119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
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
|
130
131
132
133
134
135
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
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
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
|
-
+
-
+
-
+
-
-
+
+
-
+
-
+
-
-
-
-
-
-
-
+
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
|
"_addrBitMask": self._addrBitMask,
"nBytesOffset": self.nBytesOffset
}, ensure_ascii=False))
if bInJSModule:
hDst.write(";\n\nexports.dictionary = dictionary;\n")
def isValidToken (self, sToken):
"checks if sToken is valid (if there is hyphens in sToken, sToken is split, each part is checked)"
"checks if <sToken> is valid (if there is hyphens in <sToken>, <sToken> is split, each part is checked)"
if self.isValid(sToken):
return True
if "-" in sToken:
if sToken.count("-") > 4:
return True
return all(self.isValid(sWord) for sWord in sToken.split("-"))
return False
def isValid (self, sWord):
"checks if sWord is valid (different casing tested if the first letter is a capital)"
"checks if <sWord> is valid (different casing tested if the first letter is a capital)"
if not sWord:
return None
if "’" in sWord: # ugly hack
sWord = sWord.replace("’", "'")
if self.lookup(sWord):
return True
if sWord[0:1].isupper():
if len(sWord) > 1:
if sWord.istitle():
return bool(self.lookup(sWord.lower()))
return self.lookup(sWord.lower())
if sWord.isupper():
if self.bOptNumSigle:
return True
return bool(self.lookup(sWord.lower()) or self.lookup(sWord.capitalize()))
return bool(self.lookup(sWord[:1].lower() + sWord[1:]))
return self.lookup(sWord.lower()) or self.lookup(sWord.capitalize())
return self.lookup(sWord[:1].lower() + sWord[1:])
else:
return bool(self.lookup(sWord.lower()))
return self.lookup(sWord.lower())
return False
def lookup (self, sWord):
"returns True if sWord in dictionary (strict verification)"
"returns True if <sWord> in dictionary (strict verification)"
iAddr = 0
for c in sWord:
if c not in self.dChar:
return False
iAddr = self._lookupArcNode(self.dChar[c], iAddr)
if iAddr == None:
return False
return int.from_bytes(self.byDic[iAddr:iAddr+self.nBytesArc], byteorder='big') & self._finalNodeMask
def getSugg (self, sWord, iAddr=0, sNewWord=""):
"not finished"
# RECURSIVE FUNCTION
if not sWord:
if int.from_bytes(self.byDic[iAddr:iAddr+self.nBytesArc], byteorder='big') & self._finalNodeMask:
return bool(int.from_bytes(self.byDic[iAddr:iAddr+self.nBytesArc], byteorder='big') & self._finalNodeMask)
return [sNewWord]
return []
lSugg = []
lArc = self._getSimilarArcs(sWord[0:1], iAddr)
if lArc:
for t in lArc:
lSugg.extend(self._lookupAndSuggest(sWord[1:], t[1], sNewWord+t[0]))
else:
pass
return lSugg
def _getSimilarArcs (self, cChar, iAddr):
lArc = []
for c in st.dSimilarChars.get(cChar, cChar):
jAddr = self._lookupArcNode(self.dChar[c], iAddr)
if jAddr:
