Release history and notes
*************************

Sequence based identifiers are used for versioning (schema follows
below):

   major.minor[.revision]

* It is always safe to upgrade within the same minor version (for
  example, from 0.3 to 0.3.4).

* Minor version changes might be backwards incompatible. Read the
  release notes carefully before upgrading (for example, when
  upgrading from 0.3.4 to 0.4).

* All backwards incompatible changes are mentioned in this document.


0.6.1
=====

2026-04-22

* Improve OGL patterns.


0.6
===

2026-04-11

* Added jurisdiction and scope detection for Creative Commons license
  URLs (e.g., "http://creativecommons.org/licenses/by-nc/2.0/uk" →
  jurisdiction "uk", "http://creativecommons.org/licenses/by-
  nc/3.0/igo" → scope "igo").

* Added "jurisdiction" and "scope" optional fields to "LicenceVersion"
  model.

* Added prose patterns for CC URLs embedded in text (e.g., "Article
  under https://creativecommons.org/licenses/by-sa/3.0/").


0.5.2
=====

2026-04-10

* Minor fixes.


0.5.1
=====

2026-04-04

* Improved Creative Commons recognition (additional aliases, versions
  and IGO variants) and expanded free-text matching.

* Enhanced traceability: resolved licences include source line/file
  info surfaced in explain outputs.

* Added integration and trace tests for CC matching and source-line
  explanations.

* Docs updated to reflect traceability and plugin interface changes.


0.5
===

2026-03-27

**Major refactoring and data consolidation**

This release significantly simplifies the data model and maintenance
while preserving full backwards compatibility for the public API.

* **Merged** *url_map.json* and *publishers.json* into a single
  unified *aliases/aliases.json* file.

  * All URL mappings and publisher-specific shorthand aliases are now
    stored together with rich metadata (including *urls* arrays and
    *aliases* lists).

  * Removed separate *data/urls/* and *data/publishers/* directories
    (and their parsers).

  * Updated all documentation, tests, scripts, and contributing
    guidelines accordingly.

* **Simplified internal lookups**:

  * Consolidated URL handling: *LicenceNormaliser* now builds a single
    *_vkey_to_url* map (version_key → list of URLs) and selects the
    shortest/cleanest URL for display.

  * Alias and URL resolution now use the same *AliasParser* (which
    also implements *URLPlugin*).

  * Trace output now consistently points to *aliases.json* with
    correct line numbers for both aliases and URLs.

* **Improved family categorisation**:

  * Fixed wrongly categorised licences (notably *unlicense*, *wtfpl*,
    and several others now correctly map to *public-domain* where
    appropriate).

  * Updated fallback logic in *_infer_family()* and related tests.

* **Other fixes and cleanups**:

  * Minor improvements to error handling and DataSourceError wrapping.

  * Updated exception hierarchy documentation (kept
    *LicenceNormalisationError* for backwards compatibility).

  * Cleaned up unused/deprecated references across the codebase.


0.4
===

2026-03-26

* Renamed project from *license-normaliser* to *licence-normaliser*.

* Cleanup.


0.3.2
=====

2026-03-25

* Introduce sub-aliases in aliases.json.

* Introduce handy scripts for finding, automatic filling and merging
  the duplicates.


0.3.1
=====

2026-03-24

* Minor fixes.


0.3
===

2026-03-23

* **Plugin-based architecture** — *LicenceNormaliser* class accepts
  plugin CLASSES (not instances) with lazy loading; replaces old
  module-level globals.

* **6 plugin interfaces** — *BasePlugin*, *RegistryPlugin*,
  *URLPlugin*, *AliasPlugin*, *FamilyPlugin*, *NamePlugin*,
  *ProsePlugin*.

* **New `BasePlugin.refresh()`** — classmethod to fetch/refresh data
  from upstream URLs; CLI uses parser IDs (*spdx*, *opendefinition*,
  etc.) instead of class names.

* **Thread-safe singleton** — *_DefaultNormaliser* class with double-
  checked locking protects the shared *LicenceNormaliser* instance.

* **Public `registry_keys()`** — *LicenceNormaliser.registry_keys()*
  exposes known keys; *get_registry_keys()* in *_cache.py* uses it.

* **URL population fix** — inverted URL map (*version_key →
  cleaned_url*) ensures *LicenceVersion.url* is populated for resolved
  licences.

* **Removed dead code** — deleted *_registry.py*, *parsers/base.py*,
  *parsers/__init__.py*, and empty *DEFAULT_** module globals.

* **Docs updated** — ARCHITECTURE.rst, AGENTS.md, README.rst rewritten
  to reflect new plugin architecture.


0.2
===

2026-03-21

* **Architecture rewrite** — parser-based pluggable system with 8
  parsers loading from JSON data files.

* **3-level hierarchy** — *LicenceFamily → LicenceName →
  LicenceVersion* with 11 families including publisher OA/TDM.

* **New data files** — 170+ aliases, 41 prose patterns, 50+ publisher
  URLs (Elsevier, Wiley, Springer, ACS, etc.).

* **Strict mode** — *strict=True* raises *LicenceNotFoundError* on
  unknown licences.

* **Bug fixes** — *cc-pdm*/*cc0* family inference, *gpl-3.0+*
  *+*-suffix stripping, false-positive prose pattern removed, ALIASES-
  before-REGISTRY lookup order.

* **Tests** — 386 tests (up from ~80), including 165-case integration
  matrix.

* **Docs** — *AGENTS.md*, *ARCHITECTURE.rst*, *CONTRIBUTING.rst*
  rewritten; stale *literalinclude* targets fixed.


0.1.1
=====

2026-03-17

* Remove Pydantic as dependency.


0.1
===

2026-03-17

* Initial beta release.