open-gram 词典与语言模型数据
双拼内置词典改编自与 SunPinyin 相关的 open-gram 项目所发布的资源。双拼未嵌入 SunPinyin 输入法引擎。
词库来自 open-gram 贡献者，并包含 Android Pinyin IME 及其他贡献的词条。语言模型的单字词频提取自 lm_sc.t3g.arpa-20121025；SunPinyin 的历史软件包元数据将该模型归功于 Kefu Chai。
open-gram 的归档项目说明明确指出，其收集的中文词库和 n-gram 统计数据均依据 Apache License 2.0 提供。归档项目的许可证亦标记为 Apache-2.0，且 2012 年模型下载未指定其他文件级许可条款。
双拼所作修改
修改包括删除模型元数据、未知项和标点项；提取并以整数缩放 unigram 概率；将全拼读音转换为小鹤双拼；以及将结果合并为 SQLite/SQLCipher 格式。
来源与许可
FMDB 2.7.12
Copyright © 2008–2014 Flying Meat Inc.
双拼使用 FMDB 访问本地数据库。FMDB 依据 MIT License 分发。
SQLCipher Community Edition 4.17.0
Copyright © 2025, ZETETIC LLC. All rights reserved.
双拼使用 SQLCipher Community Edition 加密本地词典数据库。SQLCipher Community Edition 依据 BSD-style license 分发。
完整声明
Open-source acknowledgements
open-gram dictionary and language-model data
ShuangPin’s bundled dictionary adapts resources published by open-gram, a data project associated with SunPinyin. ShuangPin does not embed the SunPinyin input-method engine.
The lexicon comes from open-gram contributors and includes entries from Android Pinyin IME and other contributions. Unigram statistics were extracted from lm_sc.t3g.arpa-20121025; historical SunPinyin packaging metadata credits the model to Kefu Chai.
The archived open-gram project description explicitly makes its collected Chinese lexicon and n-gram statistics available under the Apache License 2.0. ShuangPin’s changes include removing model metadata and unknown/punctuation records, extracting and integer-scaling unigram probabilities, converting QuanPin readings to Xiaohe ShuangPin, and combining the data in SQLite/SQLCipher format.
Sources and complete Apache-2.0 license
FMDB 2.7.12
Copyright © 2008–2014 Flying Meat Inc. FMDB is distributed under the MIT License. Read the complete notice.
SQLCipher Community Edition 4.17.0
Copyright © 2025, ZETETIC LLC. All rights reserved. SQLCipher Community Edition is distributed under a BSD-style license. Read the complete notice.
Download all acknowledgements and license terms as plain text.