Hugging Face's logo

Datasets:
mschonhardt
/
bdd-abbreviations-augmented

Tasks:
Translation
Modalities:
Text
Formats:
csv
Sub-tasks:
text2text-generation
Languages:
Latin
Size:
100K - 1M
Libraries:
Datasets
pandas
License:
Dataset card Data Studio Files
xet
Community
1
mschonhardt commited on
Commit
0e60245
·
verified ·
1 Parent(s): 26f408e

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +166 -0
README.md ADDED
@@ -0,0 +1,166 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language: la
3
+ license: cc-by-sa-4.0
4
+ multilinguality: monolingual
5
+ size_categories:
6
+ - 10K<n<100K
7
+ source_datasets:
8
+ - original
9
+ task_categories:
10
+ - translation
11
+ task_ids:
12
+ - text2text-generation
13
+ pretty_name: "Burchard's Decretum Dataset Abbreviations Augmented"
14
+ annotations_creators:
15
+ - expert-generated
16
+ ---
17
+
18
+ # Dataset Card for bdd-abbreviations-augmented
19
+
20
+ ## Dataset Description
21
+
22
+ ### Dataset Summary
23
+
24
+ - **Burchards Dekret Digital (BDD):** [Website](https://www.burchards-dekret-digital.de/)
25
+ - **Source data**: [BDD Gitlab](https://gitlab.rlp.net/adwmainz/projekte/burchards-dekret-digital/data)
26
+ - **Trainingscripts:** [Abbreviationes GitHub](https://github.com/michaelscho/Abbreviationes)
27
+ - **Zenodo:** [Zenodo](https://doi.org/10.5281/zenodo.16735498)
28
+ - **Contact:** Michael Schonhardt ([email protected], [ORCID](https://orcid.org/0000-0002-2750-1900))
29
+
30
+ The **BDD Abbreviations Augmented** dataset is a parallel corpus for **automated abbreviation expansion**
31
+ in transcriptions from 11th-century Latin manuscripts. It provides paired examples of abbreviated text (source)
32
+ and its corresponding full expansion (target).
33
+
34
+ The dataset was generated from the five key manuscripts of the *Decretum Burchardi*, a major work of
35
+ medieval canon law, transcribed by the Burchards Dekret Digital (BDD) project.
36
+ It is specifically designed for fine-tuning byte-level sequence-to-sequence models like **ByT5**
37
+ to assist in the creation of digital critical editions.
38
+
39
+ The core task is to translate a "graphemic" transcription, which preserves scribal abbreviations
40
+ using special Unicode characters (MUFI), into a fully readable, expanded Latin text.
41
+
42
+ ### Supported Tasks
43
+
44
+ - **`abbreviation-expansion`**: The dataset is intended for training models on a sequence-to-sequence basis. The model takes the `source_text` as input and generate the `target_text`.
45
+
46
+ ### Languages
47
+
48
+ The text is in Medieval Latin (`la`).
49
+
50
+ ## Dataset Structure
51
+
52
+ ### Data Instances
53
+
54
+ A typical data instance consists of a source string with abbreviations and a target string with the expansions.
55
+
56
+ ```json
57
+ {
58
+ "source_text": "om̅s posteri eorū cuncta sibi uendicarent sed semꝑ maiores causę sicut s̅ ep̅oꝝ et potiorū curę negotiorū ad unā beati principis apꝉoꝝ petri sedē confluerent ut inde",
59
+ "target_text": "omnes posteri eorum cuncta sibi uendicarent sed semper maiores causę sicut sunt episcoporum et potiorum curę negotiorum ad unam beati principis apostolorum petri sedem confluerent ut inde"
60
+ }
61
+ ```
62
+
63
+ ### Data Fields
64
+
65
+ **source_text**: A string containing the abbreviated Latin text from the manuscript, represented using standard and special Unicode characters.
66
+
67
+ **target_text**: A string containing the corresponding fully expanded Latin text.
68
+
69
+ ### Data Splits
70
+
71
+ The dataset contains a single train split.
72
+
73
+ ## Source Data
74
+
75
+ The data originates from the final, manually corrected TEI-XML transcriptions of five key 11th-century manuscripts of Burchard's Decree, produced in the scriptorium of Worms:
76
+
77
+ * B: Bamberg, Staatsbibliothek, Msc. Can. 6
78
+
79
+ * F: Frankfurt, Universitätsbibliothek, Ms. Barth. 50
80
+
81
+ * K: Köln, Erzbischöfliche Diözesan- und Dombibliothek, Cod. 119
82
+
83
+ * V2: Vatican, Biblioteca Apostolica Vaticana, Pal. lat. 585
84
+
85
+ * V3: Vatican, Biblioteca Apostolica Vaticana, Pal. lat. 586
86
+
87
+ ## Annotations
88
+
89
+ The dataset was generated automatically from tei [source files](https://gitlab.rlp.net/adwmainz/projekte/burchards-dekret-digital/data). The process was as follows:
90
+
91
+ **Annotation**: Editors from the BDD project transcribed and expanded all abbreviations in the source manuscripts, encoding the results in TEI-XML using the <choice> element with <abbr> (abbreviation) and <expan> (expansion) children.
