Grammalecte  Check-in [f11e2ad39c]

Overview
Comment:[core] ibdawg: suggest() use sets instead of lists
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | trunk | core
Files: files | file ages | folders
SHA3-256: f11e2ad39c0469b0b44e29b754fd0899a99bdbe4d7f7eed70a6307d698f50e38
User & Date: olr on 2017-07-02 19:53:02
Other Links: manifest | tags
Context
2017-07-03
02:35
[core] ibdawg: code cleaning check-in: f4c85fbe0b user: olr tags: trunk, core
2017-07-02
19:53
[core] ibdawg: suggest() use sets instead of lists check-in: f11e2ad39c user: olr tags: trunk, core
18:12
[fr] pt: d’évidence (faux positif) + bug de suggestion check-in: e8450a49af user: olr tags: trunk, fr
Changes

Modified gc_core/py/ibdawg.py from [5262800a60] to [48f4ad6627].

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

225
226
227
228

229
230
231
232
233
234
235


236
237
238
239
240
241
242
243
244
245

246
247
248
249
250


251
252
253
254
255
256


257
258
259
260
261
262
263
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
225
226
227

228
229
230
231
232
233


234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249


250
251

252
253
254


255
256
257
258
259
260
261
262
263







-
-
-
-
-
-
-
-
+
+
+
+
+
+
+
+



+


-
-
-
+
+
+

-



-
+




-
+


-
+


-
+



-
+





-
-
+
+










+



-
-
+
+
-



-
-
+
+







                l.extend(self.morph(sWord.capitalize()))
        return l

    def suggest (self, sWord):
        "returns a set of similar words"
        # first, we check for similar words
        #return set(self._suggestWithCrushedUselessChars(cp.clearWord(sWord)))
        lSugg = self._suggest(sWord)
        if not lSugg:
            lSugg.extend(self._suggest(sWord[1:]))
            lSugg.extend(self._suggest(sWord[:-1]))
            lSugg.extend(self._suggest(sWord[1:-1]))
            if not lSugg:
                lSugg.extend(self._suggestWithCrushedUselessChars(cp.clearWord(sWord)))
        return set(lSugg)
        aSugg = self._suggest(sWord)
        if not aSugg:
            aSugg.update(self._suggest(sWord[1:]))
            aSugg.update(self._suggest(sWord[:-1]))
            aSugg.update(self._suggest(sWord[1:-1]))
            if not aSugg:
                aSugg.update(self._suggestWithCrushedUselessChars(cp.clearWord(sWord)))
        return aSugg

    def _suggest (self, sWord, nDeep=0, iAddr=0, sNewWord="", bAvoidLoop=False):
        # RECURSIVE FUNCTION
        aSugg = set()
        if not sWord:
            if int.from_bytes(self.byDic[iAddr:iAddr+self.nBytesArc], byteorder='big') & self._finalNodeMask:
                show(nDeep, "___" + sNewWord + "___")
                return [sNewWord]
            return []
                #show(nDeep, "___" + sNewWord + "___")
                aSugg.add(sNewWord)
            return aSugg
        #show(nDeep, "<" + sWord + ">  ===>  " + sNewWord)
        lSugg = []
        cCurrent = sWord[0:1]
        for cChar, jAddr in self._getSimilarArcs(cCurrent, iAddr):
            #show(nDeep, cChar)
            lSugg.extend(self._suggest(sWord[1:], nDeep+1, jAddr, sNewWord+cChar))
            aSugg.update(self._suggest(sWord[1:], nDeep+1, jAddr, sNewWord+cChar))
        if not bAvoidLoop: # avoid infinite loop
            #show(nDeep, ":no loop:")
            if cCurrent == sWord[1:2]:
                # same char, we remove 1 char without adding 1 to <sNewWord>
                lSugg.extend(self._suggest(sWord[1:], nDeep+1, iAddr, sNewWord))
                aSugg.update(self._suggest(sWord[1:], nDeep+1, iAddr, sNewWord))
            for sRepl in cp.d1toX.get(cCurrent, ()):
                #show(nDeep, sRepl)
                lSugg.extend(self._suggest(sRepl + sWord[1:], nDeep+1, iAddr, sNewWord, True))
                aSugg.update(self._suggest(sRepl + sWord[1:], nDeep+1, iAddr, sNewWord, True))
            for sRepl in cp.d2toX.get(sWord[0:2], ()):
                #show(nDeep, sRepl)
                lSugg.extend(self._suggest(sRepl + sWord[2:], nDeep+1, iAddr, sNewWord, True))
                aSugg.update(self._suggest(sRepl + sWord[2:], nDeep+1, iAddr, sNewWord, True))
            if len(sWord) == 2:
                for sRepl in cp.dFinal2.get(sWord, ()):
                    #show(nDeep, sRepl)
                    lSugg.extend(self._suggest(sRepl, nDeep+1, iAddr, sNewWord, True))
                    aSugg.update(self._suggest(sRepl, nDeep+1, iAddr, sNewWord, True))
            elif len(sWord) == 1:
                #show(nDeep, ":end of word:")
                # end of word
                for sRepl in cp.dFinal1.get(sWord, ()):
                    #show(nDeep, sRepl)
                    lSugg.extend(self._suggest(sRepl, nDeep+1, iAddr, sNewWord, True))
        return lSugg
                    aSugg.update(self._suggest(sRepl, nDeep+1, iAddr, sNewWord, True))
        return aSugg

    def _getSimilarArcs (self, cChar, iAddr):
        "generator: yield similar char of <cChar> and address of the following node"
        for c in cp.d1to1.get(cChar, [cChar]):
            if c in self.dChar:
                jAddr = self._lookupArcNode(self.dChar[c], iAddr)
                if jAddr:
                    yield (c, jAddr)

    def _suggestWithCrushedUselessChars (self, sWord, nDeep=0, iAddr=0, sNewWord="", bAvoidLoop=False):
        aSugg = set()
        if not sWord:
            if int.from_bytes(self.byDic[iAddr:iAddr+self.nBytesArc], byteorder='big') & self._finalNodeMask:
                show(nDeep, "!!! " + sNewWord + " !!!")
                return [sNewWord]
            return []
                aSugg.add(sNewWord)
            return aSugg
        lSugg = []
        cCurrent = sWord[0:1]
        for cChar, jAddr in self._getSimilarArcsAndCrushedChars(cCurrent, iAddr):
            show(nDeep, cChar)
            lSugg.extend(self._suggestWithCrushedUselessChars(sWord[1:], nDeep+1, jAddr, sNewWord+cChar))
        return lSugg
            aSugg.update(self._suggestWithCrushedUselessChars(sWord[1:], nDeep+1, jAddr, sNewWord+cChar))
        return aSugg

    def _getSimilarArcsAndCrushedChars (self, cChar, iAddr):
        "generator: yield similar char of <cChar> and address of the following node"
        for nVal, jAddr in self._getArcs(iAddr):
            if self.dCharVal.get(nVal, None) in cp.aUselessChar:
                yield (self.dCharVal[nVal], jAddr)
        for c in cp.d1to1.get(cChar, [cChar]):