mark_potential_errors.py 19 KB
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 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 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484
#-*- coding:utf-8 -*-

import re
import sys
import time

import jsonpickle
import natsort

from operator import itemgetter

from django.core.management.base import BaseCommand
from multiservice.facade import Multiservice
from multiservice.facade.ttypes import *
from multiservice.types.ttypes import *
from thrift.transport import TSocket

from webapp.models import Meaning, MeaningStatus, Segment


PORT = 20000
HOST = 'test.multiservice.nlp.ipipan.waw.pl'

NOUN_TAGS = ['subst', 'ger', 'depr']
VERB_TAGS = ['pred', 'fin', 'praet', 'bedzie', 'inf', 'imps',
             'impt', 'winien', 'aglt']


class Command(BaseCommand):
    help = 'Mark potential errors in the database.'

    def handle(self, *args, **options):
        mark_potential_errors()


def mark_potential_errors():
    for meaning in Meaning.objects.all():
        expressions = meaning.expressions.filter(main_expression=None)
        for expression in expressions:
            if not missing_head(meaning, expression):
                head_is_not_nominal(meaning, expression)
                remove_kolega(meaning, expression)
                multiservice_checks(meaning, expression)
            first_word_is_prep(meaning, expression)
            first_word_is_verb(meaning, expression)
            contains_interp(meaning, expression)
            contains_owczesny(meaning, expression)
            contains_nazwisko(meaning, expression)
            contains_lub(meaning, expression)
            contains_itp(meaning, expression)
            contains_suspension_point(meaning, expression)
            contains_skrot(meaning, expression)
            contains_np(meaning, expression)
            starts_with_interp(meaning, expression)
            ends_with_interp(meaning, expression)


def missing_head(meaning, expression):
    is_missing = False
    try:
        expression.segments.get(is_head=True)
    except Segment.DoesNotExist:
        print (u'Missing head:', meaning, '---->', expression)
        if not meaning.status:
            meaning.comment = u'Brak głowy: %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='modify')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nBrak głowy: %s' % (meaning.comment, expression.text)
            meaning.save()
        is_missing = True
    return is_missing


def head_is_not_nominal(meaning, expression):
    head = expression.segments.get(is_head=True)
    if head.ctag not in NOUN_TAGS:
        print (u'Head is not nominal:', meaning, '---->', expression)
        if not meaning.status:
            meaning.comment = u'Głowa nie jest nominalna: %s --> %s' % (expression.text, head.orth)
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nGłowa nie jest nominalna: %s --> %s' % (meaning.comment,
                                                                            expression.text,
                                                                            head.orth)
            meaning.save()


def remove_kolega(meaning, expression):
    if expression.segments.get(is_head=True).base == 'kolega':
        print (u'Kolega:', meaning)
        if not meaning.status:
            meaning.comment = u'Głowa to "kolega": %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nGłowa to "kolega": %s' % (meaning.comment, expression.text)
            meaning.save()


def first_word_is_prep(meaning, expression):
    if expression.segments.order_by('position_in_expr')[0].ctag == 'prep':
        print (u'First word is prep:', meaning, '---->', expression)
        if not meaning.status:
            meaning.comment = u'Pierwsze słowo to przyimek: %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nPierwsze słowo to przyimek: %s' % (meaning.comment, expression.text)
            meaning.save()


def first_word_is_verb(meaning, expression):
    if expression.segments.order_by('position_in_expr')[0].ctag in VERB_TAGS:
        print (u'First word is verb:', meaning, '---->', expression)
        if not meaning.status:
            meaning.comment = u'Pierwsze słowo to czasownik: %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nPierwsze słowo to czasownik: %s' % (meaning.comment, expression.text)
            meaning.save()


def contains_interp(meaning, expression):
    if expression.segments.filter(ctag='interp').exclude(orth=',').exists():
        print (u'Contains interp:', meaning, '---->', expression)
        if not meaning.status:
            meaning.comment = u'Zawiera znak interpunkcyjny, nie przecinek: %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nZawiera znak interpunkcyjny, nie przecinek: %s' % (meaning.comment, expression.text)
            meaning.save()

# zamienic na "zawiera date"
# def contains_digit(meaning, expression):
#     _digits = re.compile('\d')
#     if _digits.search(expression.orth_text):
#         print (u'Contains digits:', meaning, '---->', expression)
#         if not meaning.status:
#             meaning.comment = u'Zawiera cyfrę: %s' % expression.text
#             meaning.status = MeaningStatus.objects.get(key='check')
#             meaning.save()
#         elif meaning.status.key != 'ok':
#             meaning.comment = u'%s\n\nZawiera cyfrę: %s' % (meaning.comment, expression.text)
#             meaning.save()


def contains_owczesny(meaning, expression):
    if expression.segments.filter(base='ówczesny').exists():
        print (u'Contains ówczesny:', meaning)
        if not meaning.status:
            meaning.comment = u'Zawiera słowo "ówczesny": %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nZawiera słowo "ówczesny": %s' % (meaning.comment, expression.text)
            meaning.save()


