OSDU Data Definitions Join Smart Data Models: 6 Candidate Models for Wells, Wellbores and Subsurface Data

Smart Data Models’ Candidates repository now includes a new candidate standard from the energy industry: osdu-data-definitions, with 6 candidate models translated from the OSDU Data Definitions of The Open Group OSDU Forum. They cover organisations, fields, wells, wellbores, wellbore trajectories and well logs. The OSDU schemas are published under Apache-2.0. Like every candidate, each model comes with a JSON Schema, an NGSI-LD @context and all four example serializations (NGSI-v2 and NGSI-LD, key-values and normalized), each one validated against its schema.

About the OSDU Data Definitions

OSDU (Open Subsurface Data Universe) is a forum of The Open Group. It builds an open, vendor-neutral data platform for the energy industry. Its Data Definitions are the JSON schemas of the platform’s records, called kinds. There are master data (wells, wellbores, fields, organisations…), work-product components (well logs, trajectories, markers, seismic, reservoir models…), reference data and datasets. Every record has the same envelope (id, kind, access control, legal tags, lineage, frame of reference) around a data block with the domain attributes. Many of these attributes come from PPDM, Energistics and other industry standards. This translation uses release v0.30.1, the latest one.

The models

Model OSDU kind What it describes Attributes
Organisation master-data–Organisation:1.2.0 Operators, service companies, regulators 33
Field master-data–Field:1.1.0 A hydrocarbon deposit that has been exploited commercially 29
Well master-data–Well:1.5.0 The surface origin of a set of wellbores, with operators, states and events 54
Wellbore master-data–Wellbore:1.5.1 A hole from the surface to its deepest point, with depths, trajectory type, products and bottom-hole location 66
WellboreTrajectory work-product-component–WellboreTrajectory:1.4.0 A directional survey of a wellbore 85
WellLog work-product-component–WellLog:1.5.0 Measurements along a wellbore against depth or time, with their curves 58

Real open data in the examples

The examples all describe one real well: AKM-11 (Oudega-Akkrum-11), an onshore gas exploration well in Friesland, the Netherlands, spudded in 1978. The records come from the TNO dataset that OSDU prepared for its open test data (Apache-2.0; the data originally come from the Dutch NLOG portal, which allows redistribution). The set covers the well, its wellbore with coordinates in the original ED50 / UTM 31N system and in WGS 84, its trajectory, its 1978 composite well log (gamma ray, sonic, density and neutron curves), the field and the current operator. The test data were written for an older schema version, so the translator migrated each record to the current one. Each example lists these changes in its source attribute.

Translation notes

  • The data block becomes the entity. Every attribute of an OSDU record’s data block, including those inherited from OSDU’s shared abstract schemas, becomes an entity attribute with its OSDU name unchanged (FacilityName, VerticalMeasurements, SpatialLocation, Curves…).
  • The envelope is kept where it carries meaning. The OSDU id becomes osduId. The NGSI-LD id urn:ngsi-ld:Well:osdu:1013 is built from it and can be turned back into it. kind, meta (the units and coordinate reference systems of the values), ancestry, tags and legal stay. Platform access control and system audit fields are left out.
  • Links between models are Relationships. Operators and service companies point to Organisation, a wellbore to its Well, logs and trajectories to their Wellbore. Links to OSDU reference data (well status, trajectory type, units, CRS…) keep the OSDU record ids, because those code lists are extensible and their namespace depends on the deployment.
  • Nested where OSDU nests. Spatial locations, vertical measurements, facility events and log curves keep their full OSDU structure. The shared abstract schemas are converted once per model into $defs.
  • Ready for geo-queries. Each model gets an NGSI-LD location, the GeoJSON geometry from OSDU’s own WGS 84 coordinates. It is the only addition to the OSDU content.
  • Current attributes only. Attributes that OSDU has deprecated are left out. The @context of each model maps every term to its definition in the published OSDU schema files.

Try it, and tell us what’s missing

These are Candidates: carefully sourced first translations, not yet official Smart Data Models subjects. Points to review: the mapping of the OSDU record envelope, the choice of Relationships versus OSDU ids, and the overlap of Organisation with the existing dataModel.Organization. Natural next steps are Basin, wellbore markers, rigs and hole sections, and the OSDU production and seismic kinds.

Browse the models, open an issue, or send a PR:
👉 https://github.com/smart-data-models/Candidates/tree/master/standards/osdu-data-definitions

InChI Joins Smart Data Models: 1 Candidate Model for Chemical Structure Identifiers

Smart Data Models’ Candidates repository now includes a new candidate standard from chemistry: inchi, with 1 candidate model translated from InChI, the IUPAC International Chemical Identifier. The InChI software and documentation are published under the MIT licence. Like every candidate, the model comes with a JSON Schema, an NGSI-LD @context and all four example serializations (NGSI-v2 and NGSI-LD, key-values and normalized), each validated against its schema.

