Build Entities

This chapter explains how to build NGSI-LD compliant entities.

Cookbook shows realistic examples - taken from the Smart Data Models Initiative - built with ngsildclient.

Create an entity

Let’s start by creating an empty entity - without any properties yet.

We’ll add properties in a next step.

A NGSI-LD entity is characterized by :

• a unique and fully qualified identifier

• the type the entity belongs to

• a context

Don’t worry about the context for now. We’ll speak about it later.

1. The preferred way

Provide :

• the type the entity belongs to

• a fully qualified identifier

from ngsilclient import Entity

entity = Entity("AirQualityObserved", "urn:ngsi-ld:AirQualityObserved:Madrid-28079004-2016-03-15T11:00:00Z")

Let’s display our newly created entity

entity.pprint()

AirQualityObserved empty NGSI-LD Entity

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:AirQualityObserved:Madrid-28079004-2016-03-15T11:00:00Z",
   "type": "AirQualityObserved"
}

Note that the library has assigned a default context.

To simplify the code we can omit the scheme and namespace identifier.

entity = Entity("AirQualityObserved", "AirQualityObserved:Madrid-28079004-2016-03-15T11:00:00Z")

By convention the identifier string is prefixed with the type.

We can simplify once again to obtain the following line.

 entity = Entity("AirQualityObserved", "Madrid-28079004-2016-03-15T11:00:00Z")

1. Alternate way

One could provide the qualified identifier and let the library infer the type.

entity = Entity("urn:ngsi-ld:AirQualityObserved:Madrid-28079004-2016-03-15T11:00:00Z")

Which can be written.

entity = Entity("AirQualityObserved:Madrid-28079004-2016-03-15T11:00:00Z")

Specify a context

One can provide the Entity constructor with a context.

If no context is provided it defaults to the NGSI-LD Core Context.

The context variable is an array of strings.

1. Using the context property

from ngsilclient import Entity

entity = Entity("AirQualityObserved", "Madrid-28079004-2016-03-15T11:00:00Z")
entity.context.append("https://github.com/smart-data-models/dataModel.Environment/blob/master/context.jsonld")

2. Using the constructor

from ngsilclient import Entity, CORE_CONTEXT

ctx = [ "https://github.com/smart-data-models/dataModel.Environment/blob/master/context.jsonld",
         CORE_CONTEXT ]
entity = Entity("AirQualityObserved", "Madrid-28079004-2016-03-15T11:00:00Z", ctx=ctx)

Add properties

Here we talk about properties in the broad sense, including relationships.

Primitives

The Entity class provides primitives, whose purpose is to generate JSON content representing a property.

The latter property is attached to the Entity instance.

Four primitives are available : prop(), gprop(), tprop() and rel() that allow respectively to set a Property, GeoProperty, Temporal Property and RelationShip.

For convenience these methods can be chained in order to shorten the code.

Properties have at least a name and value and can optionally carry metadata and userdata.

property	primitive
Property	prop()
GeoProperty	gprop()
Temporal Property	tprop()
Relationship	rel()

Example

 from ngsilclient import Entity

 entity = Entity("AirQualityObserved", "Madrid-28079004-2016-03-15T11:00:00Z")
 entity.prop("CO", 500).prop("NO", 45).prop("NO2", 69)

AirQualityObserved NGSI-LD Entity with some pollutant concentrations

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:AirQualityObserved:Madrid-28079004-2016-03-15T11:00:00Z",
   "type": "AirQualityObserved",
   "CO": {
      "type": "Property",
      "value": 500
   },
   "NO": {
      "type": "Property",
      "value": 45
   },
   "NO2": {
      "type": "Property",
      "value": 69
   }
}

Metadata

Metadata aka “properties of properties” are : unitCode, datasetId and observedAt.

Primitives accept metadata as arguments.

Which ones are available depend on which property you’re building.

For example unitCode is available for a Property but not for a GeoProperty.

