v0.1.1

Highlights

• Prepare the Rust and Python packages for the v0.1.1 release.
• Includes Pages demo dictionary loading, Unicode whitespace normalization, fullwidth ASCII aliases, Python process helper alignment, CLI cleanup, UniDic pron defaults, and romaji lattice visualization.

Packages

• Rust crates are published on crates.io as 0.1.1.
• Python wheels and sdist are published on PyPI as 0.1.1.

Dictionary Artifacts

Dictionary artifacts remain separate GitHub Releases and will be refreshed separately.

UniDic-CWJ 2025.12 Reading Index v0.1.1

UniDic-CWJ 2025.12 Reading Index v0.1.1

Release of the moine UniDic-CWJ reading-index artifact for moine v0.1.1.

This release rebuilds the artifact with the pron reading field. It keeps the
same source dictionary version, artifact name, indexed payload format, and
bounded expansion settings as the v0.1.0 artifact release.

Assets:

• moine-unidic-cwj-202512.tar.gz
• moine-unidic-cwj-202512.tar.zst

Source:

• Dictionary: UniDic-CWJ
• Source version: 2025.12
• Reading field: pron
• Runtime payload: indexed FST/mmap-friendly normalized surface -> readings
• Artifact name: moine-unidic-cwj-202512

Build options:

max_readings_per_surface: 16
max_readings_per_segment: 16
max_paths: 128
longest_match_only: true
exclude_ascii_surfaces: true
exclude_symbol_pos: true

Asset checksums:

d37570e7b5a6cfa871f83f202765682ea492efd49f8ff21e98aa0794feb8ada5  moine-unidic-cwj-202512.tar.gz
dbbe7bdae52d6b9e7a68f321976ce19a935fcffdbbdbb4182287abd475d53fdd  moine-unidic-cwj-202512.tar.zst

Payload checksums inside metadata.yaml:

sha256-file-v1:      62b6fb5bc24e1a46be65e86e30c5a3ec23a4016f49e50bc3251698aa10525dec
sha256-canonical-v1: 21788ed133d29acffab7047264575ebd06a4549b712105e4b64e2600a29abcab

Maintainer verification:

cargo run -q -p moine-cli -- unidic-artifact-verify \
  --metadata moine-unidic-cwj-202512/metadata.yaml

License:

The generated artifact is derived from UniDic-CWJ and must carry the UniDic
license references separately from the moine code license. The archive
includes UniDic license references under license/BSD and license/COPYING.

moine-cedict-20260520-v0.1.1

moine-cedict-20260520-v0.1.1

Release of the moine CC-CEDICT no-tone pinyin artifact for moine v0.1.1.

The payload content is unchanged from the v0.1.0 artifact release. This release
keeps the same source dictionary version, artifact name, indexed payload format,
and bounded expansion settings under a v0.1.1 artifact tag.

Assets:

• moine-cedict-20260520.tar.gz
• moine-cedict-20260520.tar.zst

Source:

• Dictionary: CC-CEDICT
• Source version: 2026-05-20
• Pinyin view: no-tone
• Runtime payload: indexed FST/mmap-friendly normalized surface -> pinyin readings
• Artifact name: moine-cedict-20260520

Build options:

max_readings_per_surface: 16
max_readings_per_segment: 16
max_paths: 128
longest_match_only: true

Asset checksums:

75f2822f212928f4f4f2ddfbc1736f6b44f3c64f8d3f94dcdfbd53859a60d19b  moine-cedict-20260520.tar.gz
8ade18cbf74241123c38a09bedf1d4369d3150751641dd6688a61bdbb4d70887  moine-cedict-20260520.tar.zst

Payload checksums inside metadata.yaml:

sha256-file-v1:      f6087a4f47af64f39f668c76a36c668ca4a47b5b9ddc4b00d65e12bc121e76e9
sha256-canonical-v1: 31829b86c7fb9c80029b0638b434184041c37c749568251d0f1f8b570f8d361e

Maintainer verification:

cargo run -q -p moine-cli -- zh-artifact-verify \
  --metadata moine-cedict-20260520/metadata.yaml

License and attribution:

The generated artifact is derived from CC-CEDICT and must carry CC BY-SA 4.0
attribution and source metadata separately from the moine code license. Raw
CC-CEDICT glosses are not included in the runtime payload.

