propp-press / README.md
regicid's picture
Add dataset card with characters/entities configs (incl. lemonde)
e80d602 verified
|
Raw History Blame Contribute Delete
1.95 kB
metadata
configs:
  - config_name: characters
    data_files:
      - split: lacroix
        path: data/characters/lacroix/*.parquet
      - split: le_journal_du_dimanche
        path: data/characters/le_journal_du_dimanche/*.parquet
      - split: lefigaro
        path: data/characters/lefigaro/*.parquet
      - split: lemonde
        path: data/characters/lemonde/*.parquet
      - split: leparisien
        path: data/characters/leparisien/*.parquet
      - split: lesechos
        path: data/characters/lesechos/*.parquet
      - split: liberation
        path: data/characters/liberation/*.parquet
      - split: mediapart
        path: data/characters/mediapart/*.parquet
      - split: ouestfrance
        path: data/characters/ouestfrance/*.parquet
      - split: paris_match
        path: data/characters/paris_match/*.parquet
      - split: sud_ouest
        path: data/characters/sud_ouest/*.parquet
      - split: valeurs_actuelles
        path: data/characters/valeurs_actuelles/*.parquet
  - config_name: entities
    data_files:
      - split: lacroix
        path: data/entities/lacroix/*.parquet
      - split: le_journal_du_dimanche
        path: data/entities/le_journal_du_dimanche/*.parquet
      - split: lefigaro
        path: data/entities/lefigaro/*.parquet
      - split: lemonde
        path: data/entities/lemonde/*.parquet
      - split: leparisien
        path: data/entities/leparisien/*.parquet
      - split: lesechos
        path: data/entities/lesechos/*.parquet
      - split: liberation
        path: data/entities/liberation/*.parquet
      - split: mediapart
        path: data/entities/mediapart/*.parquet
      - split: ouestfrance
        path: data/entities/ouestfrance/*.parquet
      - split: paris_match
        path: data/entities/paris_match/*.parquet
      - split: sud_ouest
        path: data/entities/sud_ouest/*.parquet
      - split: valeurs_actuelles
        path: data/entities/valeurs_actuelles/*.parquet

propp-press

Character (characters) and entity-mention (entities) annotations produced with propp_fr on French press articles, one split per newspaper. Tables join on doc_id.