Corresponding arguments in lower case.

metadata	argument
unitCode	unitcode
datasetId	datasetid
observedAt	observedat

unitCode

SO2 concentration with its measurement unit code

 entity.prop("SO2", 11, unitcode="GP") # milligram per cubic metre (UNECE/CEFACT)

datasetId

SO2 concentration with a datasetId

 entity.prop("SO2", 11, datasetid="dataset:01") # urn prefix omitted

observedAt

SO2 concentration with the observation date

 from datetime import datetime, timezone

 entity.prop("SO2", 11, observedat=datetime(2016, 3, 15, 11, tzinfo=timezone.utc))

 # Alternatively one could pass directly an ISO8601 string
 # entity.prop("SO2", 11, observedat="2016-03-15T11:00:00Z")

The library will always convert datetimes to UTC as expected by NGSI-LD.

It’s recommended to work with aware datetimes 1.

SO2 concentration with a CET observation date

 from datetime import datetime
 from zoneinfo import ZoneInfo

 CET = ZoneInfo("CET") # UTC+1
 entity.prop("SO2", 11, observedat=datetime(2016, 3, 15, 12, tzinfo=CET))

 # Alternatively one could pass directly an ISO8601 string
 # entity.prop("SO2", 11, observedat="2016-03-15T11:00:00Z")

Often an observation date is the same that is used in the entity identifier to make it unique.

And often the whole set of measures share the same observation date.

The library caches the first datetime it encounters and allows to reuse it.

For the sake of example let’s rewrite our entity.

Example

 from ngsildclient import Entity, Auto

 entity = Entity("AirQualityObserved", "Madrid-28079004-2016-03-15T11:00:00Z")
 entity.prop("CO", 500, unitcode="GP", observedat=Auto)
 entity.prop("NO", 45, unitcode="GP", observedat=Auto)
 entity.prop("NO2", 69, unitcode="GP", observedat=Auto)
 entity.prop("SO2", 11, unitcode="GP", observedat=Auto)

AirQualityObserved NGSI-LD Entity with all measures

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:AirQualityObserved:Madrid-28079004-2016-03-15T11:00:00Z",
   "type": "AirQualityObserved",
   "CO": {
      "type": "Property",
      "value": 500,
      "unitCode": "GP",
      "observedAt": "2016-03-15T11:00:00Z"
   },
   "NO": {
      "type": "Property",
      "value": 45,
      "unitCode": "GP",
      "observedAt": "2016-03-15T11:00:00Z"
   },
   "NO2": {
      "type": "Property",
      "value": 69,
      "unitCode": "GP",
      "observedAt": "2016-03-15T11:00:00Z"
   },
   "SO2": {
      "type": "Property",
      "value": 11,
      "unitCode": "GP",
      "observedAt": "2016-03-15T11:00:00Z"
   }
}

User data

User data are additional custom metadata the user wants to be included in the property.

You can use the userdata argument to provide your own dictionary.

For example imagine you’d like to add to our AirQualityObserved entity a NOx measure with an accuracy indice.

But you don’t want to express this information as a property.

More on nested properties later.

Example

 entity.prop("NOx", 119, userdata={"accuracy": 0.95})

The NOx property with userdata

"NOx": {
   "type": "Property",
   "value": 119,
   "accuracy": 0.95
}

Property

We have already spoken of the Property in previous examples.

That is a general Property that has a name and a value of any Python type that can be serialized to JSON.

A temperature property with a float value

 entity.prop("temperature", 22.5)

A description property with a List value

 entity.prop("description", [
    "https://example.org/concept/clay",
    "https://datamodel.org/example/clay"]
  }

A description property with a string value

 entity.prop("description", "Corn farm")

Considering strings, some characters are prohibited by the NGSI-LD broker.

In this case you can set the escape argument to to escape the string.

A description property with a string value including forbidden characters

 entity.prop("description", "Corn farm (organic)", escape=True)

GeoProperty

The GeoProperty as it name implies represents a geographic property.

ngsildclient relies on the geojson library.

As of now allowed geojson types are Point, LineString, Polygon and MultiPoint.

