CHANGELOG.md

Changelog

0.8.0 - 2026-06-04

New schemas:

• Added Identifiers service schema (exported as iscc_schema.Identifiers) for registry/gateway responses listing external identifiers for an asset, with optional iscc subject binding.

Metadata:

• Widened IsccMeta.identifier from str | list[str] to accept a typed identifier object or a mixed list of strings and objects. Legacy string forms remain valid.
• Added the iscc_schema.Identifier item model for constructing typed identifier objects with scheme, code, optional scope, and optional primary.
• Identifier objects are bare when nested in IsccMeta or Identifiers: no nested @context, $schema, or @type is emitted.

Service metadata:

• Added optional iscc to standalone GenAI, aligning it with TDM and Identifiers as a subject-bearing service record.
• Made all Service objects forward-compatible with extension fields (additionalProperties: true), including standalone and inline GenAI.
• Preserved falsy extension-field values on Service objects, so values such as 0.0, [], {}, "", and false are no longer normalized away.
• Made recover_context(..., schema="<name>-<version>.json") work with bare versioned standalone schema names.
• Rejected empty string entries inside IsccMeta.identifier lists; the published JSON Schema also marks identifier string branches with minLength: 1.
• Tightened the published JSON Schema's identifier array branch to validate every element instead of only the first; strict external validators may now reject payloads they previously accepted. The Pydantic model already validated all elements, so runtime behavior is unchanged.

0.7.0 - 2026-06-01

This release adds the Protocol schema category (ISCC Discovery Protocol wire records) alongside the existing Metadata, Seed, and Service categories, introduces an STM seed schema for scholarly works, makes compact self-describing JSON the default for seed and protocol records, and version-pins the $schema and @context of every standalone schema. The $schema reference makes compact data recoverable to JSON-LD on demand, conformant with IEP-0002 which accepts both application/json and application/ld+json as meta element formats.

Breaking changes:

• IsccMeta is unaffected. It still defaults to ld=True and serializes full JSON-LD exactly as before. The breaking changes below apply only to the standalone Seed/Service/Protocol models.
• Seed models (ISBN, ISRC, STM) now default to compact JSON (ld=False): .dict(), .json(), and .jcs() emit $schema only, with no @context/@type. Pass ld=True to restore full JSON-LD.
• Standalone schema $schema URLs are now version-pinned (e.g. …/isbn-0.7.0.json, …/tdm-0.7.0.json, …/iscc-note-0.7.0.json). Code that matched the old unversioned $schema string must account for the -X.Y.Z suffix. recover_context() resolves both forms, so v0.6.0 records still recover.
• $schema is now a required field on the ISBN, ISRC, STM, and IsccNote schemas (auto-populated from its const default on construction, so existing construction code keeps working).

New schemas:

• Added IsccNote protocol schema (exported as iscc_schema.IsccNote) — the permanent ISCC Declaration log record for the ISCC Discovery Protocol HUB, and the first member of the new Protocol category. Serializes to compact JSON by default and requires $schema as the sole version anchor; $schema is part of the signed jcs() bytes, pinning the schema version into the signed record. Required: $schema, iscc_code, datahash, nonce, signature.
• Added STM seed schema (exported as iscc_schema.STM) — Scientific/Technical/Medical seed metadata for DOI-identified scholarly works, for interoperable Meta-Code generation (IEP-0002). Required: doi, resource_type, title, publisher, pubyear. Optional: version_type (NISO JAV stage), container_title, issn. resource_type carries 17 DataCite-style work-type tokens mapped to resolvable schema.org/FaBiO class IRIs.

Serialization:

• Added ld parameter to .dict(), .json(), .jcs() for controlling JSON-LD output
• Seed models (ISBN, ISRC, STM) default to ld=False (compact JSON with $schema only)
• Service models (TDM, GenAI) default to ld=True (full JSON-LD for registry/gateway interop)
• Protocol model (IsccNote) defaults to ld=False (compact JSON, $schema as version anchor)
• IsccMeta defaults to ld=True (full JSON-LD, backward compatible)
• Changed seed metadata examples to compact recoverable JSON format (no @context/@type)
• Added "Recommended Format" documentation to ISBN and ISRC schema pages
• Added code-gen post-processing for const field defaults and _default_ld class attribute
• .json() now accepts an indent argument for pretty-printing compact records and forwards exclude_unset, so the compact path offers the same formatting controls as the JSON-LD path

Fixes:

• End-anchored the IsccNote gateway URL pattern ($) so values with embedded whitespace or trailing characters are rejected rather than silently accepted into a permanent declaration record
• Corrected the IsccNote @type description ("service metadata" → "protocol metadata")

Versioning:

• All standalone schemas (seed, service, protocol) now uniformly version both $schema and the default @context, and write a version-pinned archive (<name>-0.7.0.json) next to the latest file; the latest file keeps an unversioned $id
• Added x-iscc-jsonld extension to every standalone JSON Schema documenting the JSON-LD upgrade path (versioned context URL, @type, human-readable recipe) for compact records
• Versioned the @context/$schema URLs shown in the main IsccMeta docs example

Field status & tooling:

• Promoted $schema, nonce, signature, generator, and tdm_reservation to stable
• Documented the field-stability contract (stable/draft, one-way promotion) and the bump-first release workflow in the versioning guide
• Added a release-archive build guard: the schema and context builds abort if a git tag already exists for the current version, preventing silent mutation of released archives (fails open when git is unavailable)

Documentation:

• Templated the version string in hand-written docs (guide, examples, versioning) via Zensical's macros extension, so version-pinned example URLs derive from iscc_schema.__version__ and never go stale on a release bump; the {{ version }} variable is supplied by tools/docs_macros.py
• Made the README version-free (it renders on GitHub/PyPI where macros do not run) and taught the llms-full.txt generator to resolve the version token

0.6.0 - 2026-05-26

• Replaced five per-category TDM fields (train, inference, derive, search, analyze) with two W3C TDMRep-conformant fields (tdm_reservation, tdm_policy)
• Added tdm_reservation integer field (0/1) bound to http://www.w3.org/ns/tdmrep#reservation
• Added tdm_policy URI field bound to http://www.w3.org/ns/tdmrep#policy
• Changed embedded tdm wrapper to use JSON-LD @nest semantics (no intermediate blank node in RDF expansion)
• Added iscc field to standalone TDM schema for binding declarations to content by ISCC-CODE or ISCC-ID
• Embedded and standalone TDM objects now allow additional properties for backward compatibility with v0.5.0 payloads
• Fixed empty_string_to_default validator coercing integer 0 to None (now only exempts int, not bool/list/dict)
• Per-category fields deferred until IETF AIPREF vocabulary stabilizes (target August 2026)

0.5.0 - 2026-03-17

• Migrated documentation from mkdocs-material to zensical
• Added ISCC brand theming (header, footer, dark mode, logo)
• Added copy-page button for LLM-friendly markdown export
• Added ISCC-AI copilot widget integration
• Added llms.txt and per-page markdown generation for LLM consumption
• Added "For Coding Agents" reference page
• Added GitHub Pages deployment workflow
• Added navigation icons and short titles to all doc pages
• Added Open Graph and Twitter Card meta tags
• Added Plausible analytics via theme override
• Added x-iscc-status (stable/draft) to all fields across YAML schemas
• Rendered x-iscc-status in generated schema docs, vocabulary page, and terms includes
• Added datasize field for data processed size when ISCC is computed over sub-file data (e.g., bioimage planes)
• Added GenAI Service Metadata schema for generative AI disclosure signals (involvement, ai_system, digital_source_type)
• Added genai field to IsccMeta for embedding AI transparency signals in content metadata
• Moved tdm field from iscc-embeddable to iscc-extended (structured objects don't belong in media-embeddable metadata)
• Added form field for content-kind classification using Schema.org CreativeWork subtypes
• Pydantic models now emit versioned $schema and @context URLs (e.g., http://purl.org/iscc/schema/0.5.0.json)
• Standalone schema models (ISBN, ISRC, TDM) now emit versioned @context URLs
• Added versioning documentation page
• Embedded JSON-LD @context directly into JSON Schema files for self-contained validation and semantic mapping
• Added schema-specific @context to standalone schemas (isbn, isrc, tdm) with only their relevant terms
• Patched @context property in JSON Schemas to accept both URI string and inline object per JSON-LD spec
• Added recover_context() function for schema-driven JSON-LD context recovery from plain JSON data
• Added generated contexts.py module with bundled context data and type-to-schema mappings
• Reordered build pipeline so buildcontext runs before buildschema
• Added ISBN and ISRC Seed Metadata schemas for interoperable Meta-Code generation (IEP-0002)
• Added TDM Service Metadata schema for machine-readable TDM reservation signals (train, inference, derive, search, analyze)
• Added tdm field to IsccMeta for embedding TDM reservation metadata in content descriptions
• Reframed TDM schema as reservation policy signals aligned with EU DSM Directive Art. 4 opt-out
• Made TDM fields optional (omitted field = reservation status not determined)
• Added additionalProperties: false to TDM schema for strict validation
• Replaced legally loaded terminology in TDM field descriptions (e.g., "derivative works" → "content transformation")
• Introduced three-category schema framework: ISCC Metadata, Seed Metadata, Service Metadata
• Added per-schema documentation pages with JSON Schema links and field reference tables
• Added Seed and Service Metadata terms to JSON-LD context and vocabulary documentation
• Fixed misleading ISCC Metadata description (content vocabulary, not declaration schema)
• Added x-iscc-standard extension field to mark ISO 24138:2024 properties in YAML schemas
• Surfaced ISO 24138:2024 annotations in generated schema docs, vocabulary page, and terms includes
• Added iscc_code field as explicit alternative to the compact iscc field
• Added nonce field for cryptographic replay protection
• Added signature field to IsccMeta for iscc-crypto compatibility (EdDSA/JCS signing)
• Added units field for individual ISCC-UNITs that make up a composite ISCC-CODE
• Added text field for extracted plaintext of digital content
• Widened parts field to accept both strings and objects
• Added minLength: 1 to name and description fields
• Fixed keywords description typo ("sting" → "string")
• Eliminated iscc-collection.yaml duplication; JSON Schema is now auto-generated from individual schema files
• Fixed JSON-LD identifier collision: iscc_id and media_id now have distinct ISCC term mappings
• Fixed URI-typed fields in JSON-LD context to use @type: @id for proper linked data processing
• Added missing mode field to JSON-LD context
• Fixed stale 0.3.2 version defaults in generator reference schema
• Fixed missing x-iscc-status and x-iscc-context annotations in JSON Schema output
• Exported Signature model from package
• Migrated from Poetry to uv with hatchling build backend
• Dropped Pydantic v1, migrated to native Pydantic v2 (closes #36)
• Require Python >=3.10,<3.15
• Added Python 3.13 and 3.14 to CI test matrix
• Switched CI to ubuntu-latest
• Suppressed datamodel-code-generator warnings and cleaned up build output
• Updated .gitignore to comprehensive Python template

