Grammalecte  Check-in [eeef098bd9]

Overview
Comment:[build][core] graph calls within regex rules
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | core | build | rg
Files: files | file ages | folders
SHA3-256: eeef098bd9f125077bc432ade4b12eed176ae7dc4d6e139b3987bbfc14bb16b7
User & Date: olr on 2018-06-11 13:21:42
Other Links: branch diff | manifest | tags
Context
2018-06-11
18:20
[misc] SublimeText syntax rules update check-in: cac524f6db user: olr tags: misc, rg
13:21
[build][core] graph calls within regex rules check-in: eeef098bd9 user: olr tags: core, build, rg
09:26
[build][core] named graphs check-in: 7e92a17d42 user: olr tags: core, build, rg
Changes
Hide Diffs Side-by-Side Diffs Ignore Whitespace Patch

Modified compile_rules.py from [394c512707] to [8c11a24e22].

︙
108
109
110
111
112
113
114
115
116
117









118
119
120
121
122
123
124

108
109
110
111
112
113
114

115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132








-


+
+
+
+
+
+
+
+
+










def createRule (s, nIdLine, sLang, bParagraph, dOptPriority):
    "returns rule as list [option name, regex, bCaseInsensitive, identifier, list of actions]"
    global dJSREGEXES
    global nRULEWITHOUTNAME

    #### OPTIONS
    sLineId = str(nIdLine) + ("p" if bParagraph else "s")
    sRuleId = sLineId

    #### GRAPH CALL
    if s.startswith("@@@@"):
        if bParagraph:
            print("Error. Graph call can’t be made only after the first pass (sentence by sentence)")
            exit()
        return ["@@@@", s[4:], sLineId]

    #### OPTIONS
    sOption = False         # False or [a-z0-9]+ name
    nPriority = 4           # Default is 4, value must be between 0 and 9
    tGroups = None          # code for groups positioning (only useful for JavaScript)
    cCaseMode = 'i'         # i: case insensitive,  s: case sensitive,  u: uppercasing allowed
    cWordLimitLeft = '['    # [: word limit, <: no specific limit
    cWordLimitRight = ']'   # ]: word limit, >: no specific limit
    m = re.match("^__(?P<borders_and_case>[[<]\\w[]>])(?P<option>/[a-zA-Z0-9]+|)(?P<ruleid>\\(\\w+\\)|)(?P<priority>![0-9]|)__ *", s)
︙
341
342
343
344
345
346
347

348
349


350
351
352
353
354
355
356

349
350
351
352
353
354
355
356


357
358
359
360
361
362
363
364
365








+
-
-
+
+








        print("# Unknown action at line " + sIdAction)
        return None


def _calcRulesStats (lRules):
    d = {'=':0, '~': 0, '-': 0, '>': 0}
    for aRule in lRules:
        if aRule[0] != "@@@@":
        for aAction in aRule[6]:
            d[aAction[1]] = d[aAction[1]] + 1
            for aAction in aRule[6]:
                d[aAction[1]] = d[aAction[1]] + 1
    return (d, len(lRules))


def displayStats (lParagraphRules, lSentenceRules):
    print("  {:>18} {:>18} {:>18} {:>18}".format("DISAMBIGUATOR", "TEXT PROCESSOR", "GRAMMAR CHECKING", "REGEX"))
    d, nRule = _calcRulesStats(lParagraphRules)
    print("§ {:>10} actions {:>10} actions {:>10} actions  in {:>8} rules".format(d['='], d['~'], d['-'], nRule))
︙
434
435
436
437
438
439
440
441
442
443
444
445

446
447
448

449
450

451

452
453
454


455

456
457
458
459
460
461
462

463
464

465
466

467
468

469
470
471


472
473
474
475

476

477

478
479
480
481
482
483
484

443
444
445
446
447
448
449


450
451
452
453
454
455
456
457
458
459
460

461
462
463

464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486

487
488
489
490
491
492
493

494
495
496
497
498
499
500
501
502
503








-
-



+



+


+
-
+


-
+
+

+







+


+


+


+


-
+
+




+
-
+

+








    # removing comments, zeroing empty lines, creating definitions, storing tests, merging rule lines
    print("  parsing rules...")
    global dDEF
    lLine = []
    lRuleLine = []
    lTest = []
    lOpt = []
    zBookmark = re.compile("^!!+")
    zGraphLink = re.compile(r"^@@@@GRAPHLINK>(\w+)@@@@")