def contains_nazwisko(meaning, expression):
    if expression.segments.filter(base='nazwisko').exists():
        print (u'Contains nazwisko:', meaning)
        if not meaning.status:
            meaning.comment = u'Zawiera słowo "nazwisko": %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nZawiera słowo "nazwisko": %s' % (meaning.comment, expression.text)
            meaning.save()


def contains_lub(meaning, expression):
    if expression.segments.filter(orth='lub').exists():
        print (u'Contains lub:', meaning)
        if not meaning.status:
            meaning.comment = u'Zawiera słowo "lub": %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nZawiera słowo "lub": %s' % (meaning.comment, expression.text)
            meaning.save()


def contains_itp(meaning, expression):
    if expression.segments.filter(orth='itp').exists():
        print (u'Contains itp:', meaning)
        if not meaning.status:
            meaning.comment = u'Zawiera słowo "itp": %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nZawiera słowo "itp": %s' % (meaning.comment, expression.text)
            meaning.save()


def contains_suspension_point(meaning, expression):
    if '..' in expression.orth_text:
        print (u'Contains ...:', meaning)
        if not meaning.status:
            meaning.comment = u'Zawiera wielokropek: %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nZawiera wielokropek: %s' % (meaning.comment, expression.text)
            meaning.save()


def contains_skrot(meaning, expression):
    if expression.segments.filter(base='skrót').exists():
        print (u'Contains skrót:', meaning)
        if not meaning.status:
            meaning.comment = u'Zawiera słowo "skrót": %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nZawiera słowo "skrót": %s' % (meaning.comment, expression.text)
            meaning.save()


def contains_np(meaning, expression):
    if (expression.segments.filter(orth='np').exists() or
            expression.segments.filter(orth='Np').exists() or
            expression.segments.filter(orth='NP').exists()):
        print (u'Contains np:', meaning)
        if not meaning.status:
            meaning.comment = u'Zawiera słowo "np": %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nZawiera słowo "np": %s' % (meaning.comment, expression.text)
            meaning.save()


def starts_with_interp(meaning, expression):
    if expression.segments.order_by('position_in_expr')[0].ctag == 'interp':
        print (u'Interp at start:', meaning, '---->', expression)
        if not meaning.status:
            meaning.comment = u'Rozpoczyna się od znaku interpunkcyjnego: %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nRozpoczyna się od znaku interpunkcyjnego: %s' % (meaning.comment, expression.text)
            meaning.save()


def ends_with_interp(meaning, expression):
    if list(expression.segments.order_by('position_in_expr'))[-1].ctag == 'interp':
        print (u'Ends with interp:', meaning, '---->', expression)
        if not meaning.status:
            meaning.comment = u'Kończy się na znaku interpunkcyjnym: %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nKończy się na znaku interpunkcyjnym: %s' % (meaning.comment, expression.text)
            meaning.save()


def multiservice_checks(meaning, expression):
    process_chain = ['Concraft', 'Nerf', 'DependencyParser']
    expression_orth = expression.orth_text
    json_response = get_json_response(expression_orth, process_chain)

    forename_only(meaning, expression, json_response)
    head_marked_properly(meaning, expression, json_response)

    if expression.is_catchword:
        definition_structure(meaning, expression)


def get_json_response(expression, process_chain):
    transport, client = getThriftTransportAndClient(HOST, PORT)
    request = createRequest(expression, process_chain)
    jsonObj = None
    try:
        token = client.putObjectRequest(request)
        status = None
        while status not in [RequestStatus.DONE, RequestStatus.FAILED]:
            status = client.getRequestStatus(token)
            time.sleep(0.1)
        if status == RequestStatus.DONE:
            result = client.getResultObject(token)
            jsonStr = jsonpickle.encode(result, unpicklable=False)
            jsonObj = jsonpickle.decode(jsonStr)
        else:
            print (client.getException(token))
            sys.exit("Stopped loading data!")
    finally:
        transport.close()
        return jsonObj


def getThriftTransportAndClient(host, port):
    transport = TSocket.TSocket(host, port)
    try:
        transport = TTransport.TBufferedTransport(transport)
        protocol = TBinaryProtocol.TBinaryProtocol(transport)
        client = Multiservice.Client(protocol)
        transport.open()
        return (transport, client)
    except:
        transport.close()
        raise


def createRequest(text, serviceNames):
    ttext = TText(paragraphs=[TParagraph(text=chunk)
                              for chunk in re.split(r'\n\n+', text)])
    chain = [RequestPart(serviceName=name) for name in serviceNames]
    request = ObjectRequest(ttext, chain)
    return request


def forename_only(meaning, expression, json_response):
    if is_forename(expression, json_response):
        print (u'Is forename:', meaning)
        if not meaning.status:
            meaning.comment = u'Jest imieniem: %s' % expression.text
            meaning.status = MeaningStatus.objects.get(key='delete')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\nJest imieniem: %s' % (meaning.comment, expression.text)
            meaning.save()


def is_forename(expression, json_response):
    for para in json_response['paragraphs']:
        for sent in para['sentences']:
            for ne in sent['names']:
                if (ne['subtype'] == 'forename' and
                            expression.orth_text.lower() == ne['orth'].lower()):
                    return True
    return False