About InChI

InChI is the open, structure-based identifier for chemical substances, developed by IUPAC and maintained by the InChI Trust. It is computed from a chemical structure and written as a version prefix followed by layers (formula, connections, hydrogen atoms, charge, stereochemistry, isotopes…). Its hashed form, the 27-character InChIKey, is used to index and search chemical databases such as PubChem, ChEBI, ChEMBL and Wikidata. This translation is based on the Technical Manual and User Guide shipped with InChI Software v1.07.5.

The model

InChI is an identifier, not a data model: it defines no entities with state. So this candidate is small on purpose. It gives NGSI-LD a way to state, without ambiguity, which chemical an observation, analysis or product refers to.

Model What it records Attributes Example
ChemicalStructureIdentifier The InChI of one chemical structure: InChI string, InChIKey, AuxInfo, the layers parsed out of the string, the InChIKey blocks and the options used 22 L-alanine (PubChem CID 5950)

Real data

The example is L-alanine, with the InChI and InChIKey exactly as published by PubChem (US government public domain). The AuxInfo line was computed with the InChI library from PubChem’s structure and reproduces PubChem’s identifiers.

Translation notes

  • Source names. InChI, InChIKey and AuxInfo are kept verbatim; the layers use the labels of the Technical Manual (formula, connections, H_atoms, stereo, isotopic, fixed_H…), nested where the manual nests them.
  • Layers for querying. The layers are parsed from the InChI string, so you can filter by formula, by molecular skeleton (the first InChIKey block) or by stereo. The InChI string remains the authoritative value.
  • Linked identifiers. InChI and InChIKey map to the CHEMINF ontology terms; the other attributes point to the relevant page of the InChI Technical Manual.
  • Additions. The entity name, identifierKind (standard, non-standard or beta), the names of the InChIKey parts and softwareVersion are Smart Data Models additions, and the metadata documents them.

Try it, and tell us what’s missing

This is a Candidate: a first, carefully sourced translation that is not yet an official Smart Data Models subject. No official subject defines a chemical identifier today. The natural next step is to link to it from models that talk about chemicals, such as air quality observations or industrial process chemical analyses. InChI extensions for reactions (RInChI), mixtures (MInChI) or nanomaterials could follow if there is demand.

Browse the model, open an issue, or send a PR:
👉 https://github.com/smart-data-models/Candidates/tree/master/standards/inchi

FIBO Joins Smart Data Models: 5 Candidate Models for Legal Entities, LEIs, Ownership, Shares and Loans

Smart Data Models’ Candidates repository now includes a new candidate standard translated from the Financial Industry Business Ontology (FIBO) of the EDM Council: fibo, with 5 candidate models for legal entities, their Legal Entity Identifiers (LEI), parent relationships, listed shares and loans. FIBO is published under the MIT licence. Like every candidate, each model comes with a JSON Schema, an NGSI-LD @context and all four example serializations (NGSI-v2 and NGSI-LD, key-values and normalized), each validated against its schema.

About FIBO

FIBO is the EDM Council’s formal OWL ontology of the financial industry: business entities and identifiers (including the ISO 17442 LEI), ownership and control, contracts, financial instruments and securities, loans, derivatives, indices and corporate actions. It is released quarterly; this translation uses the Production release 2026Q3, which builds on the OMG Commons Ontology Library.

The models

Model FIBO class Attributes Example
LegalEntity legal entity (Commons, with FIBO’s LEI axioms) 24 Real (GLEIF)
LegalEntityIdentifierRegistryEntry LEI registry entry: status, dates, managing LOU, validation 9 Real (GLEIF)
RelationshipRecord LEI relationship record (direct / ultimate parent, branch) 8 Real (GLEIF Level 2)
ListedShare listed share 23 Synthetic
Loan loan 24 Synthetic

Grounded in open LEI data

Three examples use one real record from the GLEIF Global LEI Index (CC0): Siemens Healthineers AG (LEI 529900QBVWXMWANH7H45), its legal and headquarters addresses, legal form (ISO 20275 code 6QQB, Aktiengesellschaft), its LEI registration (issued, fully corroborated), and its Level 2 relationship to its direct parent Siemens AG, with the accounting and filing periods. The share and loan examples are synthetic and use fictitious companies.

