README.md

metadata

language:
  - ar
configs:
  - config_name: annotations
    default: true
    data_files:
      - split: dev
        path: annotations.csv
  - config_name: contexts_nopnx
    data_files:
      - split: dev
        path: contexts_nopnx.csv
  - config_name: contexts_pnx
    data_files:
      - split: dev
        path: contexts_pnx.csv

ArabiGEE Data

Paper title: ArabiGEE: A Hierarchical Taxonomy for Arabic Grammatical Error Explanation

annotations.csv

• context_id: Links the annotation row to either context file.
• pair_id: Word-pair identifier in this dataset.
• original_pair_id: Original pair identifier in the source data.
• error_id: Error number within the same pair_id.
• erroneous_word: Erroneous word or phrase.
• target_word: Target word or phrase.
• areta_label: ARETA edit label.
• lex_code: Lexical explanation code.
• lex_explanation_text: Lexical explanation text.
• orth_code: Orthographic explanation code.
• orth_explanation_text: Orthographic explanation text.
• morph_code: Morphological explanation code.
• morph_explanation_text: Morphological explanation text.
• synt_code: Syntactic explanation code.
• synt_explanation_text: Syntactic explanation text.

contexts_nopnx.csv

• context_id: Context identifier in this dataset.
• source_dataset: Source dataset and split label.
• original_sentence_id: Original sentence identifier in the source data.
• src_context: Erroneous sentence context without punctuation.
• target_context: Corrected sentence context without punctuation.

contexts_pnx.csv

• context_id: Context identifier in this dataset.
• source_dataset: Source dataset and split label.
• original_sentence_id: Original sentence identifier in the source data.
• src_context: Erroneous sentence context with punctuation.
• target_context: Corrected sentence context with punctuation.