92
+
93
+ **Pair Generation**: A [script](https://github.com/michaelscho/Abbreviationes) processed the final TEI files. It used a 3-line sliding window to provide context. For each window, it extracted the text from the <abbr> tags to create the source_text and the text from the <expan> tags to create the target_text.
94
+ Rare abbreviations were automatically multiplied to represent them well in the final model.
95
+
96
+ ## Considerations for Using the Data
97
+
98
+ The dataset is highly specialized. It reflects the scribal practices of a single scriptorium (Worms) from a specific time period (early 11th century). The system of abbreviations is consistent within the corpus but may not be representative of other medieval scripts, regions, or time periods. A model trained on this data will perform best on similar manuscripts and will likely require further fine-tuning to be used on texts from other traditions. As the dataset is augmented to better represent rare brevigraphs, the trained model might fail in instances where distribution of brevigraphs differs.
99
+
100
+ The model trained on this dataset is designed to expand a fixed set of brevigraphs and special characters recommended by the Medieval Unicode Font Initiative (MUFI). It will not be able to expand abbreviations that are not represented in the training data. The main characters are listed below:
101
+
102
+
103
+ | Glyph | Codepoint | MUFI Name | MUFI link |
104
+ | ----- | --------- | --------- | --------- |
105
+ | ƀ | \u0180 | LATIN SMALL LETTER B WITH STROKE | https://mufi.info/q.php?p=mufi/chars/unichar/384 |
106
+ | đ | \u0111 | LATIN SMALL LETTER D WITH STROKE | https://mufi.info/q.php?p=mufi/chars/unichar/273 |
107
+ | Ę | \u0118 | LATIN CAPITAL LETTER E WITH OGONEK | https://mufi.info/q.php?p=mufi/chars/unichar/280 |
108
+ | ę | \u0119 | LATIN SMALL LETTER E WITH OGONEK | https://mufi.info/q.php?p=mufi/chars/unichar/281 |
109
+ | ħ | \u0127 | LATIN SMALL LETTER H WITH STROKE | https://mufi.info/q.php?p=mufi/chars/unichar/295 |
110
+ | Ꝁ | \ua740 | LATIN CAPITAL LETTER K WITH STROKE | https://mufi.info/q.php?p=mufi/chars/unichar/42816 |
111
+ | ꝁ | \ua741 | LATIN SMALL LETTER K WITH STROKE | https://mufi.info/q.php?p=mufi/chars/unichar/42817 |
112
+ | Ꝉ | \ua748 | LATIN CAPITAL LETTER L WITH HIGH STROKE | https://mufi.info/q.php?p=mufi/chars/unichar/42824 |
113
+ | ꝉ | \ua749 | LATIN SMALL LETTER L WITH HIGH STROKE | https://mufi.info/q.php?p=mufi/chars/unichar/42825 |
114
+ | P̄ [Per] | \u0050+\u0304 | - | - |
115
+ | p̄ [per] | \u0070+\u0304 | - | - |
116
+ | Ꝑ | \ua750 | LATIN CAPITAL LETTER P WITH STROKE THROUGH DESCENDER | https://mufi.info/q.php?p=mufi/chars/unichar/42832 |
117
+ | ꝑ | \ua751 | LATIN SMALL LETTER P WITH STROKE THROUGH DESCENDER | https://mufi.info/q.php?p=mufi/chars/unichar/42833 |
118
+ | Ꝓ | \ua752 | LATIN CAPITAL LETTER P WITH FLOURISH | https://mufi.info/q.php?p=mufi/chars/unichar/42834 |
119
+ | ꝓ | \ua753 | LATIN SMALL LETTER P WITH FLOURISH | https://mufi.info/q.php?p=mufi/chars/unichar/42835 |
120
+ | Ꝗ | \ua756 | LATIN CAPITAL LETTER Q WITH STROKE THROUGH DESCENDER | https://mufi.info/q.php?p=mufi/chars/unichar/42838 |
121
+ | ꝗ | \ua757 | LATIN SMALL LETTER Q WITH STROKE THROUGH DESCENDER | https://mufi.info/q.php?p=mufi/chars/unichar/42839 |
122
+ | Ꝙ | \ua758 | LATIN CAPITAL LETTER Q WITH DIAGONAL STROKE | https://mufi.info/q.php?p=mufi/chars/unichar/42840 |
