Skip to content

module enigma

solverpy_learn.builder.enigma

Enigma

Bases: EnigmaModel

EnigmaModel combining the sel and gen feature sets, training either or both depending on which feature options are set in setup.

Source code in packages/solverpy-learn/src/solverpy_learn/builder/enigma.py 
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
class Enigma(EnigmaModel):
   """
   [`EnigmaModel`][solverpy_learn.builder.enigma.EnigmaModel] combining the
   `sel` and `gen` feature sets, training either or both depending on which
   feature options are set in `setup`.
   """

   def __init__(
      self,
      setup: Setup,
      tunesel: (dict[str, Any] | None) = None,
      tunegen: (dict[str, Any] | None) = None,
      templates: (list[str] | None) = None,
   ):
      templates = templates or ["coop", "solo", "gen"]
      AutoTuner.__init__(
         self,
         setup,
         tunesel,
         templates=templates,
      )
      assert "evals" in setup
      trains_evalset = setup["evals"]
      assert "plugin" in trains_evalset
      assert "sel_features" in setup
      assert "gen_features" in setup
      sel = setup["sel_features"]
      gen = setup["gen_features"]
      self._features = sel or gen or ""

      setup_sel = setup
      setup_gen = setup
      if sel and gen:
         # split the multi train data for sel and gen sub-builders
         assert isinstance(trains_evalset["plugin"], enigma.EnigmaMultiTrains)
         plugin = trains_evalset["plugin"]
         setup_sel = Setup(setup)
         setup_sel["evals"] = Evalset(trains_evalset, plugin=plugin._sel)
         setup_gen = Setup(setup)
         setup_gen["evals"] = Evalset(trains_evalset, plugin=plugin._gen)
         if "devels" in setup:
            devels_evalset = setup["devels"]
            assert "plugin" in devels_evalset
            assert isinstance(devels_evalset["plugin"], enigma.EnigmaMultiTrains)
            setup_sel["devels"] = Evalset(devels_evalset, plugin=devels_evalset["plugin"]._sel)
            setup_gen["devels"] = Evalset(devels_evalset, plugin=devels_evalset["plugin"]._gen)

      self._sel = EnigmaSel(setup_sel, tunesel, templates) if sel else None
      self._gen = EnigmaGen(setup_gen, tunegen, templates) if gen else None

   def reset(self, dataname: str) -> None:
      if self._sel:
         self._sel.reset(dataname)
      if self._gen:
         self._gen.reset(dataname)
      Builder.reset(self, dataname)

   def build(self, talker: Talker = Talker()) -> None:
      self._strats = []
      if self._sel:
         self._sel.build(talker)
         self._strats.extend(self._sel.strategies)
      if self._gen:
         self._gen.build(talker)
         self._strats.extend(self._gen.strategies)
      if self._sel and self._gen:
         trains = self._setup["evals"]
         assert "refs" in trains
         refs = trains["refs"]
         self._strats.extend(self.applies(refs, self._dataname))

   def apply(self, sid: str, model: str) -> list[str]:
      assert self._sel and self._gen
      (base, args) = sids.split(sid)
      sidsolo = f"{base}-solo"
      sidcoop = f"{base}-coop"
      sidsologen = self.template(sidsolo, "gen", gen)
      sidcoopgen = self.template(sidcoop, "gen", gen)
      args = dict(args, sel=self._sel._dataname, gen=self._gen._dataname)
      sidsologen = sids.fmt(sidsologen, args)
      sidcoopgen = sids.fmt(sidcoopgen, args)
      news = []
      if "solo" in self._templates:
         news.append(sidsologen)
      if "coop" in self._templates:
         news.append(sidcoopgen)
      logger.debug(f"new strategies: {news}")
      self.makemap(os.path.join(bids.dbpath(NAME), model, "enigma.map"))
      return news

EnigmaGen

Bases: EnigmaModel

EnigmaModel fixed to the gen (clause generation) feature variant.

