Grammalecte  Check-in [3348432a36]

Overview
Comment:[graphspell] normalize characters before spell checking
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | trunk | graphspell
Files: files | file ages | folders
SHA3-256: 3348432a36ef9de5c9093131397ae5ebd3d5b151fb108f44812abcbb1920f8e2
User & Date: olr on 2018-02-19 09:07:58
Other Links: manifest | tags
Context
2018-02-19
09:11
[fr] version 0.6.2 check-in: 18027d1022 user: olr tags: trunk, fr
09:07
[graphspell] normalize characters before spell checking check-in: 3348432a36 user: olr tags: trunk, graphspell
07:30
[fr] fichier des affixes: ICONV ? -> s, REP: suppression de cad -> c’est-à-dire check-in: b9474152b2 user: olr tags: trunk, fr
Changes

Modified gc_lang/fr/perf_memo.txt from [cec037999d] to [190717e473].

19
20
21
22
23
24
25

19
20
21
22
23
24
25
26







+
0.5.12      2016.10.14 18:58    4.51895     1.0843      0.772805    0.22387     0.249411    0.261593    0.628802    0.339303    0.0570326   0.00805416  
0.5.15      2017.01.22 11:44    4.85204     1.16134     0.770762    0.227874    0.244574    0.253305    0.58831     0.319987    0.0603996   0.00694786  
0.5.15      2017.01.22 11:47    4.85593     1.15248     0.762924    0.22744     0.243461    0.254609    0.586741    0.317503    0.0588827   0.00701016  (unicode normalisation NFC)
0.5.15      2017.01.31 12:06    4.88227     1.18008     0.782217    0.232617    0.247672    0.257628    0.596903    0.32169     0.0603505   0.00695196  
0.5.15      2017.02.05 10:10    4.90222     1.18444     0.786696    0.233413    0.25071     0.260214    0.602112    0.325235    0.0609932   0.00706897  
0.5.16      2017.05.12 07:41    4.92201     1.19269     0.80639     0.239147    0.257518    0.266523    0.62111     0.33359     0.0634668   0.00757178  
0.6.1       2018.02.12 09:58    5.25924     1.2649      0.878442    0.257465    0.280558    0.293903    0.686887    0.391275    0.0672474   0.00824723  
0.6.1       2018.02.19 09:06    6.20116     1.44334     1.02936     0.272956    0.311561    0.362367    0.812705    0.419061    0.0773003   0.00845671  (spelling normalization)

Modified graphspell-js/char_player.js from [c9b14a8774] to [c171c18615].

1
2
3
4
5
6
7
8
9













10
11
12
13
14
15
16

17
18
19
20
21
22
23
24
25

26
27

28
29
30
31
32
33
34
1
2
3
4
5
6
7
8

9
10
11
12
13
14
15
16
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








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







+








-
+

-
+







// list of similar chars
// useful for suggestion mechanism

${map}


var char_player = {

    _dTransChars: new Map([
    _xTransCharsForSpelling: new Map([
        ['ſ', 's'],  ['ffi', 'ffi'],  ['ffl', 'ffl'],  ['ff', 'ff'],  ['ſt', 'ft'],  ['fi', 'fi'],  ['fl', 'fl'],  ['st', 'st']
    ]),

    spellingNormalization: function (sWord) {
        let sNewWord = "";
        for (let c of sWord) {
            sNewWord += this._xTransCharsForSpelling.gl_get(c, c);
        }
        return sNewWord.normalize("NFC");
    },