Translation notes

  • Derived from OWL restrictions. FIBO defines classes through restrictions. Each model’s attributes come from the restrictions on the class and its superclasses, and every description quotes the restriction it comes from.
  • FIBO names, verbatim. Classes and properties keep their FIBO names (hasLegalAddress, isDirectlyConsolidatedBy, hasPrincipalAmount…), and each @context maps them to the FIBO or Commons IRI.
  • Parties as relationships. Lenders, borrowers, issuers and owners are links to LegalEntity entities. Identifiers, addresses, legal forms, periods, amounts and payment terms stay nested.
  • FIBO code lists. Registration status, entity status, validation level and similar values are FIBO individuals (compact IRIs), with the GLEIF codes listed alongside.
  • Few additions. The only non-FIBO terms are a relationshipType on relationship records and the class of nested identifiers and classifiers.

Try it, and tell us what’s missing

These are Candidates: first, carefully sourced translations that are not yet official Smart Data Models subjects. Before that step, the main open questions are the overlap with the LegalEntity model of dataModel.VerifiableCredentials and with the Beneficial Ownership (BODS) candidate, and how strictly FIBO’s open-world axioms should become required fields. Natural next steps are accounts, listings and exchanges, bonds, collateral, and the ACTUS contract types.

Browse the models, open an issue, or send a PR:
👉 https://github.com/smart-data-models/Candidates/tree/master/standards/fibo

CAMARA APIs Join Smart Data Models: 7 Candidate Models for Telco Network Data in Smart Cities

Smart Data Models’ Candidates repository now includes a new candidate standard from the telecom world: camara-apis, with 7 candidate models translated from the CAMARA network APIs: quality on demand, device location and geofencing, crowd data and device reachability. The CAMARA specifications are published under Apache-2.0. Like every candidate, each model comes with a JSON Schema, an NGSI-LD @context and all four example serializations (NGSI-v2 and NGSI-LD, key-values and normalized), each validated against its schema.

About CAMARA

CAMARA is the open-source project of the Linux Foundation, run with the GSMA Open Gateway initiative, that defines network APIs exposed in the same way by mobile operators around the world. Each API is an OpenAPI specification in its own GitHub repository, released in coordinated meta-releases. This candidate uses the latest public release of each API on 13 November 2026: the Fall25 releases of QualityOnDemand, DeviceLocation, RegionDeviceCount and DeviceReachabilityStatus, and the Sync26 release (1.0.0) of PopulationDensityData.

The models

Model CAMARA API What it describes Attributes
QosProfile qos-profiles 1.1.0 A network quality profile: target and maximum rates, latency, jitter, loss, priority 24
SessionInfo quality-on-demand 1.1.0 A quality-on-demand session between a device and an application server 18
DeviceLocation location-retrieval 0.5.0 The network-based location area of a device 10
GeofencingSubscription geofencing-subscriptions 0.5.0 A subscription to area-entered / area-left events for a device 15
CellPopulationDensityData population-density-data 1.0.0 Estimated number of people in a geohash cell for an hourly slot 14
RegionDeviceCount region-device-count 0.2.0 Number of devices in an area and time window, with filters 13
DeviceReachabilityStatus device-reachability-status 1.1.0 Whether a device is reachable, for data and/or SMS 10

One consistent example scenario

CAMARA has no open data (operators expose the APIs only to registered consumers), so the examples are synthetic, and say so. They follow an evening event at Puerta del Sol in Madrid: a QoD session with a QOS_L profile carries a live-video uplink, a field team’s device is located, watched by a geofence and checked for reachability, and the crowd is described by the population density of geohash cell ezjmgtw and a device count for the square. No real phone numbers are used: device identifiers are CAMARA’s own example number +123456789 and documentation IP addresses.

Translation notes

  • CAMARA names, verbatim. Every field keeps its CAMARA name. Entities are named after the CAMARA component, or after the API when the component is generic or a response wrapper (DeviceLocation, GeofencingSubscription, RegionDeviceCount, DeviceReachabilityStatus). Each @context maps terms to their place in the released OpenAPI file.
  • Linked. A QoD session points to its QosProfile as an NGSI-LD Relationship.
  • Ready for geo-queries. Location, geofence, region and geohash cell entities get an NGSI-LD location built from CAMARA’s own areas. It is the only addition to the CAMARA content.
  • No secrets. Delivery credentials (access tokens) of sessions are left out. Device identifiers are personal data, so deployments should use pseudonymous ids where possible.

Try it, and tell us what’s missing

These are Candidates: first, carefully sourced translations that are not yet official Smart Data Models subjects. Before that step, the main open questions are overlap with CrowdFlowObserved (Transportation), Device and NetworkConnection. Natural next steps are the Connected Network Type and Device Roaming Status APIs, QoS provisioning, Edge Cloud and Network Slice Booking.

