Coverage for /pythoncovmergedfiles/medio/medio/usr/local/lib/python3.11/site-packages/nameparser/config/conjunctions.py: 100%

1from nameparser.config._invariants import assert_normalized 

2 

3CONJUNCTIONS = frozenset({ 

4 '&', 

5 'and', 

6 'et', 

7 'e', 

8 'of', 

9 'the', 

10 'und', 

11 'y', 

12 # #397: the Catalan/Polish surname link ("Carod i Rovira", 

13 # "Kowalski i Nowak"). Single-letter like 'y'/'и'/'e', and the one 

14 # entry that is ALSO generational vocabulary -- 'i' is the roman 

15 # numeral I, a bare entry of SUFFIX_WORDS -- so the carve-out 

16 # counts it as a name word for its own sake and it joins only 

17 # with a name word on each side (rules.md#P3). 

18 'i', 

19 # #269: Cyrillic (ru/uk/bg) "and": и, і, та. Ukrainian writes і and 

20 # й for the same conjunction, alternating on the surrounding 

21 # vowel/consonant for euphony ("Олесь і Олена", "Марія й Петро"), 

22 # so real data carries both spellings and neither alone suffices. 

23 'и', 

24 'і', 

25 'й', 

26 'та', 

27 # #269: Greek "and": και. 

28 'και', 

29 # #269 follow-up: Arabic "and". Formal script attaches و to the 

30 # following word (وفاطمة), so a standalone و token appears only in 

31 # informal spacing -- common in real data. Single-character like 

32 # 'y'/'и': the single-letter carve-out (Google Code issue 11, 

33 # the "john e smith" bug) protects short names (joins 

34 # only with enough rootname pieces). 

35 'و', 

36}) 

37""" 

38Pieces that should join to their neighboring pieces, e.g. "and", "y" and "&". 

39"of" and "the" are also include to facilitate joining multiple titles, 

40e.g. "President of the United States". 

41""" 

42 

43CONJUNCTIONS_AMBIGUOUS = frozenset({ 

44 # #383/#479: single letters that read as an INITIAL rather than a 

45 # connective when the input carries no case evidence -- a name 

46 # written wholly in one case, upper or lower alike. A letter 

47 # outside this set joins there, bare capital included. 

48 # 

49 # 'e' and not 'y', measured: a bare E initial is common (Edward, 

50 # Elizabeth) and an 'e' between two surnames is rare outside couple 

51 # listings; a bare Y initial is rare and 'y' between two surnames is 

52 # the commonest Hispanic compound and this library's oldest fixture 

53 # ('Velasquez y Garcia'). Cyrillic и/і/й follow y, not e: #267 

54 # blessed their joining and nothing here narrows it. 

55 # 

56 # 'i' (Catalan/Polish) ships here beside 'e' for the same reason 

57 # (#397): a bare I initial is as common as a bare E, so a name 

58 # written wholly in one case reads the letter as an initial and 

59 # reports the fork rather than joining in silence. 

60 # 

61 # A caller edits the vocabulary rather than a switch: remove 'e' to 

62 # restore joining for Portuguese data, remove 'i' for Catalan or 

63 # Polish data, add 'y' for a Dutch-style "every single letter is an 

64 # initial". 

65 'e', 

66 'i', 

67}) 

68""" 

69Single-letter entries of :data:`CONJUNCTIONS` that read as an initial, 

70not a connective, in a name written wholly in one case. 

71""" 

72 

73 

74assert_normalized("CONJUNCTIONS", CONJUNCTIONS) 

75 

76# Guard the invariant the docstring promises, so a future edit that 

77# breaks it fails at import time (same rationale as suffixes.py). 

78# This holds the SHIPPED constant and nothing else, and two things 

79# follow. `assert` is stripped under `python -O`, so under -O even 

80# that much is gone. And unlike the other marker subsets, the pair is 

81# deliberately absent from Lexicon's subset checks, so a CALLER'S 

82# orphan is never rejected -- it is inert instead, the classify fork 

83# and its emitter both requiring the base entry before they read the 

84# marker (see decisions.md#P3, the 2026-09-13 entries). 

85assert CONJUNCTIONS_AMBIGUOUS <= CONJUNCTIONS, \ 

86 "CONJUNCTIONS_AMBIGUOUS must stay a subset of CONJUNCTIONS" 

87assert all(len(w) == 1 and w.upper() != w.lower() 

88 for w in CONJUNCTIONS_AMBIGUOUS), \ 

89 "CONJUNCTIONS_AMBIGUOUS holds CASED single letters; the fork only " \ 

90 "reads a cased single-letter token, so a caseless letter (Arabic " \ 

91 "و) or a multi-letter entry would be silently inert" 

92assert_normalized("CONJUNCTIONS_AMBIGUOUS", CONJUNCTIONS_AMBIGUOUS)