mappings.yaml

# This file contains mappings from various annotation schemas to common numeric values.
# Such a conversion facilitates the application of algorithms to datasets and ensures their conceptual compatibility.
# Currently, we support the following schemas:
# PER LOC MISC; used by the UGARIT/flair_grc_multi_ner model and the corresponding dataset
# https://www.researchgate.net/publication/365131651_Transformer-Based_Named_Entity_Recognition_for_Ancient_Greek
per_loc_misc:
  # add both the letter O and the number 0 because some values are not used consistently
  "O": 0
  "0": 0
  # persons
  "PER": 1
  # locations
  "LOC": 2
  # miscellaneous, such as peoples
  "MISC": 3
# PER LOC NORP; used by the LatinCy NER tagger
# https://arxiv.org/pdf/2305.04365
per_loc_norp:
  # add an empty string because this schema only provides explicit values for named entities
  "": 0
  # persons
  "PERSON": 1
  # locations
  "LOC": 2
  # groups of people
  "NORP": 3
# PRS GEO GRP; used by the Herodotos dataset
# https://github.com/Herodotos-Project/Herodotos-Project-Latin-NER-Tagger-Annotation
prs_geo_grp:
  # add both the letter O and the number 0 because some values are not used consistently
  "O": 0
  "0": 0
  # persons
  "PRS": 1
  # locations
  "GEO": 2
  # groups
  "GRP": 3