Browse the models, open an issue, or send a PR:
👉 https://github.com/smart-data-models/Candidates/tree/master/standards/camara-apis

TM Forum Open APIs Join Smart Data Models: 174 Candidate Models for the Whole Telecom Digital-Service Stack

Smart Data Models’ Candidates repository already had a first slice of the TM Forum Open APIs (five inventory, order and trouble-ticket models). The candidate tmforum-open-apis now covers every TM Forum Open API published under Apache-2.0: 86 APIs and 174 candidate models, one per managed resource, linked by 213 NGSI-LD Relationships. As with every candidate, each model has a JSON Schema, an NGSI-LD @context and four example serializations (NGSI-v2 and NGSI-LD, key-values and normalized), each validated against its schema.

About the TM Forum Open APIs

TM Forum is the global industry association of communication service providers and their suppliers. Its Open APIs are the REST interfaces of the Open Digital Architecture (ODA), used across the industry for catalogues, ordering, inventory, assurance, billing, party management, testing, IoT, AI and partner APIs (Open Gateway). Through its API Apache 2.0 project, TM Forum publishes the OpenAPI files on github.com/tmforum-apis. We checked the licence of every repository at the exact release we used. We left out the one API without a licence file (TMF667 Document), the repository template and one repository that has no OpenAPI file yet. We used the latest release of each API: 38 are at v5, the others at v4 or older.

The models

Area TM Forum APIs Models Model names
Catalogues TMF620, TMF633, TMF634, TMF662, TMF767, TMF936 22 ProductCatalog, Category, ProductOffering, ProductOfferingPrice, ProductSpecification, ServiceCatalog, ServiceCategory, ServiceCandidate, ServiceSpecification, ResourceCatalog, ResourceCategory, ResourceCandidate, ResourceSpecification, EntityCatalog, EntityCategory, EntityCatalogItem, EntitySpecification, AssociationSpecification, Association, ProductUsageSpecification, OpenGatewayProductOffering, OpenGatewayProductSpecification
Ordering, qualification and sales TMF622, TMF641, TMF652, TMF648, TMF663, TMF645, TMF679, TMF760, TMF680, TMF699, TMF651, TMF687, TMF700, TMF697, TMF716 21 ProductOrder, ServiceOrder, ServiceOrderSpecification, ResourceOrder, Quote, ShoppingCart, CheckServiceQualification, QueryServiceQualification, CheckProductOfferingQualification, QueryProductOfferingQualification, CheckProductConfiguration, QueryProductConfiguration, QueryProductRecommendation, SalesLead, SalesOpportunity, Agreement, AgreementSpecification, ProductStock, ShippingOrder, WorkOrder, ResourceReservation
Inventory TMF637, TMF638, TMF639, TMF664, TMF685, TMF703, TMF640, TMF702, TMF730 7 Product, Service, Resource, ResourceFunction, ResourcePool, Reservation, TmfEntity
Assurance and performance TMF621, TMF642, TMF656, TMF653, TMF657, TMF649, TMF628, TMF724, TMF655 18 TroubleTicket, TroubleTicketSpecification, Alarm, ServiceProblem, ServiceProblemEventRecord, ServiceTest, ServiceTestSpecification, ServiceLevelObjective, ServiceLevelSpecification, Threshold, ThresholdRule, ThresholdJob, MeasurementCollectionJob, PerformanceIndicatorGroupSpecification, PerformanceIndicatorSpecification, TrackingRecord, Incident, ChangeRequest
Party, customer and identity TMF629, TMF632, TMF669, TMF668, TMF644, TMF683, TMF717, TMF720, TMF691, TMF672, TMF681, TMF646, TMF658, TMF671 27 Customer, Individual, Organization, PartyRole, PartyRoleSpecification, Partnership, PartnershipSpecification, PartyPrivacyAgreement, PartyPrivacyProfile, PartyPrivacyProfileSpecification, PartyInteraction, Customer360, DigitalIdentity, Userinfo, PermissionSet, PermissionSpecification, PermissionSpecificationSet, CommunicationMessage, Appointment, ProgramProductSpec, LoyaltyRule, LoyaltyAction, LoyaltyCondition, LoyaltyEventType, LoyaltyMember, LoyaltyBalance, Promotion
Accounts, billing, payment and usage TMF666, TMF670, TMF676, TMF678, TMF654, TMF635, TMF677, TMF771, TMF735, TMF736, TMF737, TMF738 24 PartyAccount, BillingAccount, SettlementAccount, FinancialAccount, BillFormat, BillPresentationMedia, BillingCycleSpecification, PaymentMethod, Payment, Refund, CustomerBill, AppliedCustomerBillingRate, BillCycle, Bucket, AccumulatedBalance, BalanceActionHistory, ProductUsage, UsageConsumptionReport, ResourceUsage, ResourceUsageSpecification, CdrTransaction, PartyRevSharingAlgorithm, PartyRevSharingReport, PartyRevSharingModel
Places and shipping TMF673, TMF674, TMF675, TMF684 4 GeographicAddress, GeographicSite, GeographicLocation, ShipmentTracking
Test management TMF704, TMF705, TMF706, TMF707, TMF708, TMF709, TMF710 19 TestCase, TestSuite, NonFunctionalTestModel, AbstractEnvironment, ConcreteEnvironmentMetaModel, TestResourceAPI, ProvisioningArtifact, TestDataInstance, TestDataSchema, TestCaseResult, TestSuiteResult, NonFunctionalTestResult, TestEnvironmentAllocationExecution, TestEnvironmentProvisioningExecution, TestCaseExecution, TestSuiteExecution, NonFunctionalTestExecution, TestScenario, GeneralTestArtifact
Process, intent, events and risk TMF701, TMF921, TMF688, TMF696 14 Process, ProcessSpecification, Task, TaskSpecification, Intent, IntentSpecification, IntentReport, Topic, Event, ProductOfferingRiskAssessment, PartyRoleRiskAssessment, PartyRoleProductOfferingRiskAssessment, ShoppingCartRiskAssessment, ProductOrderRiskAssessment
Component suites (IoT, AI, Open Gateway, NaaS, self-care) TMF908, TMF914, TMF915, TMF931, TMF909, TMF910 18 IotDevice, IotDeviceSpecification, DataAccessEndpoint, IotDataEvent, IotManagementEvent, IotService, IotServiceSpecification, User, AiModel, AiModelSpecification, AiContract, AiContractSpecification, AiContractViolation, Rule, ApiProduct, ApiProductOrder, Application, ApplicationOwner

