krdict_json_to_obj.py

import orjson
from typing import NamedTuple


IN_FILE="전체 내려받기_한국어기초사전_json_20230901/3_5000_20230901.json"


class WordEquivalent(NamedTuple):
  language: str = None
  lemma: str = None
  definition: str = None

class WordExample(NamedTuple):
  type: str = None
  example: list[str] = []

class WordForm(NamedTuple):
  type: str = None
  pronunciation: str = None
  sound: str = None

class WordType(NamedTuple):
  homonym_number: str = None
  lexicalUnit: str = None
  partOfSpeech: str = None
  vocabularyLevel: str = None
  semanticCategory: str = None
  subjectCategiory: str = None
  annotation: str = None
  origin: str = None

class WordMeaning(NamedTuple):
  id: int = None
  definition: str = None
  syntacticPattern: str = None
  annotation: str = None
  equivalent: list[WordEquivalent] = []
  form: list[WordForm] = []
  type: list[WordType] = []
  examples: list[WordExample] = []

class Word(NamedTuple):
  writtenForm: str = None
  variant: str = None
  meanings: list[WordMeaning] = []


# 리스트가 아닐 경우 리스트를 씌워주기
def flatten_arrays(arr_or_obj):
  arr = None
  if not isinstance(arr_or_obj, list):
    arr = [ arr_or_obj ]
  else:
    arr = arr_or_obj
  return arr

# 사전 내 원본 feat = { "feat": [ { "att": KEY, "val": VALUE } ] }
# Flatten 결과 feat = { key: VALUE }
def flatten_feats(feats):
  if feats is None or len(feats) <= 0:
    return None

  flat_feats = dict()

  # 비정상 feat : feat이 배열로 묶여 있는 경우. 풀어서 하나씩 추가
  # 비정상 예시 : [ { "feat" : {} }, { "feat" : {} } ]
  if isinstance(feats, list) and len(feats) > 0 and feats[0].get("feat", None) is not None:
    for a in feats:
      flat_feats = flat_feats | flatten_feats(a)

  # 정상 feat
  # 정상 예시 : [ "feat" : [ {}, {} ]
  else:
    feats = feats["feat"]
    feats = flatten_arrays(feats) 

    for feat in feats:
      flat_feats[feat["att"]] = feat["val"]

  return flat_feats


# JSON을 Word 객체로 변환
def json_to_word(json):
  words = list[Word]
  for lexres_k, lexres_v in json.get("LexicalResource").items():

    if lexres_k == "GlobalInformation":
      print(flatten_feats(lexres_v))

    if lexres_k == "Lexicon":
      print(lexres_k)

      for lex_k, lex_v in lexres_v.items():
        if lex_k == "LexicalEntry":
          print(lex_k)

          entries = lex_v

          for entry in entries:
            lemma = flatten_feats(entry["Lemma"])

            writtenForm = lemma.get("writtenForm", None)
            variant = lemma.get("variant", None)

            meanings = []
            senses = flatten_arrays(entry["Sense"])
            for sense in senses:

              equivalent = []
              for eq_feat in sense.get("Equivalent", []):
                equivalent.append(flatten_feats(eq_feat))

              examples = []
              exs = sense.get("SenseExample", [])
              if not isinstance(exs, list):
                exs = flatten_arrays(exs)
              for ex_feat in exs:
                examples.append(flatten_feats(ex_feat))

              form = []
              for fo_feat in sense.get("WordForm", []):
                fo_feats.append(flatten_feats(fo_feat))

              type = flatten_feats({"feat": entry.get("feat")})

              feats = flatten_feats(sense)
              meaning = WordMeaning(
                                    id=int(sense["val"]),
                                    definition=feats.get("definition", None),
                                    syntacticPattern=feats.get("syntacticPattern", None),
                                    annotation=feats.get("annotation", None),
                                    equivalent=equivalent,
                                    form=form,
                                    type=type,
                                    examples=examples
                                    )
              meanings.append(meaning)


            w = Word(
                  writtenForm=writtenForm,
                  variant=variant,
                  meanings=meanings)

            print(w)



with open(IN_FILE, "r", encoding="utf-8") as f:
  json_obj = orjson.loads(f.read())

json_to_word(json_obj)