    for i, sLine in enumerate(lRules, 1):
        if sLine.startswith('#END'):
            # arbitrary end
            printBookmark(0, "BREAK BY #END", i)
            break
        elif sLine.startswith("#"):
            # comment
            pass
        elif sLine.startswith("@@@@"):
            # rules graph call
            m = re.match(r"^@@@@GRAPHLINK>(\w+)@@@@", sLine.strip())
            m = re.match(r"@@@@GRAPH: *(\w+)@@@@", sLine.strip())
            if m:
                #lRuleLine.append(["@GRAPHLINK", m.group(1)])
                printBookmark(1, "@GRAPHLINK: " + m.group(1), i)
                printBookmark(1, "@GRAPH: " + m.group(1), i)
                lRuleLine.append([i, "@@@@"+m.group(1)])
        elif sLine.startswith("DEF:"):
            # definition
            m = re.match("DEF: +([a-zA-Z_][a-zA-Z_0-9]*) +(.+)$", sLine.strip())
            if m:
                dDEF["{"+m.group(1)+"}"] = m.group(2)
            else:
                print("Error in definition: ", end="")
                print(sLine.strip())
        elif sLine.startswith("TEST:"):
            # test
            lTest.append("r{:<7}".format(i) + "  " + sLine[5:].strip())
        elif sLine.startswith("TODO:"):
            # todo
            pass
        elif sLine.startswith(("OPTGROUP/", "OPTSOFTWARE:", "OPT/", "OPTLANG/", "OPTDEFAULTUILANG:", "OPTLABEL/", "OPTPRIORITY/")):
            # options
            lOpt.append(sLine)
        elif re.match("[  \t]*$", sLine):
            # empty line
            pass
        elif sLine.startswith("!!"):
            m = zBookmark.search(sLine)
            # bookmark
            m = re.match("!!+", sLine)
            nExMk = len(m.group(0))
            if sLine[nExMk:].strip():
                printBookmark(nExMk-2, sLine[nExMk:].strip(), i)
        elif sLine.startswith(("    ", "\t")):
            # rule (continuation)
            lRuleLine[len(lRuleLine)-1][1] += " " + sLine.strip()
            lRuleLine[-1][1] += " " + sLine.strip()
        else:
            # new rule
            lRuleLine.append([i, sLine.strip()])

    # generating options files
    print("  parsing options...")
    try:
        dOptions, dOptPriority = prepareOptions(lOpt)
    except:
︙

Modified compile_rules_js_convert.py from [5ad87f3f46] to [f2cc9f3e39].

︙
115
116
117
118
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

115
116
117
118
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








+
+
+















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










    if not lNegLookBeforeRegex:
        lNegLookBeforeRegex = None
    return (sRegex, lNegLookBeforeRegex)


def pyRuleToJS (lRule, dJSREGEXES, sWORDLIMITLEFT):
    lRuleJS = copy.deepcopy(lRule)
    # graph rules
    if lRuleJS[0] == "@@@@":
        return lRuleJS
    del lRule[-1] # tGroups positioning codes are useless for Python
    # error messages
    for aAction in lRuleJS[6]:
        if aAction[1] == "-":
            aAction[2] = aAction[2].replace(" ", " ") # nbsp --> nnbsp
            aAction[4] = aAction[4].replace("« ", "« ").replace(" »", " »").replace(" :", " :").replace(" :", " :")
    # js regexes
    lRuleJS[1], lNegLookBehindRegex = regex2js(dJSREGEXES.get(lRuleJS[3], lRuleJS[1]), sWORDLIMITLEFT)
    lRuleJS.append(lNegLookBehindRegex)
    return lRuleJS


def writeRulesToJSArray (lRules):
    sArray = "[\n"
    for sOption, aRuleGroup in lRules:
        if sOption != "@@@@":
        sArray += '  ["' + sOption + '", [\n'  if sOption  else  "  [false, [\n"
        for sRegex, bCaseInsensitive, sLineId, sRuleId, nPriority, lActions, aGroups, aNegLookBehindRegex in aRuleGroup:
            sArray += '    [' + sRegex + ", "
            sArray += "true, " if bCaseInsensitive  else "false, "
            sArray += '"' + sLineId + '", '
            sArray += '"' + sRuleId + '", '
            sArray += str(nPriority) + ", "
            sArray += json.dumps(lActions, ensure_ascii=False) + ", "
            sArray += json.dumps(aGroups, ensure_ascii=False) + ", "
            sArray += json.dumps(aNegLookBehindRegex, ensure_ascii=False) + "],\n"
        sArray += "  ]],\n"
            sArray += '  ["' + sOption + '", [\n'  if sOption  else  "  [false, [\n"
            for sRegex, bCaseInsensitive, sLineId, sRuleId, nPriority, lActions, aGroups, aNegLookBehindRegex in aRuleGroup:
                sArray += '    [' + sRegex + ", "
                sArray += "true, " if bCaseInsensitive  else "false, "
                sArray += '"' + sLineId + '", '
                sArray += '"' + sRuleId + '", '
                sArray += str(nPriority) + ", "
                sArray += json.dumps(lActions, ensure_ascii=False) + ", "
                sArray += json.dumps(aGroups, ensure_ascii=False) + ", "
                sArray += json.dumps(aNegLookBehindRegex, ensure_ascii=False) + "],\n"
            sArray += "  ]],\n"
        else:
            sArray += '  ["' + sOption + '", [\n'
            for sGraphName, sLineId in aRuleGroup:
                sArray += '    ["' + sGraphName + '", "' + sLineId + '"],\n"'
            sArray += "  ]],\n"
    sArray += "]"
    return sArray


