de-francophones
commited on
Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- fr
|
4 |
+
viewer: false
|
5 |
+
---
|
6 |
+
|
7 |
+
> [!NOTE]
|
8 |
+
> Dataset origin: https://zenodo.org/records/3383658
|
9 |
+
|
10 |
+
# Automatic TEI encoding of manuscripts catalogues with GROBID-Dictionaries
|
11 |
+
|
12 |
+
## Credits
|
13 |
+
|
14 |
+
GROBID dictionaries is developed by Mohamed Khemakhem ([GitHub](https://github.com/MedKhem)).
|
15 |
+
|
16 |
+
More info on GROBID technologies can be found [here](https://grobid.readthedocs.io).
|
17 |
+
|
18 |
+
## Corpus
|
19 |
+
|
20 |
+
The corpus is based on scans of catalogues published by Auguste Laverdet, the *Revue des autographes*, directed by Gabriel Charavay ([data.bnf](http://data.bnf.fr/10429866/gabriel_charavay/)) and different auction sales catalogues published by the Maison Charavay.
|
21 |
+
|
22 |
+
## Methodology
|
23 |
+
|
24 |
+
PDF are OCRised with [Transkribus](https://transkribus.eu).
|
25 |
+
|
26 |
+
The GROBID dictionaries model is trained on four different types of catalogues. For each type, 15 annotated pages are available : 10 as training data and 5 as evaluation data.
|
27 |
+
|
28 |
+
The training data are extracted from the following catalogs issues:
|
29 |
+
+ Gabriel Charavay, _Revue des Autographes_, first series : 25, 35, 50, 80.
|
30 |
+
+ Gabrielle Charavay _Revue des Autographes_, second series : 24, 56.
|
31 |
+
+ Auguste Laverdet, _Catalogue de lettres autographes et manuscrits_ : 1, 22.
|
32 |
+
+ Etienne Charavay, _Catalogue d’une intéressante collection de lettres autographes…_ (14 décembre 1908).
|
33 |
+
|
34 |
+
The source of the training files are references in the file names using the following acronyms:
|
35 |
+
+ RDA_1 for Revue des Autographes, first series
|
36 |
+
+ RDA_2 for Revue des Autographes, second series
|
37 |
+
+ LAV for catalogues written by Auguste Laverdet
|
38 |
+
+ ETI for auction catalogues written by Etienne Charavay
|
39 |
+
|
40 |
+
In compliance with the
|
41 |
+
PDF are OCRised with [recommendations of Grobid⁻Dictionaries](https://github.com/MedKhem/grobid-dictionaries/wiki/How-to-Annotate%3F), the separators, whether there are composed of one or two punctuation signs, are tagged with the neutral label <pc>.
|
42 |
+
|
43 |
+
Examples :
|
44 |
+
|
45 |
+
<entry>
|
46 |
+
<num>350</num>
|
47 |
+
<pc>.</pc>
|
48 |
+
<form>
|
49 |
+
<name>STANISLAS</name>
|
50 |
+
<pc>,</pc>
|
51 |
+
<desc>roi de Pologne N. 1677. M. 1766</desc>
|
52 |
+
</form>
|
53 |
+
<pc>. -</pc>
|
54 |
+
<sense>
|
55 |
+
<subSense>L. aut. sig. à madame de Boufïlers. Lunéville, 5 janvier 1741. 1 page in-4, 10</subSense>
|
56 |
+
</sense>
|
57 |
+
</entry>
|
58 |
+
<pc>»</pc>
|
59 |
+
<entry>
|
60 |
+
|
61 |
+
## Files
|
62 |
+
Training data and evaluation data are available in `ToyData`
|
63 |
+
|
64 |
+
## Licence
|
65 |
+
|
66 |
+
Regarding `GROBID`, cf. [here](https://github.com/MedKhem/grobid-dictionaries).
|
67 |
+
|
68 |
+
Regarding the corpus: extracted data is CC-BY.
|
69 |
+
|
70 |
+
<a rel="license" href="https://creativecommons.org/licenses/by/2.0"><img alt="Creative Commons License" style="border-width:0" src="https://i.creativecommons.org/l/by/2.0/88x31.png" /></a><br />
|