A simple GeoProperty

 from geojson import Point
 from ngsildclient import Entity

 entity = Entity("AirQualityObserved", "Madrid-28079004-2016-03-15T11:00:00Z")
 entity.gprop("location", Point((-3.703790, 40.416775))) # Madrid

A location geoproperty illustrated

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:AirQualityObserved:Madrid-28079004-2016-03-15T11:00:00Z",
   "type": "AirQualityObserved",
   "location": {
      "type": "Property",
      "value": {
         "type": "Point",
         "coordinates": [
         -3.70379,
         40.416775
         ]
      }
   }
}

The Point is by far the most common geo type.

ngsildclient accepts a tuple (lat, lon) to represent a Point, avoiding the need to invoke geojson for this simple case.

entity.gprop("location", (40.416775, -3.703790)) # Madrid

The loc() alias can be used to set this very common location GeoProperty.

entity.loc((40.416775, -3.703790)) # Madrid

Temporal Property

The Temporal Property accepts values of following types : datetime, date and time, or ISO8601 string representations of these latter.

Temporal Property illustrated

 from ngsildclient import Entity

 entity = Entity("AirQualityObserved", "Madrid-28079004-2016-03-15T11:00:00Z")
 entity.tprop("dateObserved", "2016-03-15T11:00:00Z")

In the above example the library has cached at the entity creation time the datetime that is part of the identifier.

The Auto directive can be used to benefit from the cached datetime.

Temporal Property using the Auto keyword

 from ngsildclient import Entity, Auto

 entity = Entity("AirQualityObserved", "Madrid-28079004-2016-03-15T11:00:00Z")
 entity.tprop("dateObserved", Auto)  # We could omit Auto as it's the default

Note

Auto means the cached datetime if any else defaults to the current datetime utcnow().

Auto is the default value for the tprop() primitive.

The obs() alias can be used to set this very common dateObserved Temporal Property.

The library will always convert datetimes to UTC as expected by NGSI-LD.

It’s recommended to work with aware datetimes 1.

entity.obs() # use a cached datetime if any, else current datetime

Relationship

The Relationship Property points to one or many JSON-LD objects.

from ngsildclient import Entity

entity = Entity("Vehicle", "A4567")
entity.rel("isParked", "OffStreetParking:Downtown1", observedat="2017-07-29T12:00:04Z")

Relationship illustrated

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:Vehicle:A4567",
   "type": "Vehicle",
   "isParked": {
      "type": "Relationship",
      "object": "urn:ngsi-ld:OffStreetParking:Downtown1",
      "observedAt": "2017-07-29T12:00:04Z"
   }
}

Import the Rel Enum to access well-known relationship names, such as observedBy or hasPart.

Implement nested properties

Sometimes properties are composed of properties.

Single level

You might want to add a nested property to provide information about the quality check status.

Use the NESTED keyword or set the nested argument to True.

The property will be nested into the latest added property.

from ngsildclient import Entity, NESTED

entity = Entity("AirQualityObserved", "Madrid-28079004-2016-03-15T11:00:00Z")
entity.prop("NO2", 22, unitcode="GP").prop("qc", "checked", NESTED)

Nested property illustrated

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:AirQualityObserved:Madrid-28079004-2016-03-15T11:00:00Z",
   "type": "AirQualityObserved",
   "NO2": {
      "type": "Property",
      "value": 22,
      "unitCode": "GP",
      "qc": {
         "type": "Property",
         "value": "checked"
      }
   }
}

Another nested example from the Guidelines for Modelling with NGSI-LD [ETSI_WP42].

from ngsildclient import Entity, NESTED, Rel

room = Entity("Room", "01")
room.prop("temperature", 17).rel(Rel.OBSERVED_BY, "Sensor:01", NESTED)

Example from the ETSI White Paper

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:Room:01",
   "type": "Room",
   "temperature": {
      "type": "Property",
      "value": 17,
      "observedBy": {
         "type": "Relationship",
         "object": "urn:ngsi-ld:Sensor:01"
      }
   }
}

Multilevel

You can chain nested properties in order to obtain several nesting levels.

from ngsildclient import Entity, NESTED

entity = Entity("AirQualityObserved", "Madrid-28079004-2016-03-15T11:00:00Z")
entity.prop("NO2", 22, unitcode="GP").prop("qc", "checked", NESTED).prop("status", "discarded", NESTED)

Multilevel nested property illustrated

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:AirQualityObserved:Madrid-28079004-2016-03-15T11:00:00Z",
   "type": "AirQualityObserved",
   "NO2": {
      "type": "Property",
      "value": 22,
      "unitCode": "GP",
      "qc": {
         "type": "Property",
         "value": "checked",
         "status": {
            "type": "Property",
            "value": "discarded"
         }
      }
   }
}

Anchoring

By default a property is added to the entity’s root.

