Mgitau/original-language-bibles-greek

该数据集编译自STEP Bible（由廷代尔之家的圣经学者制作），采用CC BY 4.0许可。它包含希腊新约的高质量注释文本，并附带词级翻译。数据集中的每一行包含以下列：`reference`（单词的唯一位置，格式为<书>.<章>.<节>.<词号>）、`text`（希腊语文本）、`transliteration`（英语音译）、`translation`（词级翻译）、`dStrongs`（Strongs编号，根据STEP文档进行消歧，并包含语法解析）和`manuscript_source`（记录该单词的抄本来源）。数据集还总结了单词类型及其在不同抄本中的使用情况，例如：NK(O)类型包含133,608个在几乎所有抄本中相同的单词，占总数94%；其他类型如N(K)(O)、K(O)、N(O)和O分别表示在传统抄本、古代抄本或其他抄本中的差异单词，这些差异可能影响翻译。该数据集的创建动机是促进圣经NLP研究，解决现有翻译数据（如KJV）的局限性，并提供高质量的原始语言数据，以支持作者归属、问答和主题建模等任务。This data is compiled from STEP Bible (https://www.stepbible.org/), produced by Biblical scholars at Tyndale House. Usage under CC BY 4.0. This dataset contains high-quality, annotated text of the Greek New Testament with accompanying word-level translations. Each line in the dataset contains the columns: `reference`: the unique location of the word as `<book>.<chapter>.<verse>.<word_number>`; `text`: the Koine Greek text; `transliteration`: the English transliteration; `translation`: word-level translation; `dStrongs`: Strongs number, disambiguated as described at https://www.tinyURL.com/STEP-Disambiguated as well as grammatical parsing; `manuscript_source`: which manuscript(s) record the word. A key to the options, from STEPs documentation, follows. Summary of words included in the Amalgamated Greek NT: Types such as NK(O) (133,608 words identical in virtually all manuscripts), N(K)(O) (3,084 words different in Traditional manuscripts), K(O) (4,164 words found in Traditional but not Ancient manuscripts), N(O) (896 words found in Ancient but not Traditional manuscripts), and O (284 words in manuscripts different from Ancient and Traditional), with details on how these affect translations. The motivation for this dataset is to facilitate NLP tasks on Biblical text by addressing issues like reliance on translations (e.g., KJV), incomplete digitized corpora, and lack of word-level annotations for manuscript attestation, encouraging research in Biblical NLP.