0.4.1 - 2024-01-21

• Added credentials-field
• Updated dependencies
• Add pydantic v2 compatibility

0.4.0 - 2022-11-24

• Added test for error on extra fields
• Added typing information to IsccMeta.iscc_obj property
• Updated definition of meta-field
• Updated dependencies

0.3.9 - 2022-07-03

• Fixed identifier element to support stings and list of strings
• Add string/list support for creators and keywords fields
• Updated dependencies

0.3.8 - 2022-06-08

• Updated dependencies
• Refactor NFT reference to support CAIP-22 and CAIP-29 standards

0.3.7 - 2022-04-10

• Added new nft_chain, nft_contract, nft_token fields
• Moved changelog
• Updated examples

0.3.6 - 2022-03-19

• Use customized BaseModel for IsccMeta
• Convert empty strings to None
• Exclude None and unset in IsccMeta.dict() by default
• Exclude None and use by_alis in IsccMeta.json()
• Added IsccMeta.jcs() serialization
• Added IsccMEta.iscc_obj property
• Fixed JSON Schema code generation
• Moved thumbnail field to the bottom
• Added media_id vendor identifier
• Added iscc_id field
• Updated OpenAPI definitions.

0.3.5 - 2022-03-11

• Added custom field AnyUrl to support Data-URLs
• Fix typo in NftFrozen
• Change endpoint /freeze to /nft/freeze

0.3.4 - 2022-03-08

• Mark stable fields
• Change iscc field to not required
• Add schema.org context to thumbnail field
• Updated dependencies

0.3.3 - 2022-03-06

• Fixed typo in acquire field
• Renamed main schema class ISCC to IsccMeta
• Import schemas to package top-level
• Updated dependencies

0.3.2 - 2022-03-01

• Redesigned Generator API model
• Added mode element
• Added thumbnail element
• Added versioned context and schema URIs
• Added $schema element to context
• Renamed verify to verifications

0.3.1 - 2022-02-10

• Fixed packaging error

0.3.0 - 2022-02-10

• Added draft API for ISCC Generator Service
• Added new collection schema
• Updated dependencies
• Added new terms: verify, original, redirect

0.2.1 - 2022-01-19

• Tweak code generator
• Cleanup dependencies

0.2.0 - 2022-01-17

• Added generator field
• Changed properties field to support base64
• Changed iscc validation to support Semantic-Code

0.1.0 - 2022-01-05

• Initial release