sequoia.dom 4.71 KB
Newer Older
bguillaum's avatar
bguillaum committed
1 2 3 4 5 6 7 8 9 10 11 12
% ====================================================================================================
features {
  % phon and lemma features are built-in CONLL format features (phon is field 2 and lemma is field 3)
  phon: *;
  lemma: *;

  % cat and pos features are built-in CONLL format features (phon is field 4 and lemma is field 5)
  cat:  V, N, C, CL, P, I, PONCT, A, ADV, PRO, D, ET, "P+D", "P+PRO", PREF;
  pos:  V, VS, VINF, VPP, VPR, VIMP, N, NC, NPP, C, CS, CC, CL, CLS, CLO, CLR, P, "P+D", "P+PRO", I, PONCT, ET, ADJ, ADJWH, ADV, ADVWH, PRO, PROREL, PROWH, DET, DETWH, PREF;

  mwehead: V, VS, VINF, VPP, VPR, VIMP, N, NC, NPP, C, CS, CC, CL, CLS, CLO, CLR, P, "P+D", "P+PRO", I, PONCT, ET, ADJ, ADJWH, ADV, ADVWH, PRO, PROREL, PROWH, DET, DETWH, PREF;
  mwelemma: *;
Bruno Guillaume's avatar
Bruno Guillaume committed
13

bguillaum's avatar
bguillaum committed
14 15
  % In Sequoia, sentence identifier are given through a feature sentid on the first node of the CONLL description.
  sentid: *;
Bruno Guillaume's avatar
Bruno Guillaume committed
16

bguillaum's avatar
bguillaum committed
17 18 19 20 21
  % basic features
  g: m,f;
  n: s,p;
  p: 1,2,3, 12, 13; % 13 added for Talismane output
  s: c, card, def, dem, exel, ind, int, neg, obj, ord, p, part, pers, poss, qual, refl, rel, s, suj, w;
Bruno Guillaume's avatar
Bruno Guillaume committed
22
  m: ind, inf, part, subj, imp, pastp;
bguillaum's avatar
bguillaum committed
23 24 25 26 27 28
  t: pst, impft, past, fut, cond,

  % output of Talismane, should be converted !!!
  K, J, P, II, W, C, G, T, PS, F, S, Y, PJ;
  poss: s, p;

Bruno Guillaume's avatar
Bruno Guillaume committed
29 30
  component:y;

bguillaum's avatar
bguillaum committed
31 32 33 34 35 36 37
  % Features introduced in deep-sequoia
  diat: caust, caust_seobj, causi, sefaire, passif, impers, passif_impers, demsuj;
  def:y;
  intrinsimp: y;
  se: ppdc;
  cltype : int, imp, excl;
  lexicalisation: causi_fige, caust_fige;
Bruno Guillaume's avatar
Bruno Guillaume committed
38

bguillaum's avatar
bguillaum committed
39 40
  % fctpath encode information needed to deal with non-projectivity of some sentences
  fctpath:*;
Bruno Guillaume's avatar
Bruno Guillaume committed
41

bguillaum's avatar
bguillaum committed
42 43 44 45 46 47
  % --------------------------------------------------
  % features that are used only in the deep representation
  % --------------------------------------------------
  dl: *;
  dm: ind, inf, part, subj, imp, pastpart, pstpart;
  void:y;
48

49
  agent:y;
50 51 52
  det:y;
  type: *;
  adv_man:true;
bguillaum's avatar
bguillaum committed
53
}
Bruno Guillaume's avatar
Bruno Guillaume committed
54

bguillaum's avatar
bguillaum committed
55 56
% ====================================================================================================
labels {
Bruno Guillaume's avatar
Bruno Guillaume committed
57

bguillaum's avatar
bguillaum committed
58 59 60 61
  % --------------------------------------------------
  % surface labels
  % --------------------------------------------------
  root,
Bruno Guillaume's avatar
Bruno Guillaume committed
62

bguillaum's avatar
bguillaum committed
63 64
  % function used in diathesis alternation: they are not used like this in deep-sequoia but with the double function notation X:Y
  suj, obj, de_obj, a_obj, ats, ato, p_obj.agt,
Bruno Guillaume's avatar
Bruno Guillaume committed
65

bguillaum's avatar
bguillaum committed
66 67 68 69 70
  p_obj.o,
  obj.p, obj.cpl,
  aux.tps, aux.pass, aux.caus,
  dis,
  aff, aff.demsuj,
Bruno Guillaume's avatar
Bruno Guillaume committed
71

bguillaum's avatar
bguillaum committed
72 73 74 75
  det, mod,
  mod, mod.app, mod.cleft, mod.inc, mod.rel, mod.voc,
  ponct, coord, dep.coord,
  dep, arg,
Bruno Guillaume's avatar
Bruno Guillaume committed
76