When NESTED is set the property is nested into the latest added property.

Sometimes you need to nest properties into a given and fixed property.

Here comes the anchor() method that allows to set a property into which subsequent properties will be nested.

Until the unanchor() method is called to return to the default behaviour.

from datetime import datetime
from ngsildclient import Entity

parking = Entity("OffStreetParking", "Downtown1")
parking.prop("availableSpotNumber", 121, observedat=datetime(2017, 7, 29, 12, 5, 2).anchor()
parking.prop("reliability", 0.7).rel("providedBy", "Camera:C1").unanchor()
parking.prop("description", "Municipal car park located near the Trindade metro station and the Town Hall")

Multilevel nested property illustrated

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:OffStreetParking:Downtown1",
   "type": "OffStreetParking",
   "availableSpotNumber": {
      "type": "Property",
      "value": 121,
      "observedAt": "2017-07-29T12:05:02Z",
      "reliability": {
         "type": "Property",
         "value": 0.7
      },
      "providedBy": {
         "type": "Relationship",
         "object": "urn:ngsi-ld:Camera:C1"
      }
   },
   "description": {
      "type": "Property",
      "value": "Municipal car park located near the Trindade metro station and the Town Hall"
   }
}

Update an entity

An Entity instance is backed by a dictionary.

In fact a NGSI-dedicated dictionary that extends the native Python dictionary.

You use this dictionary each time you deal with a subpart of the Entity.

It provides obviously all the native dictionary staff but also :

• the prop(), gprop(), tprop() and rel() primitives quite equivalent to those provided by the Entity

• a dot notation to access fields, i.e. room[“temperature.value”]

Let’s consider the following example.

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:Room:01",
   "type": "Room",
   "temperature": {
      "type": "Property",
      "value": 22.5,
      "observedBy": {
         "type": "Relationship",
         "object": "urn:ngsi-ld:Sensor:01"
      }
   },
   "pressure": {
      "type": "Property",
      "value": 938.8
   }
}

Update a member

Entity members are id, type and context.

The Entity class provides a Python property for each one.

Members can be updated but cannot be removed.

from ngsildclient import Entity

room.id = "urn:ngsi-ld:Room:02"

Update a value

from ngsildclient import Entity

room["temperature.value"] += 0.2

Add metadata or userdata

Use the same method.

from ngsildclient import Entity

room["temperature.unitCode"] = "CEL"

Remove any part of the Entity

It applies to properties as well.

from ngsildclient import Entity

del room["temperature.unitCode"]

Update a property

To update an Entity’s property the easiest way is to override it.

from ngsildclient import Entity

room.prop("pressure", 938.7, unitcode="A97")

Add a nested property

We can add a nested property without rebuilding the upper property.

Here we nest a qc property into the temperature property.

from ngsildclient import Entity

room["temperature"].prop("qc", "checked")

Add a multilevel nested property

Here we nest a status property into a qc property, itself nested into the temperature property.

Note that chaining the prop() automatically enables nesting.

The prop() method used here does not belong to the Entity but to the NGSI-dedicated dictionary.

from ngsildclient import Entity

room["temperature"].prop("qc", "checked").prop("status", "discarded")

Display an entity

The to_json() method returns the JSON payload as a string.

By setting the kv argument to True, it returns the simplified representation aka KeyValues format.

The pprint() method relies on to_json() in order to pretty-print the entity.

It also takes a kv argument.

