Grammalecte  Diff

Differences From Artifact [a0b40e49a1]:

To Artifact [dc355a8e9b]:


1
2
3
4
5
6
7
8
9
10
11
12
13
14
1
2
3
4
5
6

7
8
9
10
11
12
13






-







#!python3

import os
import traceback
import pkgutil
import re
from itertools import chain
from functools import wraps
import time

#import logging
#logging.basicConfig(filename="suggestions.log", level=logging.DEBUG)

from . import str_transform as st
23
24
25
26
27
28
29















































30
31
32
33
34
35
36
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
72
73
74
75
76
77
78
79
80
81
82







+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+







        fStart = time.time()
        result = func(*args, **kwargs)
        fEnd = time.time()
        print(func.__name__, fEnd - fStart)
        return result
    return wrapper


class SuggResult:
    """Structure for storing, classifying and filtering suggestions"""

    def __init__ (self, sWord, nDistLimit=-1):
        self.sWord = sWord
        self.sCleanWord = cp.cleanWord(sWord)
        self.nDistLimit = nDistLimit  if nDistLimit >= 0  else  (len(sWord) // 3) + 1
        self.nMinDist = 1000
        self.aSugg = set()
        self.dSugg = { 0: [],  1: [] }

    def addSugg (self, sSugg, nDeep=0):
        "add a suggestion"
        #logging.info((nDeep * "  ") + "__" + sSugg + "__")
        if sSugg not in self.aSugg:
            nDist = st.distanceDamerauLevenshtein(self.sCleanWord, cp.cleanWord(sSugg))
            if nDist <= self.nDistLimit:
                if nDist not in self.dSugg:
                    self.dSugg[nDist] = []
                self.dSugg[nDist].append(sSugg)
                self.aSugg.add(sSugg)
                if nDist < self.nMinDist:
                    self.nMinDist = nDist
                self.nDistLimit = min(self.nDistLimit, self.nMinDist+2)

    def getSuggestions (self, nSuggLimit=10, nDistLimit=-1):
        "return a list of suggestions"
        lRes = []
        if self.dSugg[0]:
            # we sort the better results with the original word
            self.dSugg[0].sort(key=lambda sSugg: st.distanceDamerauLevenshtein(self.sWord, sSugg))
        for lSugg in self.dSugg.values():
            lRes.extend(lSugg)
            if len(lRes) > nSuggLimit:
                break
        lRes = list(cp.filterSugg(lRes))
        if self.sWord.istitle():
            lRes = list(map(lambda sSugg: sSugg.title(), lRes))
        elif self.sWord.isupper():
            lRes = list(map(lambda sSugg: sSugg.upper(), lRes))
        return lRes[:nSuggLimit]

    def reset (self):
        self.aSugg.clear()
        self.dSugg.clear()


class IBDAWG:
    """INDEXABLE BINARY DIRECT ACYCLIC WORD GRAPH"""

    def __init__ (self, sDicName):
        self.by = pkgutil.get_data(__package__, "_dictionaries/" + sDicName)
        if not self.by:
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
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



331
332





333
334
335
336







337














338
339
340











341

342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369

370
371
372


373
374






375
376
377
378
379
380
381
382
383
384
385
386







-
+



+


+
-
+

-
+
-

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





-
+
-

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

-
+






+
+
+
+
+
+
+














-
+


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







        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

    @timethis
    #@timethis
    def suggest (self, sWord, nMaxSugg=10):
        "returns a set of suggestions for <sWord>"
        sPfx, sWord, sSfx = cp.cut(sWord)
        nMaxSwitch = max(len(sWord) // 3, 1)
        nMaxDel = len(sWord) // 5
        nMaxHardRepl = max((len(sWord) - 5) // 4, 1)
        oSuggResult = SuggResult(sWord)
        aSugg = self._suggest(sWord, nMaxDel=nMaxDel, nMaxHardRepl=nMaxHardRepl)
        self._suggest(oSuggResult, sWord, nMaxSwitch=nMaxSwitch, nMaxDel=nMaxDel, nMaxHardRepl=nMaxHardRepl)
        if sWord.istitle():
            aSugg.update(self._suggest(sWord.lower(), nMaxDel=nMaxDel, nMaxHardRepl=nMaxHardRepl))
            self._suggest(oSuggResult, sWord.lower(), nMaxSwitch=nMaxSwitch, 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]
            self._suggest(oSuggResult, sWord.title(), nMaxSwitch=nMaxSwitch, nMaxDel=nMaxDel, nMaxHardRepl=nMaxHardRepl)
        aSugg = oSuggResult.getSuggestions()
        if sSfx or sPfx:
            # we add what we removed
            return list(map(lambda sSug: sPfx + sSug + sSfx, aSugg))
        return aSugg