lArc.append((c, iAddr))
return lArc
def getMorph (self, sWord):
"retrieves morphologies list, different casing allowed"
l = self.morph(sWord)
if sWord[0:1].isupper():
l.extend(self.morph(sWord.lower()))
if sWord.isupper() and len(sWord) > 1:
l.extend(self.morph(sWord.capitalize()))
return l
def suggest (self, sWord, nMaxSugg=10):
"returns a set of suggestions for <sWord>"
# first, we check for similar words
#return self._suggestWithCrushedUselessChars(cp.clearWord(sWord))
aSugg = self._suggest(sWord)
if not aSugg:
aSugg.update(self._suggest(sWord[1:]))
if not aSugg:
aSugg.update(self._suggestWithCrushedUselessChars(cp.clearWord(sWord)))
return sorted(aSugg, key=lambda sSugg: cp.distanceBetweenWords(sWord, sSugg))
def _suggest (self, sRemain, nDeep=0, iAddr=0, sNewWord="", bAvoidLoop=False):
"returns a set of suggestions"
# recursive function
aSugg = set()
if not sRemain:
if int.from_bytes(self.byDic[iAddr:iAddr+self.nBytesArc], byteorder='big') & self._finalNodeMask:
#show(nDeep, "___" + sNewWord + "___")
aSugg.add(sNewWord)
for sTail in self._getTails(iAddr):
aSugg.add(sNewWord+sTail)
return aSugg
#show(nDeep, "<" + sRemain + "> ===> " + sNewWord)
cCurrent = sRemain[0:1]
for cChar, jAddr in self._getSimilarArcs(cCurrent, iAddr):
#show(nDeep, cChar)
aSugg.update(self._suggest(sRemain[1:], nDeep+1, jAddr, sNewWord+cChar))
if not bAvoidLoop: # avoid infinite loop
#show(nDeep, ":no loop:")
if cCurrent == sRemain[1:2]:
# same char, we remove 1 char without adding 1 to <sNewWord>
aSugg.update(self._suggest(sRemain[1:], nDeep+1, iAddr, sNewWord))
else:
# switching chars
aSugg.update(self._suggest(sRemain[1:2]+sRemain[0:1]+sRemain[2:], nDeep+1, iAddr, sNewWord, True))
for sRepl in cp.d1toX.get(cCurrent, ()):
#show(nDeep, sRepl)
aSugg.update(self._suggest(sRepl + sRemain[1:], nDeep+1, iAddr, sNewWord, True))
for sRepl in cp.d2toX.get(sRemain[0:2], ()):
#show(nDeep, sRepl)
aSugg.update(self._suggest(sRepl + sRemain[2:], nDeep+1, iAddr, sNewWord, True))
if len(sRemain) == 2:
for sRepl in cp.dFinal2.get(sRemain, ()):
#show(nDeep, sRepl)
aSugg.update(self._suggest(sRepl, nDeep+1, iAddr, sNewWord, True))
elif len(sRemain) == 1:
#show(nDeep, ":end of word:")
# end of word
aSugg.update(self._suggest("", nDeep+1, iAddr, sNewWord, True)) # remove last char and go on
for sRepl in cp.dFinal1.get(sRemain, ()):
#show(nDeep, sRepl)
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 _getTails (self, iAddr, sTail="", n=2):
"return a list of suffixes ending at a distance of <n> from <iAddr>"
aTails = set()
for nVal, jAddr in self._getArcs(iAddr):
if nVal < self.nChar:
if int.from_bytes(self.byDic[jAddr:jAddr+self.nBytesArc], byteorder='big') & self._finalNodeMask:
aTails.add(sTail + self.dCharVal[nVal])
if n and not aTails:
aTails.update(self._getTails(jAddr, sTail+self.dCharVal[nVal], n-1))
return aTails
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 + " !!!")