Import/Export

Dictionary

Suppose we have this dictionary.

payload = {
   "type": "Room",
   "id": "urn:ngsi-ld:Room:01",
   "@context": "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld",
   "temperature": {"type": "Property", "value": 22.5}
}

You can create an entity from this dictionary.

Note that id, type and @context are mandatory.

If missing an exception will be raised.

from ngsildclient import Entity

room = Entity.from_dict(payload)

The opposite operation converts your entity to a dictionary.

from ngsildclient import Entity

payload = room.to_dict()

File

Load/Save a single entity

Import and Export from/to a file is a very useful feature that allows :

• backup : just restore an entity you’ve previously saved

• testing : store an expected result for further comparison

• sharing : elaborate with others about modeling

• experimenting : load an example from the Smart Data Models Initiative and play around

• contributing : propose a NGSI-LD example to the Smart Data Models Initiative

We can load a local file.

from ngsildclient import Entity, SmartDataModels

room = Entity.load("/tmp/room1.jsonld")

And save an entity to a file.

from ngsildclient import Entity, SmartDataModels

room.save("/tmp/room2.jsonld")

We can load a remote file through HTTP.

from ngsildclient import Entity

battery = Entity.load("https://github.com/smart-data-models/dataModel.Battery/raw/master/Battery/examples/example-normalized.jsonld")

Load sample entities

For convenience some datamodel example URLs of the Smart Data Models Initiative are made available.

from ngsildclient import Entity, SmartDataModels

beach = Entity.load(SmartDataModels.SmartCities.PointOfInterest.Beach)

Batch Import

If the input file contains a JSON array made of several entities, the result will be a list of entities.

Really useful to create batch of entities in the broker.

from ngsildclient import Entity

rooms: list[Entity] = Entity.load("/tmp/rooms_all.jsonld")

Note

One could use explicitly the load_batch() method that expects a JSON array.

Batch Export

from ngsildclient import Entity

rooms = [Entity("Room", "Room1"), Entity("Room", "Room2")]
Entity.save_batch(rooms, "/tmp/rooms_all.jsonld")

Utils

ISO8601

In NGSI-LD entities dates, times and datetimes are represented as ISO8601 strings.

The iso8601 module provides you with functions to convert from Python types to ISO8601 :

• from_date()

• from_time()

• from_datetime()

• utcnow() to get the current datetime

Note that this is not needed for the tprop() primitive and observedat argument that accepts Python date types, calling these functions for you.

ISO8601 Example

from datetime import datetime
from ngsildclient import iso8601, TZ_CET

dt = datetime(2022, 3, 10, 17, 49, tzinfo=TZ_CET)
iso8601.from_datetime(dt) # '2022-03-10T16:49:00Z'

Short UUID

A UUID may be useful in some cases to create a unique Entity identifier.

But a long dash-separated string is not always suitable. The short UUID string will be 22 characters long, base64-encoded, with padding characters removed. The encoding uses the urlsafe alphabet with a slightly difference. The dash character (often used as a NGSI field separator) is replaced by the tilde character.

from ngsildclient import Entity, shortuuid

dt = datetime(2022, 3, 10, 17, 49, tzinfo=TZ_CET)
crop = Entity("AgriCrop", shortuuid())

short UUID illustrated

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:AgriCrop:ldoRJQMZSaaKoWn9g_JR~g",
   "type": "AgriCrop"
}

Helpers

Helper functions can help building complex data structures frequently used in NGSI-LD entities.

Using helper functions :

• enforces a well-constructed consistent structure

• guides you through the different options thanks to IDE autocompletion

The code may look quite long at first sight but is mainly generated by the IDE.

PostalAddress

PostalAddress as defined by Schema.org.

from ngsildclient import Entity, PostalAddressBuilder

entity = Entity("WeatherObserved", "Spain-Valladolid-2016-11-30T07:00:00.00Z")
entity.prop("address",
   PostalAddressBuilder()
   .street("C/ La Pereda 14")
   .locality("Santander")
   .region("Cantabria")
   .country("Spain")
   .build())

The addr() alias can be used to set this very common address Property.