Some APIs, mostly the component suites, only re-publish resources that another API already defines, such as Service, Resource or Alarm. Those resources are modelled once, from the API that owns them. Operations like cancelling an order or acknowledging an alarm are not separate models, and neither are the notification payloads.

Examples

For the 78 models from v5 APIs, the examples are the example payloads in TM Forum’s own OpenAPI files (Apache-2.0, illustrative, not real operator data). The older v4 files have no example payloads, so the other 96 examples are generated from the schema and its example values. Each example says which kind it is in its source. Many of these resources hold personal data in real deployments (individuals, customers, contacts, digital identities), so all examples are fictitious. Names of people are removed or replaced by “Jane Doe”, e-mail addresses use example.com, phone numbers come from the ranges reserved for fiction, and postal addresses and identity-document numbers are placeholders.

Translation notes

  • TMF names, verbatim. Every attribute keeps its TMF name. Each @context maps it to its definition in the versioned OpenAPI file on GitHub, and all 3,181 of these links were checked.
  • TMF polymorphism kept. @type, @baseType and @schemaLocation become tmfType, tmfBaseType and tmfSchemaLocation, and the TMF id becomes tmfId. A few other names that clash with NGSI-LD or with common Smart Data Models attributes also get a tmf prefix, and TMF703’s Entity becomes TmfEntity.
  • A connected graph. When a TMF attribute references a resource that is also a model here, it becomes an NGSI-LD Relationship. For example, a product offering points to its specification and prices, a product to its offering and billing account, and a service problem to the affected service, resource and trouble ticket. Characteristics, prices, related parties and order items keep their full TMF structure as nested JSON.
  • One generator. A single generic script, driven by a table of API, repository, release, file and resources, produces all 174 models. It was also used to regenerate the five original models, which move to schema version 0.0.2 because their references to catalogue, billing and intent resources are now Relationships.
  • Source issues documented. The metadata lists the small problems found in the specifications, such as invalid timestamps, example attributes missing from the schemas, dangling references and copyright lines left over in some files, for reporting upstream.

Try it, and tell us what’s missing

These are Candidates: carefully sourced translations, not yet official Smart Data Models subjects. The main open question before promotion is naming. Many TMF names are generic (Organization, Event, Alarm, Rule, Task, User, Product, Service…) and already exist in Smart Data Models or other candidates with other meanings. There is also scope overlap with dataModel.Organization, dataModel.Device and dataModel.IssueTracking.

