# Expectations for the fields of LetterInfo NOT checked by Example 1 (whose test_data.yml covers the others), written by reading the letters.
# Conventions (decided once, for all fields, and documented here, where the committee edits them):
# - strings are matched TOLERANTLY: case-, accent-, and punctuation-insensitive, the expected text must be CONTAINED in the actual one
# - a list of strings means: any of them is fine (e.g. a title, in the original language or translated)
# - `null` means: the letter does not say it, so the field must be empty (YAML's null = Python's None)
# - `attended` lists courses which must ALL be among the extracted ones (others may be there too, in any order)
- input: letter-mario-rossi.txt
  expectations:
    degree: master
    alma_mater: null                     # the letter never says where Mario studied (Bologna is the author's affiliation!)
    attended: [machine learning and data mining]
    affiliation: university of bologna
    position: associate professor
    seniority: [3]                       # admissible values (a list, in case of ambiguity)
    nationality: null                    # never stated
    application_for: data science
- input: letter-jean-dupont.txt
  expectations:
    degree: [master, informatique]
    alma_mater: [universite de lyon, university of lyon]
    attended: [apprentissage automatique, vision par ordinateur]
    affiliation: [universite de lyon, university of lyon]
    position: [maitresse de conferences, associate professor, lecturer]  # a French title, maybe translated
    seniority: [1, 2, 3]                 # ambiguous: "maîtresse de conférences" is a tenured lecturer, i.e. 1, 2, or 3 depending on the mapping
    nationality: null
    application_for: artificial intelligence
- input: letter-mohammed-ali.txt
  expectations:
    degree: software engineering
    alma_mater: [university of amman, null]   # implied ("our academic programme", the author works there), not stated: both are fine
    attended: [databases, computer networks]
    affiliation: university of amman
    position: associate professor
    seniority: [3]
    nationality: [null, jordan]          # "Amman, Jordan" is the author's address, not their nationality: a model may conflate them
    application_for: cyber security