def groupsPositioningCodeToList (sGroupsPositioningCode):
    if not sGroupsPositioningCode:
        return None
    return [ int(sCode)  if sCode.isdigit() or (sCode[0:1] == "-" and sCode[1:].isdigit())  else sCode \
             for sCode in sGroupsPositioningCode.split(",") ]

Modified gc_core/py/lang_core/gc_engine.py from [12b89317cd] to [0da090b219].

︙
87
88
89
90
91
92
93
94
95
96
97
98
99
100








101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117

118
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

87
88
89
90
91
92
93







94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117

118
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








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
















-
+




















-
-
-
-

-
+

-
-
-
-
















-
+



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










def _loadRules ():
    from . import gc_rules
    global _rules
    _rules = gc_rules
    # compile rules regex
    for lRuleGroup in chain(_rules.lParagraphRules, _rules.lSentenceRules):
        for rule in lRuleGroup[1]:
            try:
                rule[0] = re.compile(rule[0])
            except:
                echo("Bad regular expression in # " + str(rule[2]))
                rule[0] = "(?i)<Grammalecte>"
    for sOption, lRuleGroup in chain(_rules.lParagraphRules, _rules.lSentenceRules):
        if sOption != "@@@@":
            for aRule in lRuleGroup:
                try:
                    aRule[0] = re.compile(aRule[0])
                except:
                    echo("Bad regular expression in # " + str(aRule[2]))
                    aRule[0] = "(?i)<Grammalecte>"


#### Parsing

def parse (sText, sCountry="${country_default}", bDebug=False, dOptions=None, bContext=False):
    "analyses the paragraph sText and returns list of errors"
    #sText = unicodedata.normalize("NFC", sText)
    aErrors = None
    sRealText = sText
    dDA = {}        # Disambiguisator. Key = position; value = list of morphologies
    dPriority = {}  # Key = position; value = priority
    dOpt = _dOptions  if not dOptions  else dOptions
    bShowRuleId = option('idrule')

    # parse paragraph
    try:
        sNew, aErrors = _proofread(sText, sRealText, 0, True, dDA, dPriority, sCountry, dOpt, bShowRuleId, bDebug, bContext)
        sNew, aErrors = _proofread(None, sText, sRealText, 0, True, dDA, dPriority, sCountry, dOpt, bShowRuleId, bDebug, bContext)
        if sNew:
            sText = sNew
    except:
        raise

    # cleanup
    if " " in sText:
        sText = sText.replace(" ", ' ') # nbsp
    if " " in sText:
        sText = sText.replace(" ", ' ') # nnbsp
    if "'" in sText:
        sText = sText.replace("'", "’")
    if "‑" in sText:
        sText = sText.replace("‑", "-") # nobreakdash

    # parse sentences
    for iStart, iEnd in _getSentenceBoundaries(sText):
        if 4 < (iEnd - iStart) < 2000:
            dDA.clear()
            try:
                # regex parser
                _, errs = _proofread(sText[iStart:iEnd], sRealText[iStart:iEnd], iStart, False, dDA, dPriority, sCountry, dOpt, bShowRuleId, bDebug, bContext)
                aErrors.update(errs)
                # token parser
                oSentence = TokenSentence(sText[iStart:iEnd], sRealText[iStart:iEnd], iStart)
                bChange, errs = oSentence.parse(dAllGraph["test_graph"], dPriority, sCountry, dOpt, bShowRuleId, bDebug, bContext)
                _, errs = _proofread(oSentence, sText[iStart:iEnd], sRealText[iStart:iEnd], iStart, False, dDA, dPriority, sCountry, dOpt, bShowRuleId, bDebug, bContext)
                aErrors.update(errs)
                if bChange:
                    oSentence.rewrite()
                    if bDebug:
                        print("~", oSentence.sSentence)
            except:
                raise
    return aErrors.values() # this is a view (iterable)