aSugg.add(sNewWord)
return aSugg
cCurrent = sWord[0:1]
for cChar, jAddr in self._getSimilarArcsAndCrushedChars(cCurrent, iAddr):
show(nDeep, cChar)
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.aVovels:
yield (self.dCharVal[nVal], jAddr)
yield from self._getSimilarArcs(cChar, iAddr)
def drawPath (self, sWord, iAddr=0):
cChar = sWord[0:1] if sWord else " "
iPos = -1
n = 0
print(cChar + ": ", end="")
for nVal, jAddr in self._getArcs(iAddr):
if nVal in self.dCharVal:
print(self.dCharVal[nVal], end="")
if self.dCharVal[nVal] == sWord[0:1]:
iNextNodeAddr = jAddr
iPos = n
n += 1
if not sWord:
return
if iPos >= 0:
print("\n "+ " " * iPos + "|")
self.drawPath(sWord[1:], iNextNodeAddr)
def select (self, sPattern=""):
"generator: returns all entries which morphology fits <sPattern>"
zPattern = None
try:
zPattern = re.compile(sPattern)
except:
print("# Error in regex pattern")
traceback.print_exc()
yield from self._select1(zPattern, 0, "")
# def morph (self, sWord):
# is defined in __init__
# VERSION 1
def _select1 (self, zPattern, iAddr, sWord):
# recursive generator
for nVal, jAddr in self._getArcs1(iAddr):
if nVal < self.nChar:
# simple character
yield from self._select1(zPattern, jAddr, sWord + self.lArcVal[nVal])
else:
sEntry = sWord + "\t" + self.funcStemming(sWord, self.lArcVal[nVal])
for nMorphVal, _ in self._getArcs1(jAddr):
if not zPattern or zPattern.search(self.lArcVal[nMorphVal]):
yield sEntry + "\t" + self.lArcVal[nMorphVal]
def _morph1 (self, sWord):
"returns morphologies of sWord"
iAddr = 0
for c in sWord:
if c not in self.dChar:
return []
iAddr = self._lookupArcNode(self.dChar[c], iAddr)
|
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
|
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
|
-
+
+
+
+
+
+
+
+
+
+
+
-
+
|
# This value is not a char, this is a stemming code
l.append(self.funcStemming(sWord, self.lArcVal[nArc]))
iAddr = iEndArcAddr+self.nBytesNodeAddress
return l
return []
def _lookupArcNode1 (self, nVal, iAddr):
"looks if nVal is an arc at the node at iAddr, if yes, returns address of next node else None"
"looks if <nVal> is an arc at the node at <iAddr>, if yes, returns address of next node else None"
while True:
iEndArcAddr = iAddr+self.nBytesArc
nRawArc = int.from_bytes(self.byDic[iAddr:iEndArcAddr], byteorder='big')
if nVal == (nRawArc & self._arcMask):
# the value we are looking for
# we return the address of the next node
return int.from_bytes(self.byDic[iEndArcAddr:iEndArcAddr+self.nBytesNodeAddress], byteorder='big')
else:
# value not found
if (nRawArc & self._lastArcMask):
return None
iAddr = iEndArcAddr+self.nBytesNodeAddress
def _getArcs1 (self, iAddr):
"generator: return all arcs at <iAddr> as tuples of (nVal, iAddr)"
while True:
iEndArcAddr = iAddr+self.nBytesArc
nRawArc = int.from_bytes(self.byDic[iAddr:iEndArcAddr], byteorder='big')
yield (nRawArc & self._arcMask, int.from_bytes(self.byDic[iEndArcAddr:iEndArcAddr+self.nBytesNodeAddress], byteorder='big'))
if (nRawArc & self._lastArcMask):
break
iAddr = iEndArcAddr+self.nBytesNodeAddress
def _writeNodes1 (self, spfDest):
"for debugging only"
print(" > Write binary nodes")
with codecs.open(spfDest, 'w', 'utf-8', newline="\n") as hDst:
iAddr = 0
hDst.write("i{:_>10} -- #{:_>10}\n".format("0", iAddr))
while iAddr < len(self.byDic):
iEndArcAddr = iAddr+self.nBytesArc
nRawArc = int.from_bytes(self.byDic[iAddr:iEndArcAddr], byteorder='big')
nArc = nRawArc & self._arcMask
hDst.write(" {:<20} {:0>16} i{:>10} #{:_>10}\n".format(self.lArcVal[nArc], bin(nRawArc)[2:], "?", \
int.from_bytes(self.byDic[iEndArcAddr:iEndArcAddr+self.nBytesNodeAddress], \
byteorder='big')))
iAddr = iEndArcAddr+self.nBytesNodeAddress
if (nRawArc & self._lastArcMask) and iAddr < len(self.byDic):
hDst.write("\ni{:_>10} -- #{:_>10}\n".format("?", iAddr))
hDst.close()
# VERSION 2
def _morph2 (self, sWord):
"returns morphologies of sWord"
"returns morphologies of <sWord>"
iAddr = 0
for c in sWord:
if c not in self.dChar:
return []
iAddr = self._lookupArcNode(self.dChar[c], iAddr)
if iAddr == None:
return []
|