def head_marked_properly(meaning, expression, json_response):
    tokens = get_tokens(json_response)
    root = get_root(tokens, json_response)
    if root['tok']['orth'].lower() != expression.segments.get(is_head=True).orth.lower():
        print (u'Head is not same as in dependency tree:', expression,
               u'dependency parser --> %s' % root['tok']['orth'],
               u'database --> %s' % expression.segments.get(is_head=True).orth)
        if not meaning.status:
            meaning.comment = u'Brak zgodności głów w wyrażeniu %s: ' \
                              u'parser zależnościowy --> %s, ' \
                              u'baza --> %s' % (expression.text,
                                                root['tok']['orth'],
                                                expression.segments.get(is_head=True).orth)
            meaning.status = MeaningStatus.objects.get(key='check')
            meaning.save()
        elif meaning.status.key != 'ok':
            meaning.comment = u'%s\n\n' \
                              u'Brak zgodności głów w wyrażeniu %s: ' \
                              u'parser zależnościowy --> %s, ' \
                              u'baza --> %s' % (meaning.comment,
                                                expression.text,
                                                root['tok']['orth'],
                                                expression.segments.get(is_head=True).orth)
            meaning.save()


def get_root(tokens, json_response):
    root = None
    for para in json_response['paragraphs']:
        for sent in para['sentences']:
            for link in sent['dependencyParse']:
                if link['label'] == 'root':
                    root = tokens[link['endTokenId']]
                    break
    return root


def definition_structure(meaning, catchword):
    process_chain = ['Concraft', 'DependencyParser']
    for definition in meaning.expressions.filter(main_expression=None).exclude(pk=catchword.pk):
        sentence = u'%s to %s.' % (definition.orth_text.capitalize(), catchword.orth_text)
        json_response = get_json_response(sentence, process_chain)
        if not is_definition(catchword, definition, json_response):
            print (u'Not a definition:', catchword, '-->', definition)
            if not meaning.status:
                meaning.comment = u'Nie jest definicją: %s' % sentence
                meaning.status = MeaningStatus.objects.get(key='check')
                meaning.save()
            elif meaning.status.key != 'ok':
                meaning.comment = u'%s\n\nNie jest definicją: %s' % (meaning.comment, sentence)
                meaning.save()


def is_definition(catchword, definition, json_response):
    tokens = get_tokens(json_response)
    root = get_root_and_add_links(tokens, json_response)
    if root['tok']['chosenInterpretation']['base'] != 'to' or root['tok']['chosenInterpretation']['ctag'] != 'pred':
        return False

    pd_head = get_pd_head(root)
    subj_head = get_subj_head(root)
    if not pd_head or not subj_head:
        return False

    pd_expr = get_subtree_expression(pd_head)
    subj_expr = get_subtree_expression(subj_head)

    pd_is_catchword = False
    pd_is_definition = False
    if (pd_expr.lower() == definition.orth_text.lower()):
        pd_is_definition = True
    elif (pd_expr.lower() == catchword.orth_text.lower()):
        pd_is_catchword = True

    subj_is_catchword = False
    subj_is_definition = False
    if (subj_expr.lower() == catchword.orth_text.lower()):
        subj_is_catchword = True

    if (subj_expr.lower() == definition.orth_text.lower()):
        subj_is_definition = True

    return ((pd_is_definition and subj_is_catchword) or
            (pd_is_catchword and subj_is_definition))


def get_tokens(json_response):
    tokens = {}
    for para in json_response['paragraphs']:
        for sent in para['sentences']:
            for tok in sent['tokens']:
                tokens[tok['id']] = {'tok': tok,
                                     'links': [],
                                     'root': False,
                                     'id': tok['id']}
    return tokens


def get_root_and_add_links(tokens, json_response):
    root = None
    for para in json_response['paragraphs']:
        for sent in para['sentences']:
            for link in sent['dependencyParse']:
                if link['label'] == 'root':
                    root = tokens[link['endTokenId']]
                else:
                    from_tok = tokens[link['startTokenId']]
                    from_tok['links'].append({'to': tokens[link['endTokenId']],
                                              'label': link['label']})
    return root


def get_pd_head(root):
    for link in root['links']:
        if link['label'] == 'pd':
            return link['to']
    return None


def get_subj_head(root):
    for link in root['links']:
        if link['label'] == 'subj':
            return link['to']
    return None


def get_subtree_expression(token):
    subtree_expression = ''

    subtree_tokens = get_subtree_tokens(token)
    subtree_tokens = natsort.natsorted(subtree_tokens, key=itemgetter('id'))

    for i, tok in enumerate(subtree_tokens):
        if i == 0 or tok['tok']['noPrecedingSpace'] == 'true':
            subtree_expression += tok['tok']['orth']
        else:
            subtree_expression += ' %s' % tok['tok']['orth']

    return subtree_expression


def get_subtree_tokens(token):
    subtree_tokens = []
    subtree_tokens.append(token)
    for link in token['links']:
        subtree_tokens.extend(get_subtree_tokens(link['to']))
    return subtree_tokens