dgeq 0.4.1

Provide a query string API for construction of complex django QuerySet - https://dgeq.readthedocs.io/

DGeQ

DGeQ (Django Generic Query) is a package that allows the construction of complex GET query on any Model. It implements a query string syntax allowing filtering, join, annotation, aggregation and more.

Even though this package is primarily intended for the creation of public API, allowing to set new endpoints with only few lines of code, it could also be used to communicate with other services or a front-end framework.

The documentation can be found on readthedocs.

Features

• You can choose which field of which Model can be queried.

• Natively support the follow types : int, float, string, boolean, date and None

• Natively support most of Django's lookup function : (i)exact, (i)contains, (i)startswith , (i)endswith, gt(e) and lt(e).

• Supports spanning relationship lookup.

• Support commands such as Joins, Aggregations and Annotations.

• The syntax allows choosing which field the result should contain, avoiding over-fetching.

• Highly customizable : Allow the creation of new type, lookup function and commands. Parts of the query string syntax can also be customized.

You can find more about the syntax, and the possibility it offers here.

Requirements

• python >= 3.6.0
• django >= 2.0.0
• dateutil >= 2.8.0

Test it Online!

You can test DGeQ there if you want : https://dgeq.qcoumes.com/.

Quick Example

Let's show what DGeQ can do using the following models:

# models.py

class Continent(models.Model):
    name = models.CharField(max_length=255, unique=True)
    
class Region(models.Model):
    name = models.CharField(max_length=255, unique=True)
    continent = models.ForeignKey(Continent, on_delete=models.CASCADE, related_name="regions")

class Country(models.Model):
    name = models.CharField(max_length=255, unique=True)
    area = models.BigIntegerField()
    population = models.BigIntegerField()
    region = models.ForeignKey(Region, on_delete=models.CASCADE, related_name="countries")

class River(models.Model):
    name = models.CharField(max_length=255, unique=True)
    discharge = models.IntegerField(null=True)
    length = models.IntegerField()
    countries = models.ManyToManyField(Country, related_name="rivers")

# views.py

@require_GET
def continent(request: HttpRequest):
    q = dgeq.GenericQuery(models.Continent, request.GET)
    return JsonResponse(q.evaluate())

@require_GET
def region(request: HttpRequest):
    q = dgeq.GenericQuery(models.Region, request.GET)
    return JsonResponse(q.evaluate())

@require_GET
def country(request: HttpRequest):
    q = dgeq.GenericQuery(models.Country, request.GET)
    return JsonResponse(q.evaluate())

@require_GET
def river(request: HttpRequest):
    q = dgeq.GenericQuery(models.River, request.GET)
    return JsonResponse(q.evaluate())

In the following examples, the used URL pattern is [model]/?[query]. For instance, if we want to query the continent Africa : continent/?name=Africa.

• Name of all the rivers in China : river/?countries.name=China&c:show=name

{
  "status": true,
    "rows":[
        { "name":"Wu" },
        { "name":"Huai River" },
        { "name":"Yellow River" },
        { "name":"Red (Asia)" },
        { "name":"Ghaghara" },
        { "name":"Salween" },
        { "name":"Indus" },
        { "name":"Amur" },
        { "name":"Ob" },
        { "name":"Irrawaddy River" }
    ]
}

• Name, population and area of country in Asia finishing with stan, sorted by area :
  country/?region.continent.name=Asia&name=$stan&c:hide=id,region,rivers&c:sort=area

{
  "status":true,
  "rows":[
    {
      "area":142600,
      "population":8880300,
      "name":"Tajikistan"
    },
    {
      "area":199949,
      "population":6189700,
      "name":"Kyrgyzstan"
    },
    {
      "area":448969,
      "population":31959800,
      "name":"Uzbekistan"
    },
    {
      "area":488100,
      "population":5757700,
      "name":"Turkmenistan"
    },
    {
      "area":652864,
      "population":36296100,
      "name":"Afghanistan"
    },
    {
      "area":796095,
      "population":207906200,
      "name":"Pakistan"
    },
    {
      "area":2724902,
      "population":18080000,
      "name":"Kazakhstan"
    }
  ]
}

• Join the two lengthiest rivers of France :
  country/?name=France&c:join=field=rivers|limit=2|sort=-length|hide=countries

{
  "status": true,
  "rows": [
    {
      "name": "France",
      "area": 551500,
      "population": 64842500,
      "id": 75,
      "region": 20,
      "rivers": [
        {"name": "Rhine", "discharge": 2330, "length": 1233, "id": 43},
        {"name": "Loire", "discharge": 840, "length": 1012, "id": 22}
      ]
    }
  ]
}

• Sum and Avg of the length of every river in Europe :
  river/?countries.region.continent.name=Europe&c:aggregate=field=length|to=sum_length|func=sum,field=length|to=avg_length|func=avg&c:evaluate=0

{
  "status": true,
  "sum_length": 145943,
  "avg_length": 1871.0641025641025
}

CHANGELOG

0.4.0

• Updated github workflows.

0.4.0

Breaking changes.

• Rewritten most of the project.
• A more complete documentation.

0.3.7

• Default and max c:limit can now be defined in settings.py through DGEQ_DEFAULT_LIMIT and DGEQ_MAX_LIMIT.
• Dotted path to models can now be used for DGEQ_PRIVATE_FIELDS.

0.3.7

• user argument is now mandatory only when use_permission=True in utils.serialize

0.3.6

• user argument is now mandatory only when use_permission=True

0.3.5

• Fix an error occurring when a ForeignField is None
• User's password are private by default

0.3.4

• Fix an exception when no related_name is given on a related field

0.3.3

• Reverse field's name are now correctly fetched

0.3.2

• Joins now correctly account for nullable foreign key

0.3.1

• Minor fix

0.3.0

• Instead of calling commands in a specific order, each field/value pair of the query string call the commands with a regex matching the field. The main consequence of this change is that query argument's order now matter.

• Added c:distinct command as well as distinct argument for for c:join. A new function called dcount (to make distinct count) has been added for c:aggregate and c:annotate.

0.2.1

• Added serialize() method to serialize a model instance in the same format as a GenericQuery's row.

0.2.0

• Now can allow only specific field for each Model, or can use django's permission system

0.1.1 & 0.1.2

• Added missing MANIFEST.in

0.1.0

• Initial release