Browse the models, open an issue, or send a PR:
👉 https://github.com/smart-data-models/Candidates/tree/master/standards/tmforum-open-apis

New pysmartdatamodels Version 0.8.2.0: Semantic Identification for Data Spaces

We are happy to announce a new release of our Python package, pysmartdatamodels, version 0.8.2.0. The headline feature is aimed squarely at data spaces: a new function that can give semantic meaning to a payload even when the party who sent it never provided one.

What’s New in This Version?

Semantic identification for data spaces: identify_or_draft_datamodel()

In a data space, you routinely receive payloads from other participants with no guarantee they follow any particular standard. identify_or_draft_datamodel() takes any such payload and does one of two things:

  • If it recognizes the payload as an existing Smart Data Model, it returns that model’s real schema, together with a full validation of the payload against it.
  • If it doesn’t, it drafts a brand new schema.json on the spot — following the same structural conventions as a real Smart Data Model — so you have a ready-to-review starting point instead of nothing at all.

It accepts the payload in any of the three formats actually used in practice — plain key-values, NGSI-v2 normalized, or NGSI-LD normalized — detecting and handling the right one automatically:

from pysmartdatamodels import pysmartdatamodels as sdm

payload = {
    "id": "urn:ngsi-ld:WeatherObserved:station-042",
    "type": "WeatherObserved",
    "temperature": 21.5,
    "relativeHumidity": 0.6
}

result = sdm.identify_or_draft_datamodel(payload)
print(result["schema"]["source"])   # "existing" -- it recognized WeatherObserved
print(result["schema"]["validation"]["result"])  # True

When the payload’s own “type” doesn’t match anything, you can optionally ask it to try a softer, attribute-based match against the whole catalog before giving up and drafting a new schema:

result = sdm.identify_or_draft_datamodel(payload, fuzzy=True, fuzzy_threshold=0.3)

And when nothing matches at all, the drafted schema isn’t just a bare shape guess — any attribute whose name is already established elsewhere in the catalog (over 160,000 attribute definitions) gets its real description, model reference, and units reused automatically, so only genuinely new attributes are left with a plain TODO placeholder.

A real validate_payload()

validate_payload(datamodel, subject, payload) existed before but didn’t actually validate anything. It now runs full JSON Schema validation against the live schema of the data model you name, confirms the payload’s “type” matches, and separately flags – without failing – any attribute that isn’t part of the official definition.

Reliability and performance fixes

  • generate_sql_schema() no longer crashes on attributes that can hold more than one type (e.g. a Property that can also be a Relationship), and no longer generates colliding enum type names when two unrelated models happen to share an attribute name.
  • validate_data_model_schema() no longer terminates your entire Python process on a bad input — it returns an error result like every other function in the package.
  • The ~160,000-entry attribute database is now cached in memory instead of being re-read from disk on every single function call — noticeably faster for any code calling the per-attribute lookup functions in a loop.

Get the Latest Version

Update your installation with:

pip install --upgrade pysmartdatamodels

➔️ Get the package on PyPI

New pysmartdatamodels Version 0.8.1.0: Refreshed Data (reduced size 20 times with same functionality)

We are happy to announce a new release of our Python package, pysmartdatamodels, version 0.8.1.0, with a refreshed attribute database, a smaller download, and a cleanup of some leftover dead code.

What’s New in This Version?

Refreshed Data, End to End

The package’s bundled data — the official list of data models, the attribute metadata, and the full attribute database — had gone stale. Fixing this surfaced two real issues in our own publication pipeline, now resolved:

  • One data model repository had accidentally been left private, which was silently breaking the daily metadata regeneration every time it ran.
  • The database backing the attribute export (see below) was far less complete than it should have been. It’s now fully backfilled: 164,926 attribute records across all 82 subjects.

As always, you don’t need to upgrade the package just to get fresh data — calling update_data() pulls the latest information directly:

from pysmartdatamodels import pysmartdatamodels as sdm
sdm.update_data()

Smaller Install, Compressed Attribute Database

The full attribute database is now shipped gzip-compressed inside the package (~5MB instead of ~110MB) — the installed package footprint drops from over 100MB to under 10MB. This is fully transparent if you use the documented functions (load_all_attributes(), description_attribute(), etc.) — nothing changes in how you call them.

The same compressed file is now also available for direct download, alongside the existing uncompressed one, from the attribute database export page.

Cleanup

Removed a small piece of dead code in generate_sql_schema() left over from an earlier implementation — no change in behavior, just a cleaner function. Thanks to the community member who originally reported it.

