Add CHANGELOG.md

CHANGELOG.MD

@@ -0,0 +1,27 @@
+# Changelog
+
+All notable changes to the Dicta Unvocalized Hebrew Wordlist dataset.
+
+## [attested-1.0.0] — pending upload
+
+### Added
+- **`attested` config** — a second table of *attested wordforms*: every distinct
+  `(wordform, prefix, morphology)` actually seen in a Dicta corpus, across five genres
+  (Modern, Rabbinic, Talmud, ModernPoetry, MedievalPoetry). 695,870 rows.
+- Columns: `wordform` (the seen unvocalized spelling, prefix removed), `prefix`, the morphology
+  ClassLabels (`pos`/`binyan`/`gender`/`number`/`person`/`tense`/`status`/`suffix_*`), and
+  `genres` (list of genres the form was attested in).
+- Unlike the `default` table (which enumerates every *theoretically* valid word-object),
+  `attested` records only spellings observed in corpora. It is ~99% a subset of the
+  `default` table's known spellings (`wordform` matches `default`'s `word`, `haser`, or a
+  member of `maleh_forms`); the residual is out-of-wordlist POS such as `INITIALISM` and a few other
+  pathologies.
+
+## [1.0.0] — initial upload
+
+### Added
+- **`default` config** — the Dicta Unvocalized Hebrew Wordlist: every distinct unvocalized
+  (nikud-stripped) Hebrew word form Dicta knows, with morphology, lemma, and root.
+  4,056,791 rows.
+- All categorical columns as `ClassLabel` (integer codes + names in the Parquet metadata).
+- `lemma_id` (one per unique vocalized `lex`).