Old Italian Model

This repository contains the Stanza model for Old Italian, based on the data from the UD treebank Italian-Old (version 2.14). The model is built on parts of two Cantiche of the Divine Comedy of Dante Alighieri, Inferno and Purgatorio, an Old Italian (Florentine) poem composed between approximately 1306 and 1321.

Train data

The training data is composed as follows:

train: 1 748 sentences and 57 898 tokens

Inferno: 902 sentences (OldItalian_Dante_Inferno-1; OldItalian_Dante_Inferno-902) - 28 995 tokens

Purgatorio: 846 sentences (OldItalian_Dante_Purgatorio-1; OldItalian_Dante_Purgatorio-846) - 28 903 tokens

dev: 316 sentences and 12 439 tokens

Inferno: 151 sentences (OldItalian_Dante_Inferno-903; OldItalian_Dante_Inferno-1053) - 6 232 tokens

Purgatorio: 165 sentences (OldItalian_Dante_Purgatorio-847; OldItalian_Dante_Purgatorio-1011) - 6 207 tokens

How to use

Clone this repository:

https://github.com/CIRCSE/Old_Italian_Model.git
cd Old_Italian_Model

Create and activate a python virtual environment (virtualenv, conda, minicoda, etc). Here an example with virtualenv:

virtualenv -p python3 venv
source venv/bin/activate

or if you installed virtualenv as a module:

python -m virtualenv -p python3 venv
source venv/bin/activate

install the Stanza module with:

pip install stanza

To perform a small test, launch the command :

python processText.py ./OldItalianModel "Vuolsi così colà dove si puote ciò che si vuole, e più non dimandare." old_italian_out.conllu

this commands will produce an annotated file (old_italian_out) with tok,lemma,pos,parser. If everything went correctly the conllu file should contain something similar to this:

# text = Vuolsi così colà dove si puote ciò che si vuole, e più non dimandare.
# sent_id = 0
1-2     Vuolsi  _       _       _       _       _       _       _       start_char=0|end_char=6
1       Vuole   volere  VERB    vta*2ips3       Mood=Ind|Number=Sing|Person=3|Tense=Pres|VerbForm=Fin|Voice=Act 0       root    _       _
2       si      si      PRON    pf3xim  Clitic=Yes|Person=3|PronType=Prs|Reflex=Yes     1       expl:impers     _       _
3       così    così    ADV     b       _       1       advmod  _       start_char=7|end_char=11
4       colà    colà    ADV     b       _       1       advmod:lmod     _       start_char=12|end_char=16
5       dove    dove    ADV     b       _       7       advmod:lmod     _       start_char=17|end_char=21
6       si      si      PRON    pf3ypa  Clitic=Yes|Person=3|PronType=Prs|Reflex=Yes     7       expl:impers     _       start_char=22|end_char=24
7       puote   potere  VERB    vtp2ips3        Mood=Ind|Number=Sing|Person=3|Tense=Pres|VerbForm=Fin|Voice=Pass        4       acl:relcl       _       start_char=25|end_char=30
8       ciò     cio'    PRON    pd      PronType=Dem    7       obj     _       start_char=31|end_char=34
9       che     che     PRON    pr      PronType=Rel    11      nsubj   _       start_char=35|end_char=38
10      si      si      PRON    pf3yia  Clitic=Yes|Person=3|PronType=Prs|Reflex=Yes     11      expl:impers     _       start_char=39|end_char=41
11      vuole   volere  VERB    vta*2ips3       Mood=Ind|Number=Sing|Person=3|Tense=Pres|VerbForm=Fin|Voice=Act 8       acl:relcl       _       start_char=42|end_char=47|SpaceAfter=No
12      ,       ,       PUNCT   _       _       16      punct   _       start_char=47|end_char=48
13      e       e       CCONJ   cscc    _       16      cc      _       start_char=49|end_char=50
14      più     più     ADV     b       _       16      advmod  _       start_char=51|end_char=54
15      non     non     ADV     b       PronType=Neg    16      advmod:neg      _       start_char=55|end_char=58
16      dimandare       domandare       VERB    vta1mps2        Mood=Imp|Number=Sing|Person=2|Tense=Pres|VerbForm=Fin|Voice=Act 1       conj    _       start_char=59|end_char=68|SpaceAfter=No
17      .       .       PUNCT   _       _       1       punct   _       start_char=68|end_char=69|SpaceAfter=No

Similarly, if the file to be annotated is in conllu format, use:

python processConllu.py  ./OldItalianModel <conlluFile.conllu> old_italian_out.conllu

Contact

For further details, please contact: giovanni.moretti@unicatt.it, claudia.corbetta@unibg.it