Datasets:

slone
/

finugorbib

@@ -52,8 +52,9 @@ configs:
 <!-- Provide a quick summary of the dataset. -->
-This a text of the Bible, translated into 12 Finno-Ugric languages (see below), and aligned with Russian, English and Finnish.
 This dataset was downloaded from http://finugorbib.com.
 ## Dataset Details
@@ -117,70 +118,40 @@ The dataset has been downloaded from http://finugorbib.com/.
 Its constituent books and the corresponding copyritght notices are listed in the file `LICENSE.md`
 ## Uses
 The dataset is intended (apart from its original religious purposes) to help humans and machines learn the low-resourced Finno-Ugric languages.
 ## Dataset Structure
-TBD
-## Dataset Creation
-### Curation Rationale
-<!-- Motivation for the creation of this dataset. -->
-[More Information Needed]
-### Source Data
-<!-- This section describes the source data (e.g. news text and headlines, social media posts, translated sentences, ...). -->
-#### Data Collection and Processing
-<!-- This section describes the data collection and processing process such as data selection criteria, filtering and normalization methods, tools and libraries used, etc. -->
-[More Information Needed]
-#### Who are the source data producers?
-<!-- This section describes the people or systems who originally created the data. It should also include self-reported demographic or identity information for the source data creators if this information is available. -->
-[More Information Needed]
-### Annotations [optional]
-<!-- If the dataset contains annotations which are not part of the initial data collection, use this section to describe them. -->
-#### Annotation process
-<!-- This section describes the annotation process such as annotation tools used in the process, the amount of data annotated, annotation guidelines provided to the annotators, interannotator statistics, annotation validation, etc. -->
-[More Information Needed]
-#### Who are the annotators?
-<!-- This section describes the people or systems who created the annotations. -->
-[More Information Needed]
-#### Personal and Sensitive Information
-<!-- State whether the dataset contains data that might be considered personal, sensitive, or private (e.g., data that reveals addresses, uniquely identifiable names or aliases, racial or ethnic origins, sexual orientations, religious beliefs, political opinions, financial or health data, etc.). If efforts were made to anonymize the data, describe the anonymization process. -->
-[More Information Needed]
-## Bias, Risks, and Limitations
-<!-- This section is meant to convey both technical and sociotechnical limitations. -->
-[More Information Needed]
-### Recommendations
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
-Users should be made aware of the risks, biases and limitations of the dataset. More information needed for further recommendations.

 <!-- Provide a quick summary of the dataset. -->
+This a text of the Bible, translated into 12 Finno-Ugric languages spoken in Russian Federation (see below), and aligned with Russian, English and Finnish.
 This dataset was downloaded from http://finugorbib.com.
+All rights belong to the Institute for Bible translation Helsinki, see the `LICENSE.md` file.
 ## Dataset Details
 Its constituent books and the corresponding copyritght notices are listed in the file `LICENSE.md`
 ## Uses
 The dataset is intended (apart from its original religious purposes) to help humans and machines learn the low-resourced Finno-Ugric languages.
 ## Dataset Structure
+Note that the Bible text consists of the Old and New Testaments. Each consists of multiple books (denoted by 3-letter codes, e.g. `Php`).
+Each book consists of one or more chapters, and each chapter consists of verses (usually corresponding to sentences).
+The texts are aligned across languages verse-by-verse, but these alignments are not always exact, so additional data filtering is recommended.
+The dataset is separated in two subsets. By default, both are loaded:
+1. `all_to_rus`: Old and New Testament, paired with Russian (reachable from http://finugorbib.com/index.html).
+2. `nt_to_rus_fin_eng`: New Testament, paired with English, Finnish and Russian (reachable from http://finugorbib.com/alt/alt_al.html).
+Both subsets are split into `test`, `validation` and `train` splits:
+- `test` is composed of the ultimate chapter of each book (and the chapters 15 and 16 of the Gospel by Mark)
+- `dev` is composed of the penultimate chapter of each book (and the chapters 13 and 14 of the Gospel by Mark)
+- `train` is composed of the rest of the chapters.
+This specific structure was imposed to ensure that each verse belongs to a single split, regardless of the language.
+Each data row has the following fields:
+- `text: str`: text of the current verse in the lower-resourced Finno-Ugric language
+- `other: str`: text of the current verse in another langage (English, Finnish or Russian)
+- `lang: str`: name (from the website) of the lower-resourced language
+- `lang_code: str`: language code (ISO 639-3) of the lower-resourced language
+- `other_lang: str`: language code (ISO 639-3) of the paired language (`eng`, `fin` or `rus`)
+- `verse_id: Optional[int]`: id of the verse, numeric (if parsed correctly, otherwise `null`)
+- `book_chapter: str`: id of the book and chapter (e.g. `57_Php02`)
+- `book_chapter_verse: Optional[str]`: id of the book, chapter and verse (e.g. `57_Php02_14`), string (if parsed correctly, otherwise `null`)
+## Dataset Creation
+The dataset was curates as a source of machine translation training and evaluation data.
+It consists solely of the Bible texts. They have been downloaded from the http://finugorbib.com website (using the `wget` tool)
+and parsed from `html` files into aligned text pairs (using the `beautifulsoup` package).
+No additional filtering or postprocessing has been applied.