    def _suggest (self, oSuggResult, sRemain, nMaxSwitch=0, nMaxDel=0, nMaxHardRepl=0, nDeep=0, iAddr=0, sNewWord="", sAction="", bAvoidLoop=False):
        # recursive function
        #logging.info((nDeep * "  ") + sNewWord + ":" + sRemain + " · " + sAction)
        if not sRemain:
            if int.from_bytes(self.byDic[iAddr:iAddr+self.nBytesArc], byteorder='big') & self._finalNodeMask:
                oSuggResult.addSugg(sNewWord, nDeep)
            for sTail in self._getTails(iAddr):
                oSuggResult.addSugg(sNewWord+sTail, nDeep)
            return
        cCurrent = sRemain[0:1]
        for cChar, jAddr in self._getSimilarCharArcs(cCurrent, iAddr):
            self._suggest(oSuggResult, sRemain[1:], nMaxSwitch, nMaxDel, nMaxHardRepl, nDeep+1, jAddr, sNewWord+cChar, "*")
        if not bAvoidLoop: # avoid infinite loop
            if len(sRemain) > 1:
                if cCurrent == sRemain[1:2]:
                    # same char, we remove 1 char without adding 1 to <sNewWord>
                    self._suggest(oSuggResult, sRemain[1:], nMaxSwitch, nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, cCurrent+"/2")
                else:
                    # switching chars
                    if nMaxSwitch:
                        self._suggest(oSuggResult, sRemain[1:2]+sRemain[0:1]+sRemain[2:], nMaxSwitch-1, nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, "><",True)
                    # delete char
                    if nMaxDel:
                        self._suggest(oSuggResult, sRemain[1:], nMaxSwitch, nMaxDel-1, nMaxHardRepl, nDeep+1, iAddr, sNewWord, "-"+cCurrent, True)
                # Phonetic replacements
                for sRepl in cp.d1toX.get(cCurrent, ()):
                    self._suggest(oSuggResult, sRepl + sRemain[1:], nMaxSwitch, nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, cCurrent+">"+sRepl, True)
                for sRepl in cp.d2toX.get(sRemain[0:2], ()):
                    self._suggest(oSuggResult, sRepl + sRemain[2:], nMaxSwitch, nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, sRemain[0:2]+">"+sRepl, True)
                # Hard replacements
                if nDeep > 3 and nMaxHardRepl:
                    for cChar, kAddr in self._getCharArcs(iAddr):
                        if cChar not in cp.d1to1.get(cCurrent, ""):
                            self._suggest(oSuggResult, sRemain[1:], nMaxSwitch, nMaxDel, nMaxHardRepl-1, nDeep+1, kAddr, sNewWord+cChar, "[["+cChar+"]]", True)
            # end of word
            if len(sRemain) == 2:
                for sRepl in cp.dFinal2.get(sRemain, ()):
                    self._suggest(oSuggResult, sRepl, nMaxSwitch, nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, sRemain + " >> " + sRepl, True)
            elif len(sRemain) == 1:
                self._suggest(oSuggResult, "", nMaxSwitch, nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, sRemain + " [last char removed] ", True) # remove last char and go on
                for sRepl in cp.dFinal1.get(sRemain, ()):
                    self._suggest(oSuggResult, sRepl, nMaxSwitch, nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, sRemain + " >> " + sRepl, True)

    #@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 _suggest (self, sRemain, nMaxDel=0, nMaxHardRepl=0, nDeep=0, iAddr=0, sNewWord="", bAvoidLoop=False):
    def _suggest2 (self, oSuggResult, nDeep=0, iAddr=0, sNewWord=""):
        "returns a set of suggestions"
        # recursive function
        #logging.info((nDeep * "  ") + sNewWord + ":" + sRemain)
        aSugg = set()
        if not sRemain:
            if int.from_bytes(self.byDic[iAddr:iAddr+self.nBytesArc], byteorder='big') & self._finalNodeMask:
        #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 + "__")
                aSugg.add(sNewWord)
            for sTail in self._getTails(iAddr):
            oSuggResult.addSugg(sNewWord, nDeep)
        for cChar, jAddr in self._getCharArcsWithPriority(iAddr, oSuggResult.sWord[nDeep:nDeep+1]):
                #logging.info((nDeep * "  ") + "__" + sNewWord+sTail + "__")
                aSugg.add(sNewWord+sTail)
            return aSugg
        cCurrent = sRemain[0:1]
        for cChar, jAddr in self._getSimilarArcs(cCurrent, iAddr):
            self._suggest2(oSuggResult, nDeep+1, jAddr, sNewWord+cChar)
        return

    def _getCharArcs (self, iAddr):
            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))
            else:
                # switching chars
        "generator: yield all chars and addresses from node at address <iAddr>"
                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))
            # Phonetic replacements
            for sRepl in cp.d1toX.get(cCurrent, ()):
                aSugg.update(self._suggest(sRepl + sRemain[1:], nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, True))
            for sRepl in cp.d2toX.get(sRemain[0:2], ()):
                aSugg.update(self._suggest(sRepl + sRemain[2:], nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, 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]
        for nVal, jAddr in self._getArcs(iAddr):
            if nVal < self.nChar:
                yield (self.dCharVal[nVal], jAddr)
                        if cChar not in cp.d1to1.get(cCurrent, ""):
                            aSugg.update(self._suggest(sRemain[1:], nMaxDel, nMaxHardRepl-1, nDeep+1, kAddr, sNewWord+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))
            elif len(sRemain) == 1:
                aSugg.update(self._suggest("", nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, 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))
        return aSugg

    def _getSimilarArcs (self, cChar, iAddr):
    def _getSimilarCharArcs (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 _getCharArcsWithPriority (self, iAddr, cChar):
        if not cChar:
            yield from self._getCharArcs(iAddr)
        lTuple = list(self._getCharArcs(iAddr))
        lTuple.sort(key=lambda t: 0  if t[0] in cp.d1to1.get(cChar, cChar)  else  1)
        yield from lTuple

    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 drawPath (self, sWord, iAddr=0):
        "show the path taken by <sWord> in the graph"
        cChar = sWord[0:1]  if sWord  else " "
        c1 = sWord[0:1]  if sWord  else " "
        iPos = -1
        n = 0
        print(cChar + ": ", end="")
        for nVal, jAddr in self._getArcs(iAddr):
        print(c1 + ": ", end="")
        for c2, jAddr in self._getCharArcs(iAddr):
            if nVal in self.dCharVal:
                print(self.dCharVal[nVal], end="")
                if self.dCharVal[nVal] == sWord[0:1]:
                    iNextNodeAddr = jAddr
                    iPos = n
                n += 1
            print(c2, end="")
            if c2 == 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=""):