Radixor model-data notice

Radixor-derived model data

Copyright (C) 2026, Leo Galambos.

Copyright and, where applicable, database rights are claimed in the
Radixor-specific selection, verification, cleaning, normalization,
grouping, deduplication, filtering, reformatting, metadata preparation,
and packaging of this model, to the extent protected by applicable law.

The underlying morphological data remains attributed to UniMorph and
the upstream contributors identified in this notice.

This derived model data, including Radixor's protectable contributions,
is distributed under Creative Commons Attribution-ShareAlike 3.0.

Model ID: et-ee-default
Source project: UniMorph
Official repository: https://github.com/unimorph/est
Selected upstream data:
- est [lemma-form-features], SHA-256 602b778567b4607514ee380152fe8a6ccd70dbc15190339f7f5a6d9a710b899f
Source origin: Wikipedia
Attribution: UniMorph
License:
Creative Commons Attribution-ShareAlike 3.0 Unported

Canonical license URI: https://creativecommons.org/licenses/by-sa/3.0/
Source revision: 97663990872e6bbc06701839d33462833676ebc2
Revision status: recorded
Repository evidence paths: README.md:evidence-00-README.md
Repository evidence SHA-256: evidence-00-README.md:dee0ff4e57b5133dcfed8ff147aae36fad03814cab6997de3f4aa921ee58e9c7
Import policy: radixor-unimorph-import-v1
Projection semantics: Each declared three-column source directly projects lemma and surface form.
Conflict semantics: All valid lemma-form assignments are retained. Surface ambiguity is reported explicitly and is never resolved by dropping or overwriting an assignment.
Import rows: source=39105; accepted=27454; rejected=11651
Rejection reasons: blank-row=890;form-unicode-whitespace=10761
Per-source import audit:
- est: source=39105; accepted=27454; rejected=11651; added-groups=886; added-distinct-surfaces=24811; added-source-pairs=24890; added-output-forms=24890; added-changed-pairs=24004; intrinsic-ambiguous-surfaces=72; rejection-reasons=blank-row=890;form-unicode-whitespace=10761
Surface conflict audit: cross-source-assignments=0; cross-source-surfaces=0; intrinsic-ambiguous-surfaces=72
Audited excluded upstream files:
- none
Generated dictionary: groups=886; forms=24890
Deterministic GZip SHA-256: 0d7da990392011b0312fdee181720c2778ff07581912c718b8e7c1d07267c735

Radixor modifications: Strict UTF-8 validation, NFC normalization, lowercase conversion, structural unsafe-character rejection, grouping by lemma, exact deduplication, deterministic ordering, and reproducible GZip packaging.

Neither UniMorph nor any upstream contributor endorses Radixor.

Upstream information verified: 2026-09-11