UniDic-CWJ 2025.12 Reading Index v0.1.0

UniDic-CWJ 2025.12 Reading Index v0.1.0

Updated on 2026-06-02 with a rebuilt indexed UniDic-CWJ reading artifact.

This rebuild adds halfwidth aliases for UniDic surfaces that are written with
fullwidth ASCII-compatible characters, such as entries like ＪＲ. The release
tag and asset names are unchanged.

Assets:

• moine-unidic-cwj-202512.tar.gz
• moine-unidic-cwj-202512.tar.zst

Source:

• Dictionary: UniDic-CWJ
• Source version: 2025.12
• Reading field: lform
• Runtime payload: indexed FST/mmap-friendly normalized surface -> readings
• Artifact name: moine-unidic-cwj-202512

Build options:

max_readings_per_surface: 16
max_readings_per_segment: 16
max_paths: 128
longest_match_only: true
exclude_ascii_surfaces: true
exclude_symbol_pos: true

The release is generated with:

scripts/release-unidic-cwj.sh \
  --lex-csv unidic-cwj-202512_full/lex.csv \
  --source-version 2025.12 \
  --artifact-name moine-unidic-cwj-202512 \
  --payload-format indexed \
  --dist-dir dist \
  --compression gzip

target/release/moine unidic-artifact-archive \
  --metadata dist/moine-unidic-cwj-202512/metadata.yaml \
  --output dist/moine-unidic-cwj-202512.tar.zst \
  --compression zstd

Maintainer verification:

cargo run -q -p moine-cli -- unidic-artifact-verify \
  --metadata moine-unidic-cwj-202512/metadata.yaml

Python usage after download/extraction:

from moine.ja import Dictionary

dictionary = Dictionary.load_bundle("moine-unidic-cwj-202512")
dictionary.distance("JR", "ジェーアール")

License:

The generated artifact is derived from UniDic-CWJ and must carry the UniDic
license references separately from the moine code license. The archive
includes UniDic license references under license/BSD and license/COPYING.

moine-cedict-20260520-v0.1.0

moine-cedict-20260520-v0.1.0

Updated on 2026-06-02 with regenerated release assets for the moine
CC-CEDICT no-tone pinyin artifact. The payload content is unchanged from the
previous v0.1.0 assets.

Assets:

• moine-cedict-20260520.tar.gz
• moine-cedict-20260520.tar.zst

Source:

• Dictionary: CC-CEDICT
• Source version: 2026-05-20
• Pinyin view: no-tone
• Runtime payload: indexed FST/mmap-friendly normalized surface -> pinyin readings
• Artifact name: moine-cedict-20260520

Build options:

max_readings_per_surface: 16
max_readings_per_segment: 16
max_paths: 128
longest_match_only: true

The release is generated with:

scripts/release-cedict.sh \
  --cedict cedict_1_0_ts_utf-8_mdbg.txt \
  --source-version 2026-05-20 \
  --artifact-name moine-cedict-20260520 \
  --payload-format indexed \
  --dist-dir dist \
  --compression gzip

target/release/moine zh-artifact-archive \
  --metadata dist/moine-cedict-20260520/metadata.yaml \
  --output dist/moine-cedict-20260520.tar.zst \
  --compression zstd

Maintainer verification:

cargo run -q -p moine-cli -- zh-artifact-verify \
  --metadata moine-cedict-20260520/metadata.yaml

Python usage after download/extraction:

import moine

moine.distance("weishiji", "威士忌", lang="zh")
moine.distance("布納哈奔", "布納哈本", lang="zh")

License and attribution:

The generated artifact is derived from CC-CEDICT and must carry CC BY-SA 4.0
attribution and source metadata separately from the moine code license. Raw
CC-CEDICT glosses are not included in the runtime payload.