76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
|
"ts_n_dash_start": [ ("^[-—][ ]", "– "),
("^– ", "– "),
("^[-–—](?=[\\w.…])", "– ") ],
"ts_m_dash_start": [ ("^[-–][ ]", "— "),
("^— ", "— "),
("^«[ ][—–-][ ]", "« — "),
("^[-–—](?=[\\w.…])", "— ") ],
"ts_quotation_marks": [ ('"(\\w+)"', "“$1”"),
("''(\\w+)''", "“$1”"),
("'(\\w+)'", "“$1”"),
("^(?:\"|'')(?=\\w)", "« "),
(" (?:\"|'')(?=\\w)", " « "),
("\\((?:\"|'')(?=\\w)", "(« "),
("(?<=\\w)(?:\"|'')$", " »"),
("(?<=\\w)(?:\"|'')(?=[] ,.:;?!…)])", " »"),
('(?<=[.!?…])" ', " » "),
('(?<=[.!?…])"$', " »") ],
|
|
|
|
|
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
|
"ts_n_dash_start": [ ("^[-—][ ]", "– "),
("^– ", "– "),
("^[-–—](?=[\\w.…])", "– ") ],
"ts_m_dash_start": [ ("^[-–][ ]", "— "),
("^— ", "— "),
("^«[ ][—–-][ ]", "« — "),
("^[-–—](?=[\\w.…])", "— ") ],
"ts_quotation_marks": [ ('"(\\w+)"', "“\\1”"),
("''(\\w+)''", "“\\1”"),
("'(\\w+)'", "“\\1”"),
("^(?:\"|'')(?=\\w)", "« "),
(" (?:\"|'')(?=\\w)", " « "),
("\\((?:\"|'')(?=\\w)", "(« "),
("(?<=\\w)(?:\"|'')$", " »"),
("(?<=\\w)(?:\"|'')(?=[] ,.:;?!…)])", " »"),
('(?<=[.!?…])" ', " » "),
('(?<=[.!?…])"$', " »") ],
|
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
|
("noeu", "nœu"), ("Noeu", "Nœu"),
("soeur", "sœur"), ("Soeur", "Sœur"),
("voeu", "vœu"), ("Voeu", "Vœu"),
("aequo", "æquo"), ("Aequo", "Æquo"),
("\\bCa\\b", "Ça"), (" ca\\b", " ça"),
("\\bdej[aà]\\b", "déjà"), ("\\bplutot\\b", "plutôt"),
("\\bmeme\\b", "même"), ("\\bmemes\\b", "mêmes"), ("\\bMeme\\b", "Même"),
("\\b([cC]e(?:ux|lles?|lui))-la\\b", "$1-là"),
("\\bmalgre\\b", "malgré"), ("\\bMalgre\\b", "Malgré"),
("\\betre\\b", "être"), ("\\bEtre\\b", "Être"),
("\\btres\\b", "très"), ("\\bTres\\b", "Très"),
("\\bEtai([ts]|ent)\\b", "Étai$1"),
("\\bE(tat|cole|crit|poque|tude|ducation|glise|conomi(?:qu|)e|videmment|lysée|tienne|thiopie|cosse|gypt(?:e|ien)|rythrée|pinal|vreux)", "É$1") ],
"ts_ligature_ffi_on": [("ffi", "ffi")],
"ts_ligature_ffl_on": [("ffl", "ffl")],
"ts_ligature_fi_on": [("fi", "fi")],
"ts_ligature_fl_on": [("fl", "fl")],
"ts_ligature_ff_on": [("ff", "ff")],
"ts_ligature_ft_on": [("ft", "ſt")],
"ts_ligature_st_on": [("st", "st")],
|
|
|
|
|
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
|
("noeu", "nœu"), ("Noeu", "Nœu"),
("soeur", "sœur"), ("Soeur", "Sœur"),
("voeu", "vœu"), ("Voeu", "Vœu"),
("aequo", "æquo"), ("Aequo", "Æquo"),
("\\bCa\\b", "Ça"), (" ca\\b", " ça"),
("\\bdej[aà]\\b", "déjà"), ("\\bplutot\\b", "plutôt"),
("\\bmeme\\b", "même"), ("\\bmemes\\b", "mêmes"), ("\\bMeme\\b", "Même"),
("\\b([cC]e(?:ux|lles?|lui))-la\\b", "\\1-là"),
("\\bmalgre\\b", "malgré"), ("\\bMalgre\\b", "Malgré"),
("\\betre\\b", "être"), ("\\bEtre\\b", "Être"),
("\\btres\\b", "très"), ("\\bTres\\b", "Très"),
("\\bEtai([ts]|ent)\\b", "Étai\\1"),
("\\bE(tat|cole|crit|poque|tude|ducation|glise|conomi(?:qu|)e|videmment|lysée|tienne|thiopie|cosse|gypt(?:e|ien)|rythrée|pinal|vreux)", "É\\1") ],
"ts_ligature_ffi_on": [("ffi", "ffi")],
"ts_ligature_ffl_on": [("ffl", "ffl")],
"ts_ligature_fi_on": [("fi", "fi")],
"ts_ligature_fl_on": [("fl", "fl")],
"ts_ligature_ff_on": [("ff", "ff")],
"ts_ligature_ft_on": [("ft", "ſt")],
"ts_ligature_st_on": [("st", "st")],
|