COMMA / README.md
LorMolf's picture
Upload dataset
809308d verified
metadata
dataset_info:
  - config_name: constitution
    features:
      - name: article_n
        dtype: int64
      - name: article_commas
        struct:
          - name: '1'
            dtype: string
          - name: '10'
            dtype: string
          - name: '11'
            dtype: string
          - name: '12'
            dtype: string
          - name: '2'
            dtype: string
          - name: '3'
            dtype: string
          - name: '4'
            dtype: string
          - name: '5'
            dtype: string
          - name: '6'
            dtype: string
          - name: '7'
            dtype: string
          - name: '8'
            dtype: string
          - name: '9'
            dtype: string
    splits:
      - name: it
        num_bytes: 70944
        num_examples: 139
      - name: en
        num_bytes: 69749
        num_examples: 139
      - name: fr
        num_bytes: 77000
        num_examples: 139
      - name: es
        num_bytes: 76072
        num_examples: 139
    download_size: 225640
    dataset_size: 293765
  - config_name: en
    features:
      - name: id
        dtype: string
      - name: ruling_type
        dtype: int64
      - name: epigraph
        dtype: string
      - name: body
        dtype: string
      - name: decision
        dtype: string
      - name: maxims_text
        dtype: string
      - name: maxims_title
        dtype: string
      - name: full_text
        dtype: string
      - name: num_maxims
        dtype: int64
      - name: maxims_len
        dtype: int64
      - name: full_text_len
        dtype: int64
      - name: judgment_type
        dtype: int64
      - name: constitutional_parameters
        dtype: string
      - name: maxims
        dtype: string
    splits:
      - name: train
        num_bytes: 555145830
        num_examples: 12600
      - name: test
        num_bytes: 30737608
        num_examples: 700
      - name: validation
        num_bytes: 31671019
        num_examples: 700
    download_size: 278441383
    dataset_size: 617554457
  - config_name: es
    features:
      - name: id
        dtype: string
      - name: ruling_type
        dtype: int64
      - name: epigraph
        dtype: string
      - name: body
        dtype: string
      - name: decision
        dtype: string
      - name: maxims_text
        dtype: string
      - name: maxims_title
        dtype: string
      - name: full_text
        dtype: string
      - name: num_maxims
        dtype: int64
      - name: maxims_len
        dtype: int64
      - name: full_text_len
        dtype: int64
      - name: judgment_type
        dtype: int64
      - name: constitutional_parameters
        dtype: string
      - name: maxims
        dtype: string
    splits:
      - name: train
        num_bytes: 575679719
        num_examples: 12600
      - name: test
        num_bytes: 31896832
        num_examples: 700
      - name: validation
        num_bytes: 32827830
        num_examples: 700
    download_size: 300803577
    dataset_size: 640404381
  - config_name: fr
    features:
      - name: id
        dtype: string
      - name: ruling_type
        dtype: int64
      - name: epigraph
        dtype: string
      - name: body
        dtype: string
      - name: decision
        dtype: string
      - name: maxims_text
        dtype: string
      - name: maxims_title
        dtype: string
      - name: full_text
        dtype: string
      - name: num_maxims
        dtype: int64
      - name: maxims_len
        dtype: int64
      - name: full_text_len
        dtype: int64
      - name: judgment_type
        dtype: int64
      - name: constitutional_parameters
        dtype: string
      - name: maxims
        dtype: string
    splits:
      - name: train
        num_bytes: 580985816
        num_examples: 12600
      - name: test
        num_bytes: 32177379
        num_examples: 700
      - name: validation
        num_bytes: 33152939
        num_examples: 700
    download_size: 306338176
    dataset_size: 646316134
  - config_name: it
    features:
      - name: id
        dtype: string
      - name: ruling_type
        dtype: int64
      - name: epigraph
        dtype: string
      - name: body
        dtype: string
      - name: decision
        dtype: string
      - name: maxims_text
        dtype: string
      - name: maxims_title
        dtype: string
      - name: full_text
        dtype: string
      - name: num_maxims
        dtype: int64
      - name: maxims_len
        dtype: int64
      - name: full_text_len
        dtype: int64
      - name: judgment_type
        dtype: int64
      - name: constitutional_parameters
        dtype: string
      - name: maxims
        dtype: string
    splits:
      - name: train
        num_bytes: 557553146
        num_examples: 12600
      - name: test
        num_bytes: 30850184
        num_examples: 700
      - name: validation
        num_bytes: 31775341
        num_examples: 700
    download_size: 293523614
    dataset_size: 620178671
configs:
  - config_name: constitution
    data_files:
      - split: it
        path: constitution/it-*
      - split: en
        path: constitution/en-*
      - split: fr
        path: constitution/fr-*
      - split: es
        path: constitution/es-*
  - config_name: en
    data_files:
      - split: train
        path: en/train-*
      - split: test
        path: en/test-*
      - split: validation
        path: en/validation-*
  - config_name: es
    data_files:
      - split: train
        path: es/train-*
      - split: test
        path: es/test-*
      - split: validation
        path: es/validation-*
  - config_name: fr
    data_files:
      - split: train
        path: fr/train-*
      - split: test
        path: fr/test-*
      - split: validation
        path: fr/validation-*
  - config_name: it
    data_files:
      - split: train
        path: it/train-*
      - split: test
        path: it/test-*
      - split: validation
        path: it/validation-*

Dataset Summary

COMMA is a constitutional multi-task and multi-lingual archive consisting of 14K CCIR rulings with expert-authored annotations. It embodies distinctive features that render it a valuable object of study for broader NLP research.

Languages

Italian, English, Spanish, French

Dataset

Data Fields

The dataset contains a list of instances (rulings); each instance contains the following data:

Field Description
id (str) The ruling ID
ruling_type (int) The ruling type
epigraph (str) The ruling epigraph
text (str) The ruling text
decision (str) The ruling decision
maxims_text (str) The text of ruling maxims
maxims_title (str) The title of ruling maxims
full_text (str) The ruling full_text
num_maxims (int) The number of maxims
maxims_len (int) The length of maxims
full_text_len (int) The length of the full text
judgment_type (int) The judgment type
constitutional_parameters (List[List[str]]) The constitutional parameters
maxims (dict) The maxims' numbers, texts, and titles

Please check the exemplar usage below for loading the data:

from datasets import load_dataset

comma_en = load_dataset("disi-unibo-nlp/COMMA", "en")
# Download comma_en locally and load it as a Dataset object.

example = comma_en["validation"][0]  # The first instance of the dev set 
example["full_text"]  # The full text (i.e., epigraph + text + decision) for the ruling

print(example['maxims_title'])  # The corresponding maxims title for the ruling 

Data Splits

IT Instances
Train (90%) 12,600
Test (5%) 700
Dev (5%) 700
EN Instances
Train (90%) 12,600
Test (5%) 700
Dev (5%) 700
ES Instances
Train (90%) 12,600
Test (5%) 700
Dev (5%) 700
FR Instances
Train (90%) 12,600
Test (5%) 700
Dev (5%) 700