = value
            form_metadata['properties'][field['name']] = properties
        return form_metadata

The above example is taken from the built in django-custom-table formats and can be used as is as follows

from custom_table.formats import RestSpaFormat

When creating views, you can use the Django Custom table base view classes that have helper methods for reading and writing using customizable models.

This how a view for a rest web service might look.

In a views.py

class RestMetadataListView(BaseMetadataView):
    metadata_model=RestSpaFormatMetadata
    include_metadata=True

    
    def get(self, request):
        return JsonResponse(self.get_list(), safe=False)


    def post(self, request):
        new_record = self.create(json.loads(request.body))
        return JsonResponse({'pk': new_record.pk}, status=201)


class RestMetadataDetailView(BaseMetadataView):
    metadata_model=RestSpaFormatMetadata
    include_metadata=True
    always_update_fields = ['modified']


    def get(self, request, name_or_pk):
        return JsonResponse(self.get_detail(name_or_pk), safe=False)

    
    def patch(self, request, name_or_pk):
        self.update_fields(name_or_pk, json.loads(request.body))
        return HttpResponse(status=202)


    def delete(self, request, name_or_pk):
        self.delete_record(name_or_pk)
        return HttpResponse(status=204)

In a urls.py

rest_urlpatterns = [
    path('metadata/', RestMetadataListView.as_view()),
    path('metadata/<str:name_or_pk>/', RestMetadataDetailView.as_view()),
    path('data/<str:name>/', RestCustomTableListView.as_view(),
    path('data/<str:name>/<int:pk>/', RestCustomTableDetailView.as_view(),
]

html_urlpatterns = [
    path('<str:name>/', HtmlCustomTableListView.as_view()),
    path('<str:name>/<int:pk>/', HtmlCustomTableDetailView.as_view()),
]

urlpatterns = [
    path('admin/', admin.site.urls),
    path('rest/', include(rest_urlpatterns)),
    path('html/', include(html_urlpatterns)),
]

metadata_model tells the view what metadata model to use to discover and render all customizations include_metadata is used to tell the view to always include metadata in calls to get_list() and get_detail() always_update_fields tells the view what additional fields tell Django to update in the save() method called in update_fields()