bguillaum's avatar
bguillaum committed
77 78 79 80 81 82 83 84
  dep_cpd,
  % --------------------------------------------------
  % deep only labels
  % --------------------------------------------------
  argc,
  mod.comp, mod.super, mod.rel.part,
  arg.comp, arg.cons, dep.de,
  void,
Bruno Guillaume's avatar
Bruno Guillaume committed
85

bguillaum's avatar
bguillaum committed
86 87 88 89
  % --------------------------------------------------
  % mixed labels
  % --------------------------------------------------
  S:root, D:root,
Bruno Guillaume's avatar
Bruno Guillaume committed
90

bguillaum's avatar
bguillaum committed
91 92 93 94 95 96 97 98 99 100 101
  % bilabel
  suj:suj@black, D:suj:suj@blue@bottom, S:suj:suj@red, I:suj:suj@grey@bottom,
  suj:obj@black, D:suj:obj@blue@bottom, S:suj:obj@red, I:suj:obj@grey@bottom,
  obj:obj@black, D:obj:obj@blue@bottom, S:obj:obj@red, I:obj:obj@grey@bottom,
  a_obj:a_obj@black, D:a_obj:a_obj@blue@bottom, S:a_obj:a_obj@red, I:a_obj:a_obj@grey@bottom,
  de_obj:de_obj@black, D:de_obj:de_obj@blue@bottom, S:de_obj:de_obj@red, I:de_obj:de_obj@grey@bottom,
  a_obj:suj@black, D:a_obj:suj@blue@bottom, S:a_obj:suj@red,
  p_obj.agt:suj@black, D:p_obj.agt:suj@blue@bottom, S:p_obj.agt:suj@red,
  obj:suj@black, S:obj:suj@red, D:obj:suj@blue@bottom,
  suj:_@black, D:suj:_@blue@bottom, S:suj:_@red,
  suj:argc@black, D:suj:argc@blue@bottom,
Bruno Guillaume's avatar
Bruno Guillaume committed
102

bguillaum's avatar
bguillaum committed
103 104 105
  ats:ats, S:ats:ats@red, D:ats:ats@blue@bottom,
  ato:ato, S:ato:ato@red, D:ato:ato@blue@bottom,
  ats:ato@black,
Bruno Guillaume's avatar
Bruno Guillaume committed
106

bguillaum's avatar
bguillaum committed
107 108 109
  dis:dis@black,
  dis:suj@black, S:dis:suj@red, D:dis:suj@blue@bottom,
  dis:obj@black, S:dis:obj@red, D:dis:obj@blue@bottom,
Bruno Guillaume's avatar
Bruno Guillaume committed
110

bguillaum's avatar
bguillaum committed
111 112 113 114 115 116 117
  % modifier unconsidered in surface annoation and defined in deep
  D:mod.comp@blue@bottom, I:mod.comp@grey@bottom,
  D:arg.comp@blue@bottom, I:arg.comp@grey@bottom,
  D:arg.cons@blue@bottom, I:arg.cons@grey@bottom,
  D:dep.de@blue@bottom, I:dep.de@grey@bottom,
  D:mod.super@blue@bottom,
  D:mod.rel.part @blue@bottom, % TODO comment
Bruno Guillaume's avatar
Bruno Guillaume committed
118

bguillaum's avatar
bguillaum committed
119 120
  %relation specific to surface:
  S:aux.tps@red, S:aux.pass@red, S:aux.caus@red,
Bruno Guillaume's avatar
Bruno Guillaume committed
121

bguillaum's avatar
bguillaum committed
122 123 124 125
  D:p_obj.o@blue@bottom, S:p_obj.o@red,
  D:mod@blue@bottom, S:mod@red,
  S:aff@red,
  S:aff.demsuj@red,
Bruno Guillaume's avatar
Bruno Guillaume committed
126

bguillaum's avatar
bguillaum committed
127 128 129
  S:arg@red, D:arg@blue@bottom, I:arg@grey@bottom,
  S:mod.rel@red, D:mod.rel@blue@bottom,
  D:mod.app@blue@bottom, S:mod.app@red,
Bruno Guillaume's avatar
Bruno Guillaume committed
130

bguillaum's avatar
bguillaum committed
131 132 133 134 135
  S:ponct@red, D:ponct@blue@bottom,
  S:coord@red, D:coord@blue@bottom,
  S:dep.coord@red, D:dep.coord@blue@bottom,
  S:dep@red, D:dep@blue@bottom, I:dep@grey@bottom,
  S:det@red, D:det@blue@bottom,
Bruno Guillaume's avatar
Bruno Guillaume committed
136

bguillaum's avatar
bguillaum committed
137 138 139
  D:obj.p@blue@bottom, S:obj.p@red,
  D:obj.cpl@blue@bottom, S:obj.cpl@red,
}