Source code in packages/solverpy-learn/src/solverpy_learn/builder/enigma.py 
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
class EnigmaGen(EnigmaModel):
   """[`EnigmaModel`][solverpy_learn.builder.enigma.EnigmaModel] fixed to the `gen` (clause generation) feature variant."""

   def __init__(
      self,
      setup: Setup,
      tuneargs: (dict[str, Any] | None) = None,
      templates: (list[str] | None) = None,
   ):
      EnigmaModel.__init__(
         self,
         setup,
         tuneargs,
         "gen",
         templates,
      )

   def apply(self, sid: str, model: str) -> list[str]:
      if "gen" not in self._templates:
         return []
      (base, args) = sids.split(sid)
      sidgen = self.template(base, "gen", gen)
      sidgen = sids.fmt(sidgen, dict(args, gen=model))
      news = [sidgen]
      self.makemap(os.path.join(bids.dbpath(NAME), model, "enigma.map"))
      logger.debug(f"new strategies: {news}")
      return news

EnigmaModel

Bases: AutoTuner

AutoTuner for one ENIGMA model variant (a feature set such as sel or gen). Subclasses EnigmaSel and EnigmaGen fix the variant; Enigma combines both.

Source code in packages/solverpy-learn/src/solverpy_learn/builder/enigma.py 
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
class EnigmaModel(AutoTuner):
   """
   [`AutoTuner`][solverpy_learn.builder.autotuner.AutoTuner] for one ENIGMA
   model variant (a feature set such as `sel` or `gen`). Subclasses
   [`EnigmaSel`][solverpy_learn.builder.enigma.EnigmaSel] and
   [`EnigmaGen`][solverpy_learn.builder.enigma.EnigmaGen] fix the variant;
   [`Enigma`][solverpy_learn.builder.enigma.Enigma] combines both.
   """

   def __init__(
      self,
      setup: Setup,
      tuneargs: (dict[str, Any] | None),
      variant: str,
      templates: (list[str] | None) = None,
   ):
      AutoTuner.__init__(self, setup, tuneargs, templates)
      self._variant = variant
      assert f"{variant}_features" in setup
      self._features: str = setup[f"{variant}_features"]
      self.reset(self._dataname)

   def featurepath(self) -> str:
      fpath = enigma.featurepath(self._features)
      return f"{self._variant}_{fpath}"

   def reset(self, dataname: str) -> None:
      dataname = os.path.join(dataname, self.featurepath())
      super().reset(dataname)  # does: self._dataname = dataname

   def template(
      self,
      sid: str,
      name: str,
      mk_strat: Callable[[str], str],
   ) -> str:
      sidml = f"{sid}-{name}"
      if os.path.exists(sids.path(sidml)):
         logger.debug(f"ml strategy {sidml} already exists")
         return sidml
      strat = mk_strat(sid)
      sids.save(sidml, strat)
      logger.debug(
         f"created parametric ml strategy {sidml} inherited from {sid}:\n{strat}"
      )
      return sidml

   def build(self, talker: Talker = Talker()) -> None:
      super().build(talker)
      self.makemap()

   def makemap(self, mapfile: str | None = None) -> None:
      f_map = mapfile or self.path("enigma.map")
      logger.debug(f"creating enigma map: {f_map}")
      open(f_map, "w").write(f'features("{self._features}").\n')

EnigmaSel

Bases: EnigmaModel

EnigmaModel fixed to the sel (clause selection) feature variant.

Source code in packages/solverpy-learn/src/solverpy_learn/builder/enigma.py 
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
class EnigmaSel(EnigmaModel):
   """[`EnigmaModel`][solverpy_learn.builder.enigma.EnigmaModel] fixed to the `sel` (clause selection) feature variant."""

   def __init__(
      self,
      setup: Setup,
      tuneargs: (dict[str, Any] | None) = None,
      templates: (list[str] | None) = None,
   ):
      EnigmaModel.__init__(
         self,
         setup,
         tuneargs,
         "sel",
         templates,
      )

   def apply(self, sid: str, model: str) -> list[str]:
      (base, args) = sids.split(sid)
      sidsolo = self.template(base, "solo", solo)
      sidsolo = sids.fmt(sidsolo, dict(args, sel=model))
      sidcoop = self.template(base, "coop", coop)
      sidcoop = sids.fmt(sidcoop, dict(args, sel=model))
      news = []
      if "solo" in self._templates:
         news.append(sidsolo)
      if "coop" in self._templates:
         news.append(sidcoop)
      self.makemap(os.path.join(bids.dbpath(NAME), model, "enigma.map"))
      logger.debug(f"new strategies: {news}")
      return news