Get the Latest Version

Update your installation with:

pip install --upgrade pysmartdatamodels

➡️ Get the package on PyPI

New pysmartdatamodels Version 0.8.0.11: More Reliable SQL Schema Generation

We are happy to announce a new release of our Python package, pysmartdatamodels, version 0.8.0.11, focused on making the SQL schema generation function more robust, fixing how the package exposes its functions, and keeping our GitHub source and published package in sync.

What’s New in This Version?

More Reliable SQL Schema Generation

The generate_sql_schema() function, which turns a data model’s model.yaml into a ready-to-use PostgreSQL CREATE TABLE statement, received three fixes:

  • Column names are now properly quoted, so attribute names that are reserved words or mixed case no longer produce invalid SQL.
  • Properties using anyOf are now supported, in addition to oneOf.
  • The function no longer emits a duplicate id column in the generated schema.

Example:

from pysmartdatamodels import pysmartdatamodels as sdm

sql = sdm.generate_sql_schema("https://raw.githubusercontent.com/smart-data-models/dataModel.Weather/master/WeatherForecast/model.yaml")
print(sql)

More Usable Imports

You can now import the package’s main functions directly from the top level, instead of always going through the submodule:

from pysmartdatamodels import generate_sql_schema, load_all_datamodels

The documented usage pattern below continues to work exactly as before:

from pysmartdatamodels import pysmartdatamodels as sdm

Repository Housekeeping

The source code published on GitHub and the package published on PyPI had drifted apart over the last few releases. With this version, both are back in sync, so what you see in the repository is exactly what you get when you pip install the package.

Get the Latest Version

Update your installation with:

pip install --upgrade pysmartdatamodels

➡️ Get the package on PyPI

Our Commitment to Open and Interoperable Data

We are committed to making data interoperability easier for everyone. Thanks to everyone in the community who reports issues and helps us keep improving pysmartdatamodels!

UNTP Digital Product Passport Joins Smart Data Models: Two new candidate models — DigitalProductPassport and Product

Smart Data Models’ Candidates repository now includes a new candidate standard, untp-v0.8.0, translated from the UN Transparency Protocol (UNTP) Digital Product Passport specification — a B2B data standard, issued as a W3C Verifiable Credential, for carrying product and sustainability information through supply chains. Two entities are covered: DigitalProductPassport and Product.

What’s in it

Model What it is
DigitalProductPassport The W3C Verifiable Credential (VCDM 2.0) envelope: issuer, validity period, credential status, render method, issuing software.
Product The core product data model: identification, classification, material provenance, producing facility, dimensions, packaging, performance claims, and an open industry-specific characteristics extension point.

Every model ships the full set of Candidates artifacts: a JSON Schema, an NGSI-LD @context, and all four example serializations (NGSI-v2 and NGSI-LD, key-values and normalized) — each validated against its own schema.

Two entities, one relationship — not one flattened blob

In the source, DigitalProductPassport.json‘s credentialSubject is literally $ref: "#/$defs/Product" — a 1:1 inline embedding. But Product.json is also published standalone, with its own complete identity independent of any credential wrapper. This candidate keeps that distinction: DigitalProductPassport carries only the envelope fields plus a product Relationship, and Product is its own separately-addressable entity — avoiding a ~20-property duplication and matching how the source itself chose to publish two files, not one.

The VC envelope fields (issuer DID, credential status, render method, issuing software) are modeled faithfully as real typed properties rather than simplified away — even though they’re strictly artifacts of the credential-verification layer, not “product data” per se.

Real context, real examples — not invented

Two things set this translation apart from working off prose docs alone:

  • A genuine published JSON-LD context. UNTP ships a complete, ~106KB context file using JSON-LD type-scoped contexts. This candidate’s context.jsonld IRIs are extracted from that real file — with two scope-collision bugs caught and hand-corrected along the way (a naive flatten of a type-scoped context can silently grab a same-named term from an unrelated class; here it briefly mapped the envelope’s validFrom to an unrelated ConformityProfile field, and the new product Relationship to an unrelated Digital Traceability Event term, before being fixed).
  • Real sample data. Both example.json files are drawn from UNTP’s own published sample instance — a 75 kWh Li-ion EV battery pack passport, complete with real material provenance (cobalt from DRC, lithium from Chile, nickel from Indonesia, …) and EU Battery Regulation performance claims — not fabricated placeholder values.

One naming fix was needed: the source’s own @context property (the W3C VC context array) would collide with NGSI-LD’s reserved top-level @context mechanism, so it’s renamed to vcContext — the same pattern already used for a dataProvider collision in the Gaia-X candidate.

Note on scope

Per explicit instruction, only these two entities are covered. UNTP defines several other credential types this specification references — Digital Conformity Credential, Digital Facility Record, Digital Traceability Event, Digital Identity Anchor — none of which are modeled here. Fields that reference them (e.g. producedAtFacility, a claim’s evidence links) stay exactly as the source declares them, not promoted to Relationships pointing at not-yet-modeled entity types.