    _xTransCharsForSimplification: new Map([
        ['à', 'a'],  ['é', 'e'],  ['î', 'i'],  ['ô', 'o'],  ['û', 'u'],  ['ÿ', 'i'],  ['y', 'i'],
        ['â', 'a'],  ['è', 'e'],  ['ï', 'i'],  ['ö', 'o'],  ['ù', 'u'],  ['ŷ', 'i'],
        ['ä', 'a'],  ['ê', 'e'],  ['í', 'i'],  ['ó', 'o'],  ['ü', 'u'],  ['ý', 'i'],
        ['á', 'a'],  ['ë', 'e'],  ['ì', 'i'],  ['ò', 'o'],  ['ú', 'u'],  ['ỳ', 'i'],
        ['ā', 'a'],  ['ē', 'e'],  ['ī', 'i'],  ['ō', 'o'],  ['ū', 'u'],  ['ȳ', 'i'],
        ['ñ', 'n'],  ['k', 'q'],  ['w', 'v'],
        ['œ', 'oe'], ['æ', 'ae'], 
        ['ſ', 's'],  ['ffi', 'ffi'],  ['ffl', 'ffl'],  ['ff', 'ff'],  ['ſt', 'ft'],  ['fi', 'fi'],  ['fl', 'fl'],  ['st', 'st']
    ]),

    simplifyWord: function (sWord) {
        // word simplication before calculating distance between words
        sWord = sWord.toLowerCase();
        let sNewWord = "";
        let i = 1;
        for (let c of sWord) {
            let cNew = this._dTransChars.gl_get(c, c);
            let cNew = this._xTransCharsForSimplification.gl_get(c, c);
            let cNext = sWord.slice(i, i+1)
            if (cNew != this._dTransChars.gl_get(cNext, cNext)) {
            if (cNew != this._xTransCharsForSimplification.gl_get(cNext, cNext)) {
                sNewWord += cNew;
            }
            i++;
        }
        return sNewWord.replace(/eau/g, "o").replace(/au/g, "o").replace(/ai/g, "e").replace(/ei/g, "e").replace(/ph/g, "f");
    },

Modified graphspell-js/ibdawg.js from [08ad598b63] to [73e27f350e].

206
207
208
209
210
211
212

213
214
215
216
217
218
219
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220







+







            "sByDic": this.sByDic    // binary word graph
        };
        return oJSON;
    }

    isValidToken (sToken) {
        // checks if sToken is valid (if there is hyphens in sToken, sToken is split, each part is checked)
        sToken = char_player.spellingNormalization(sToken)
        if (this.isValid(sToken)) {
            return true;
        }
        if (sToken.includes("-")) {
            if (sToken.gl_count("-") > 4) {
                return true;
            }
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
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







+












+







            }
        }
        return Boolean(this._convBytesToInteger(this.byDic.slice(iAddr, iAddr+this.nBytesArc)) & this._finalNodeMask);
    }

    getMorph (sWord) {
        // retrieves morphologies list, different casing allowed
        sWord = char_player.spellingNormalization(sWord)
        let l = this.morph(sWord);
        if (sWord[0].gl_isUpperCase()) {
            l.push(...this.morph(sWord.toLowerCase()));
            if (sWord.gl_isUpperCase() && sWord.length > 1) {
                l.push(...this.morph(sWord.gl_toCapitalize()));
            }
        }
        return l;
    }

    suggest (sWord, nSuggLimit=10) {
        // returns a array of suggestions for <sWord>
        sWord = char_player.spellingNormalization(sWord)
        let sPfx = "";
        let sSfx = "";
        [sPfx, sWord, sSfx] = char_player.cut(sWord);
        let nMaxSwitch = Math.max(Math.floor(sWord.length / 3), 1);
        let nMaxDel = Math.floor(sWord.length / 5);
        let nMaxHardRepl = Math.max(Math.floor((sWord.length - 5) / 4), 1);
        let oSuggResult = new SuggResult(sWord);

Modified graphspell/char_player.py from [82e97eae54] to [e841b9211a].

