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
264
265
266
267
268
269
270
271
272
273
274
275
276
|
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
|
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
|
if sWord.istitle():
aSugg.update(self._suggest(sWord.lower(), nMaxDel=nMaxDel, nMaxHardRepl=nMaxHardRepl))
aSugg = set(map(lambda sSugg: sSugg.title(), aSugg))
elif sWord.islower():
aSugg.update(self._suggest(sWord.title(), nMaxDel=nMaxDel, nMaxHardRepl=nMaxHardRepl))
aSugg = cp.filterSugg(aSugg)
sCleanWord = cp.cleanWord(sWord)
aSugg = sorted(aSugg, key=lambda sSugg: cp.distanceDamerauLevenshtein(sCleanWord, cp.cleanWord(sSugg)))[:nMaxSugg]
aSugg = sorted(aSugg, key=lambda sSugg: st.distanceDamerauLevenshtein(sCleanWord, cp.cleanWord(sSugg)))[:nMaxSugg]
if sSfx or sPfx:
# we add what we removed
return list(map(lambda sSug: sPfx + sSug + sSfx, aSugg))
return aSugg
def _suggest (self, sRemain, nMaxDel=0, nMaxHardRepl=0, nDeep=0, iAddr=0, sNewWord="", bAvoidLoop=False):
def _suggest (self, sRemain, nMaxDel=0, nMaxHardRepl=0, nDeep=0, iAddr=0, sNewWord="", sAction="", bAvoidLoop=False):
"returns a set of suggestions"
# recursive function
#logging.info((nDeep * " ") + sNewWord + ":" + sRemain)
#logging.info((nDeep * " ") + sNewWord + ":" + sRemain + " · " + sAction)
aSugg = set()
if not sRemain:
if int.from_bytes(self.byDic[iAddr:iAddr+self.nBytesArc], byteorder='big') & self._finalNodeMask:
#logging.info((nDeep * " ") + "__" + sNewWord + "__")
aSugg.add(sNewWord)
for sTail in self._getTails(iAddr):
#logging.info((nDeep * " ") + "__" + sNewWord+sTail + "__")
aSugg.add(sNewWord+sTail)
return aSugg
cCurrent = sRemain[0:1]
for cChar, jAddr in self._getSimilarArcs(cCurrent, iAddr):
aSugg.update(self._suggest(sRemain[1:], nMaxDel, nMaxHardRepl, nDeep+1, jAddr, sNewWord+cChar))
aSugg.update(self._suggest(sRemain[1:], nMaxDel, nMaxHardRepl, nDeep+1, jAddr, sNewWord+cChar, "*"))
if not bAvoidLoop: # avoid infinite loop
if cCurrent == sRemain[1:2]:
# same char, we remove 1 char without adding 1 to <sNewWord>
aSugg.update(self._suggest(sRemain[1:], nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord))
aSugg.update(self._suggest(sRemain[1:], nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, cCurrent+"/2"))
else:
# switching chars
aSugg.update(self._suggest(sRemain[1:2]+sRemain[0:1]+sRemain[2:], nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, True))
aSugg.update(self._suggest(sRemain[1:2]+sRemain[0:1]+sRemain[2:], nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, "><",True))
# delete char
if nMaxDel > 0:
aSugg.update(self._suggest(sRemain[1:], nMaxDel-1, nMaxHardRepl, nDeep+1, iAddr, sNewWord, True))
aSugg.update(self._suggest(sRemain[1:], nMaxDel-1, nMaxHardRepl, nDeep+1, iAddr, sNewWord, "-"+cCurrent, True))
# Phonetic replacements
for sRepl in cp.d1toX.get(cCurrent, ()):
aSugg.update(self._suggest(sRepl + sRemain[1:], nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, True))
aSugg.update(self._suggest(sRepl + sRemain[1:], nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, cCurrent+">"+sRepl, True))
for sRepl in cp.d2toX.get(sRemain[0:2], ()):
aSugg.update(self._suggest(sRepl + sRemain[2:], nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, True))
aSugg.update(self._suggest(sRepl + sRemain[2:], nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, sRemain[0:2]+">"+sRepl, True))
# Hard replacements
if nDeep > 3 and nMaxHardRepl and len(sRemain) >= 2:
for nVal, kAddr in self._getArcs1(iAddr):
if nVal in self.dCharVal:
cChar = self.dCharVal[nVal]
if cChar not in cp.d1to1.get(cCurrent, ""):
aSugg.update(self._suggest(sRemain[1:], nMaxDel, nMaxHardRepl-1, nDeep+1, kAddr, sNewWord+cChar, True))
aSugg.update(self._suggest(sRemain[1:], nMaxDel, nMaxHardRepl-1, nDeep+1, kAddr, sNewWord+cChar, "[["+cChar+"]]", True))
# end of word
if len(sRemain) == 2:
for sRepl in cp.dFinal2.get(sRemain, ()):
aSugg.update(self._suggest(sRepl, nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, True))
aSugg.update(self._suggest(sRepl, nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, sRemain + " >> " + sRepl, True))
elif len(sRemain) == 1:
aSugg.update(self._suggest("", nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, True)) # remove last char and go on
aSugg.update(self._suggest("", nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, sRemain + " [last char removed] ", True)) # remove last char and go on
for sRepl in cp.dFinal1.get(sRemain, ()):
aSugg.update(self._suggest(sRepl, nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, True))
aSugg.update(self._suggest(sRepl, nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, sRemain + " >> " + sRepl, True))
return aSugg
@timethis
def suggest2 (self, sWord, nMaxSugg=10):
"returns a set of suggestions for <sWord>"
sPfx, sWord, sSfx = cp.cut(sWord)
oSuggResult = SuggResult(sWord)
self._suggest2(oSuggResult)
aSugg = oSuggResult.getSuggestions()
if sSfx or sPfx:
# we add what we removed
return list(map(lambda sSug: sPfx + sSug + sSfx, aSugg))
return aSugg
def _suggest2 (self, oSuggResult, nDeep=0, iAddr=0, sNewWord=""):
#logging.info((nDeep * " ") + sNewWord)
if nDeep >= oSuggResult.nDistLimit:
sCleanNewWord = cp.cleanWord(sNewWord)
if st.distanceSift4(oSuggResult.sCleanWord[:len(sCleanNewWord)], sCleanNewWord) > oSuggResult.nDistLimit:
return
if int.from_bytes(self.byDic[iAddr:iAddr+self.nBytesArc], byteorder='big') & self._finalNodeMask:
#logging.info((nDeep * " ") + "__" + sNewWord + "__")
oSuggResult.addSugg(sNewWord, nDeep)
for cChar, jAddr in self._getCharArcs(iAddr):
self._suggest2(oSuggResult, nDeep+1, jAddr, sNewWord+cChar)
return
def _getCharArcs (self, iAddr):
"generator: yield all chars and addresses from node at address <iAddr>"
for nVal, jAddr in self._getArcs(iAddr):
if nVal < self.nChar:
yield (self.dCharVal[nVal], jAddr)
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:
|