Try it, and tell us what’s missing

This is a Candidate — a first, carefully-sourced translation, not yet promoted to an official Smart Data Models subject. Before that step:

  • License terms haven’t been confirmed against the spec-untp repository’s own LICENSE file yet.
  • Overlap with other Smart Data Models product/traceability vocabularies hasn’t been checked in detail.

Browse the models, open an issue, or send a PR: 👉 https://github.com/smart-data-models/Candidates/tree/master/standards/untp-v0.8.0

Gaia-X Ontology Joins Smart Data Models: 165 Candidate Models for Cloud, Connectivity & Trust

Smart Data Models’ Candidates repository now includes a new candidate standard, gaiax-ontology-v2111, translated from the Gaia-X ontology v2111 — the vocabulary behind Self-Descriptions in Gaia-X federated data spaces and cloud ecosystems. All 165 classes published at docs.gaia-x.eu/ontology/v2111/classes now have an NGSI-LD / NGSI-v2 representation.

What’s in it

Gaia-X’s ontology is broader than a typical device/observation vocabulary — it describes both the physical/technical side of an offering and the legal/governance claims attached to it. That split shows up directly in the model mix:

Category Models Examples
Legal & governance documents 50 LegalDocument, TermsAndConditions, DataUsageAgreement, ServiceAgreementOffer
Compute & software 27 CPU, GPU, Memory, Hypervisor, ComputeFunctionRuntime, Image
Security 25 Encryption, PhysicalSecurity, InformationSecurityPolicies, ProductSecurity
Network 23 Endpoint, VLANConfiguration, InterconnectionServiceOffering, PointOfPresence
Infrastructure 19 Datacenter, AvailabilityZone, Region, PhysicalResource
Service offerings 17 ServiceOffering, ComputeServiceOffering, StorageServiceOffering
QoS metrics 13 Latency, Jitter, PacketLoss, Throughput, IOPS
Storage 13 BlockStorageConfiguration, ReplicationPolicy, SnapshotPolicy
Data products 13 DataProduct, DataProductCatalogue, EvidenceTemplate
Trust framework 10 Ecosystem, EcoTrustScope, EcoTSP, CompliantCredential
+ identifiers, measurement, FaaS, environmental, and more ~19 VatID, EORI, EnergyMix, WaterUsageEffectiveness

Every model ships the full set of Candidates artifacts: a JSON Schema, an NGSI-LD @context, and all four example serializations (NGSI-v2 and NGSI-LD, key-values and normalized) — each validated against its own schema.

Translation notes

  • Names preserved verbatim. Gaia-X’s own class and property names (hasPoint-style camelCase, PascalCase classes) are kept exactly as the ontology defines them — no forced renaming.
  • Cross-references become NGSI-LD Relationships. Any property whose range is another Gaia-X class (e.g. Datacenter.aggregationOfResources → AvailabilityZone) is modeled as a Relationship, not an inlined object, so the resulting graph mirrors the ontology’s own linking structure.
  • Inheritance is flattened, not duplicated in text. A subclass’s schema only lists its own properties; inherited ones are documented on the parent model and referenced by name in the description, matching how the source ontology itself organizes properties.
  • Open vocabularies stay open. A number of Gaia-X properties reference controlled vocabularies (e.g. DiskType, HypervisorType) that aren’t enumerated on the class documentation pages themselves; those are modeled as open strings with a note in the description rather than an invented closed enum.
  • One naming collision, resolved explicitly. DataUsageAgreement‘s own dataProvider property (the participant providing a Data Product) would collide with the Candidates convention’s boilerplate dataProvider metadata field (harmonised-data-entity provider). It’s exposed as gxDataProvider instead, with the substitution documented in the schema’s own description.

Try it, and tell us what’s missing

This is a Candidate — a first, carefully-sourced translation, not yet promoted to an official Smart Data Models subject. A few things are open and would benefit from community input before that step:

  • Overlap with existing Smart Data Models subjects (dataModel.Device, generic infrastructure/building vocabularies) hasn’t been checked in detail.
  • License terms for the specific v2111 release should be confirmed before promotion out of Candidates.
  • A handful of controlled vocabularies referenced by the ontology (e.g. disk/hypervisor/firmware type enumerations) aren’t resolved to closed lists yet — see the individual schema descriptions for which ones.

Browse the models, open an issue, or send a PR: 👉 https://github.com/smart-data-models/Candidates/tree/master/standards/gaiax-ontology-v2111