123
+ | ꝙ | \ua759 | LATIN SMALL LETTER Q WITH DIAGONAL STROKE | https://mufi.info/q.php?p=mufi/chars/unichar/42841 |
124
+ |  [quia] | \ue8b3 | LATIN SMALL LETTER Q LIGATED WITH R ROTUNDA | https://mufi.info/q.php?p=mufi/chars/unichar/59571 |
125
+ | x̄ [Nasalstrich] | \u0304 | COMBINING MACRON | https://mufi.info/q.php?p=mufi/chars/unichar/772 |
126
+ | x̅ [Kürzungszeichen] | \u0305 | COMBINING OVERLINE | https://mufi.info/q.php?p=mufi/chars/unichar/773 |
127
+ | ᷒ [-us] | \u1dd2 | COMBINING US ABOVE | https://mufi.info/q.php?p=mufi/chars/unichar/7634 |
128
+ | ᷓ [^a] | \u1dd3 | COMBINING LATIN SMALL LETTER FLATTENED OPEN A ABOVE | https://mufi.info/q.php?p=mufi/chars/unichar/7635 |
129
+ | ◌ [-ur] | \uf1c2 | COMBINING ABBREVIATION MARK SUPERSCRIPT UR LEMNISKATE FORM | https://mufi.info/q.php?p=mufi/chars/unichar/1048860 |
130
+ | ◌ͥ [^i] | \u0365 | COMBINING LATIN SMALL LETTER I | https://mufi.info/q.php?p=mufi/chars/unichar/869 |
131
+ | ◌ͦ [^o] | \u0366 | COMBINING LATIN SMALL LETTER O | https://mufi.info/q.php?p=mufi/chars/unichar/870 |
132
+ | ◌ͧ [^u] | \u0367| COMBINING LATIN SMALL LETTER U | https://mufi.info/q.php?p=mufi/chars/unichar/871 |
133
+ | Ꝝ | \ua75c | LATIN CAPITAL LETTER RUM ROTUNDA | https://mufi.info/q.php?p=mufi/chars/unichar/42844 |
134
+ | ꝝ | \ua75d | LATIN SMALL LETTER RUM ROTUNDA | https://mufi.info/q.php?p=mufi/chars/unichar/42845 |
135
+ |  | \uf1f8 | DISTINCTIO |https://mufi.info/q.php?p=mufi/chars/unichar/61944|
136
+ |  | \uf1ea | PUNCTUS VERSUS |https://mufi.info/q.php?p=mufi/chars/unichar/61930|
137
+ |  | \uf1f0 | PUNCTUS ELEVATUS DIAGONAL STROKE |https://mufi.info/q.php?p=mufi/chars/unichar/61936|
138
+ |  | \uf160 | PUNCTUS INTERROGATIVUS |https://mufi.info/q.php?p=mufi/chars/unichar/61792|
139
+ |  | \uf1f5 | PUNCTUS FLEXUS |https://mufi.info/q.php?p=mufi/chars/unichar/61941|
140
+ |  | \uf160 + \uf1e2 | PUNCTUS INTERROGATIVUS |https://mufi.info/q.php?p=mufi/chars/unichar/61792|
141
+ |  | \uf1aC | LATIN ABBREVIATION SIGN SEMICOLON |https://mufi.info/q.php?p=mufi/chars/unichar/61868|
142
+ |  | \uf1e1 | PARAGRAPHUS |https://mufi.info/q.php?p=mufi/chars/unichar/61921|
143
+
144
+
145
+ ## Dataset Curators
146
+
147
+ The dataset was generated by Michael Schonhardt (TU Darmstadt / Akademie der Wissenschaften und der Literatur | Mainz), using transcriptions prepared by the Burchards Dekret Digital Project Team. Individual contributions are encoded in [TEI files](https://gitlab.rlp.net/adwmainz/projekte/burchards-dekret-digital/data).
148
+
149
+ ## Licensing Information
150
+
151
+ Creative Commons Attribution-ShareAlike 4.0 International ([CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/deed.en)).
152
+
153
+ ## Citation Information
154
+
155
+ If you use this dataset in your research, please cite the original project and the dataset.
156
+
157
+ ```bibtex
158
+ @misc{schonhardt_bdd_abbreviations_2025,
159
+ author = {Schonhardt, Michael},
160
+ title = {BDD Abbreviations Augmented: A Parallel Corpus for Abbreviation Expansion in Medieval Latin},
161
+ year = {2025},
162
+ institution = {Burchards Dekret Digital},
163
+ doi={https://doi.org/10.5281/zenodo.16735498},
164
+ howpublished = {\url{[https://huggingface.co/datasets/mschonhardt/bdd-abbreviations-augmented](https://huggingface.co/datasets/mschonhardt/bdd-abbreviations-augmented)}}
165
+ }
166
+ ```