PostalAddress illustrated

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:WeatherObserved:Spain-Valladolid-2016-11-30T07:00:00.00Z",
   "type": "WeatherObserved",
   "address": {
      "type": "Property",
      "value": {
         "streetAddress": "C/ La Pereda 14",
         "addressLocality": "Santander",
         "addressRegion": "Cantabria",
         "addressCountry": "Spain"
      }
   }
}

OpeningHours

OpeningHoursSpecification as defined by Schema.org.

from ngsildclient import Entity, OpeningHours

openinghours = OpeningHoursBuilder()
   .businessdays("10:00", "17:30")
   .saturday("10:00", "14:00")
   .build()
library = Entity("Library", "Ireland-Shannon-PublicLibrary")
library.prop("openingHours", openinghours)

OpeningHours illustrated

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:Library:Ireland-Shannon-PublicLibrary",
   "type": "Library",
   "openingHours": {
      "type": "Property",
      "value": [
         {
            "opens": "10:00",
            "closes": "17:30",
            "dayOfWeek": "Monday"
         },
         {
            "opens": "10:00",
            "closes": "17:30",
            "dayOfWeek": "Tuesday"
         },
         {
            "opens": "10:00",
            "closes": "17:30",
            "dayOfWeek": "Wednesday"
         },
         {
            "opens": "10:00",
            "closes": "17:30",
            "dayOfWeek": "Thursday"
         },
         {
            "opens": "10:00",
            "closes": "17:30",
            "dayOfWeek": "Friday"
         },
         {
            "opens": "10:00",
            "closes": "14:00",
            "dayOfWeek": "Saturday"
         }
      ]
   }
   }

This is only a basic implementation but still useful. As of now it does not support break times.

Mocking

For testing purpose you may need a lot of entities but don’t have them.

Here comes MockerNgsi. In the following example it generates from a given entity 100 new mocked entities.

from ngsildclient import Entity, MockerNgsi

entity = Entity("AirQualityObserved", "Madrid-28079004-2016-03-15T11:00:00Z")
entity.prop("NO2", 22, unitcode="GP")

# generate 100 mocked entities
mocker = MockerNgsi()
mocked_entities = mocker.mock(entity, 100)

Have a look at the first mocked entity in the list.

Mocking illustrated

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:AirQualityObserved:Madrid-28079004-2016-03-15T11:00:00Z:Mocked:a_m7D8qkQBCIjlTnkY7J~g",
   "type": "AirQualityObserved",
   "NO2": {
      "type": "Property",
      "value": 22,
      "unitCode": "GP"
   },
   "mocked": {
      "type": "Property",
      "value": true
   }
}

The id has been suffixed with a unique mock-identifier. A mocked property has been added. This is the default behaviour of the MockerNgsi class.

It’s up to you to implement your custom mocking logic by providing your own f_mock_id and f_mock_payload functions.

import random

def randomize_NO2(entity: Entity):
   entity.prop("mocked", True)
   entity["NO2.value"] += random.uniform(-3.0, 3.0)

# generate 100 mocked entities
mocker = MockerNgsi(f_mock_payload=randomize_NO2)
mocked_entities = mocker.mock(entity, 100)

Custom mocking logic illustrated

{
   "@context": [
      "https://uri.etsi.org/ngsi-ld/v1/ngsi-ld-core-context.jsonld"
   ],
   "id": "urn:ngsi-ld:AirQualityObserved:Madrid-28079004-2016-03-15T11:00:00Z:Mocked:uIzmaZfVT3ulIcR8UY8cXg",
   "type": "AirQualityObserved",
   "NO2": {
      "type": "Property",
      "value": 22.830140401969413,
      "unitCode": "GP"
   },
   "mocked": {
      "type": "Property",
      "value": true
   }
}

For further information look at the MockerNgsi class documentation.

1(1,2)

Note that all NGSI-LD datetimes are UTC. The library will always convert datetimes to UTC, either naive or aware. The downside of not specifying the timezone is that the result depends on your local environment therefore code execution is not reproducible. Behaviour may change if your code is run in a different place/timezone.

ETSI_WP42

Guidelines for Modelling with NGSI-LD ETSI WhitePaper