_zEndOfSentence = re.compile(r'([.?!:;…][ .?!… »”")]*|.$)')
_zBeginOfParagraph = re.compile(r"^\W*")
_zEndOfParagraph = re.compile(r"\W*$")

def _getSentenceBoundaries (sText):
    iStart = _zBeginOfParagraph.match(sText).end()
    for m in _zEndOfSentence.finditer(sText):
        yield (iStart, m.end())
        iStart = m.end()


def _proofread (s, sx, nOffset, bParagraph, dDA, dPriority, sCountry, dOptions, bShowRuleId, bDebug, bContext):
def _proofread (oSentence, s, sx, nOffset, bParagraph, dDA, dPriority, sCountry, dOptions, bShowRuleId, bDebug, bContext):
    dErrs = {}
    bChange = False
    for sOption, lRuleGroup in _getRules(bParagraph):
        if sOption == "@@@@":
            # graph rules
            for sGraphName, sLineId in lRuleGroup:
                if bDebug:
                    print(sGraphName, sLineId)
                bChange, errs = oSentence.parse(dAllGraph[sGraphName], dPriority, sCountry, dOptions, bShowRuleId, bDebug, bContext)
                dErrs.update(errs)
                if bChange:
                    oSentence.rewrite()
                    if bDebug:
                        print("~", oSentence.sSentence)
        if not sOption or dOptions.get(sOption, False):
        elif not sOption or dOptions.get(sOption, False):
            # regex rules
            for zRegex, bUppercase, sLineId, sRuleId, nPriority, lActions in lRuleGroup:
                if sRuleId not in _aIgnoredRules:
                    for m in zRegex.finditer(s):
                        bCondMemo = None
                        for sFuncCond, cActionType, sWhat, *eAct in lActions:
                            # action in lActions: [ condition, action type, replacement/suggestion/action[, iGroup[, message, URL]] ]
                            try:
︙
323
324
325
326
327
328
329

330
331
332



333
334
335
336
337
338
339

328
329
330
331
332
333
334
335



336
337
338
339
340
341
342
343
344
345








+
-
-
-
+
+
+








    if sFilter:
        try:
            zFilter = re.compile(sFilter)
        except:
            echo("# Error. List rules: wrong regex.")
            sFilter = None
    for sOption, lRuleGroup in chain(_getRules(True), _getRules(False)):
        if sOption != "@@@@":
        for _, _, sLineId, sRuleId, _, _ in lRuleGroup:
            if not sFilter or zFilter.search(sRuleId):
                yield (sOption, sLineId, sRuleId)
            for _, _, sLineId, sRuleId, _, _ in lRuleGroup:
                if not sFilter or zFilter.search(sRuleId):
                    yield (sOption, sLineId, sRuleId)


def displayRules (sFilter=None):
    echo("List of rules. Filter: << " + str(sFilter) + " >>")
    for sOption, sLineId, sRuleId in listRules(sFilter):
        echo("{:<10} {:<10} {}".format(sOption, sLineId, sRuleId))
︙

Modified gc_lang/fr/rules.grx from [bf0a5c739e] to [5101ac8a06].

︙
12441
12442
12443
12444
12445
12446
12447




12448
12449
12450
12451
12452
12453
12454

12441
12442
12443
12444
12445
12446
12447
12448
12449
12450
12451
12452
12453
12454
12455
12456
12457
12458








+
+
+
+








    -2>> =suggVerbMode(@, ":I", \1)
    # Après « quand » ou « lorsque », le verbe ne s’emploie pas au subjonctif mais à l’indicatif.

TEST: quand elle {{rencontrât}} son créateur
TEST: lorsqu’il y {{eût}} du grabuge, nous montâmes tous sur le pont.



@@@@GRAPH: test_graph@@@@



!!
!!
!!
!!
!!
!!
︙
16512
16513
16514
16515
16516
16517
16518
16519
16520
16521
16522
16523
16524
16525
16526
16527
16528

16516
16517
16518
16519
16520
16521
16522

16523

16524
16525
16526
16527
16528
16529
16530








-

-








TEST: Éliante, cousine de Célimène,
TEST: Arsinoé, amie de Célimène,
TEST: Acaste,
TEST: Clitandre, marquis
TEST: Basque, valet de Célimène,
TEST: Un garde de la maréchaussée de France,
TEST: Dubois, valet d’Alceste.

TEST: La scène se passe à Paris, dans la maison de Célimène.

TEST: ACTE I
TEST: SCÈNE PREMIÈRE. Philinte, Alceste.
TEST: PHILINTE. Qu’est-ce donc ? Qu’avez-vous ?
TEST: ALCESTE, assis. Laissez-moi, je vous prie.
TEST: PHILINTE. Mais encor, dites-moi, quelle bizarrerie…
TEST: ALCESTE. Laissez-moi là, vous dis-je, et courez vous cacher.
TEST: PHILINTE. Mais on entend les gens au moins sans se fâcher.
︙

Fossil 2.24 [8be0372c10] 2024-04-23 13:25:26
This page was generated in about 0.154s