1
2
3
4

5
6
7









8
9
10
11
12
13
14


15
16
17
18
19

20
21
22
23
24
25
26
1
2
3
4
5
6
7

8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

23
24
25
26
27
28

29
30
31
32
33
34
35
36




+


-
+
+
+
+
+
+
+
+
+






-
+
+




-
+







# list of similar chars
# useful for suggestion mechanism

import re
import unicodedata


_xTransChars = str.maketrans({
_xTransCharsForSpelling = str.maketrans({
    'ſ': 's',  'ffi': 'ffi',  'ffl': 'ffl',  'ff': 'ff',  'ſt': 'ft',  'fi': 'fi',  'fl': 'fl',  'st': 'st'
})

def spellingNormalization (sWord):
    return unicodedata.normalize("NFC", sWord.translate(_xTransCharsForSpelling))


_xTransCharsForSimplification = str.maketrans({
    'à': 'a',  'é': 'e',  'î': 'i',  'ô': 'o',  'û': 'u',  'ÿ': 'i',  "y": "i",
    'â': 'a',  'è': 'e',  'ï': 'i',  'ö': 'o',  'ù': 'u',  'ŷ': 'i',
    'ä': 'a',  'ê': 'e',  'í': 'i',  'ó': 'o',  'ü': 'u',  'ý': 'i',
    'á': 'a',  'ë': 'e',  'ì': 'i',  'ò': 'o',  'ú': 'u',  'ỳ': 'i',
    'ā': 'a',  'ē': 'e',  'ī': 'i',  'ō': 'o',  'ū': 'u',  'ȳ': 'i',
    'ñ': 'n',  'k': 'q',  'w': 'v',
    'œ': 'oe',  'æ': 'ae', 
    'œ': 'oe',  'æ': 'ae',
    'ſ': 's',  'ffi': 'ffi',  'ffl': 'ffl',  'ff': 'ff',  'ſt': 'ft',  'fi': 'fi',  'fl': 'fl',  'st': 'st', 
})

def simplifyWord (sWord):
    "word simplication before calculating distance between words"
    sWord = sWord.lower().translate(_xTransChars)
    sWord = sWord.lower().translate(_xTransCharsForSimplification)
    sNewWord = ""
    for i, c in enumerate(sWord, 1):
        if c != sWord[i:i+1]:
            sNewWord += c
    return sNewWord.replace("eau", "o").replace("au", "o").replace("ai", "e").replace("ei", "e").replace("ph", "f")


Modified graphspell/ibdawg.py from [3bf18d8144] to [c41b426a86].

214
215
216
217
218
219
220

221
222
223
224
225
226
227
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228







+







                            "sByDic": self.byDic.hex()  if bBinaryDictAsHexString  else [ e  for e in self.byDic ]
                        }, 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)"
        sToken = cp.spellingNormalization(sToken)
        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
256
257
258
259
260
261
262

263
264
265
266
267
268
269
270
271
272

273
274
275
276
277
278
279
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







+










+







            iAddr = self._lookupArcNode(self.dChar[c], iAddr)
            if iAddr == None:
                return False
        return bool(int.from_bytes(self.byDic[iAddr:iAddr+self.nBytesArc], byteorder='big') & self._finalNodeMask)

    def getMorph (self, sWord):
        "retrieves morphologies list, different casing allowed"
        sWord = cp.spellingNormalization(sWord)
        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
    def suggest (self, sWord, nSuggLimit=10):
        "returns a set of suggestions for <sWord>"
        sWord = cp.spellingNormalization(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)
        self._suggest(oSuggResult, sWord, nMaxSwitch=nMaxSwitch, nMaxDel=nMaxDel, nMaxHardRepl=nMaxHardRepl)
        if sWord.istitle():
326
327
328
329
330
331
332

333
334
335
336
337
338
339
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343







+







                self._suggest(oSuggResult, "", nMaxSwitch, nMaxDel, nMaxHardRepl, nDeep+1, iAddr, sNewWord, 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, True)

    #@timethis
    def suggest2 (self, sWord, nMaxSugg=10):
        "returns a set of suggestions for <sWord>"
        sWord = cp.spellingNormalization(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))
382
383
384
385
386
387
388

389
390
391
392
393
394
395
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400







+







                    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"
        sWord = cp.spellingNormalization(sWord)
        c1 = sWord[0:1]  if sWord  else " "
        iPos = -1
        n = 0
        print(c1 + ": ", end="")
        for c2, jAddr in self._getCharArcs(iAddr):
            print(c2, end="")
            if c2 == sWord[0:1]: