import_tei.py 11 KB
#! /usr/bin/python
# -*- coding: utf-8 -*-

from django.core.management.base import BaseCommand

import sys, os, shutil, codecs, copy, errno
from xml.sax import saxutils, handler, make_parser
from importer.WalentyXML import WalentyTeiHandler
from importer.WalentyPreprocessXML import WalentyPreprocessTeiHandler
from shellvalier.settings import BASE_DIR
from connections.models import POS, Status
from examples.models import ExampleOpinion, ExampleSource
from syntax.models import SchemaOpinion, Aspect, InherentSie, Negativity, Predicativity, SyntacticFunction, Control, PredicativeControl, Position
from syntax.models_phrase import (
    Case, PhraseAspect, AdverbialCategory, PhraseNegativity, PhraseInherentSie,
    Number, Gender, Degree,
    LemmaOperator, LemmaCooccur,
    ModificationType,
)
from semantics.models import FrameOpinion, ArgumentRole, SemanticRole, RoleAttribute, PredefinedSelectionalPreference, SelectionalPreferenceRelation
from common.models import ImportInProgress

class Command(BaseCommand):
    args = 'none'
    help = ''

    def handle(self, **options):
        import_tei()

def import_tei():
    #xml_file = os.path.join(BASE_DIR, 'data', 'walenty', 'walenty_20200926_smaller.xml')
    xml_file = os.path.join(BASE_DIR, 'data', 'walenty', 'walenty_20200926.xml')

    xml_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), xml_file)

    import_constants()

    parser = make_parser()
    parser.setFeature(handler.feature_external_ges, False)

    parser.setContentHandler(WalentyPreprocessTeiHandler())
    parser.parse(xml_path)

    entry_meanings = parser.getContentHandler().entry_meanings
    meanings = parser.getContentHandler().meanings
    frames = parser.getContentHandler().frames

    parser.setContentHandler(WalentyTeiHandler(entry_meanings, meanings, frames))
    parser.parse(xml_path)
    ImportInProgress.objects.all().delete()

def import_constants():
    import_poses()
    import_statuses()
    import_schema_opinions()
    import_frame_opinions()
    import_aspects()
    import_inherent_sies()
    import_negativities()
    import_predicativities()
    import_syntactic_functions()
    import_control_tags()
    import_semantic_roles()
    import_predefined_preferences()
    import_preference_relations()
    import_examples_sources()
    import_examples_opinions()
    import_phrase_attributes()
    import_lemma_operators()
    import_modification_types()
    pass

def import_poses():
    poses = [u'unk', u'adj', u'noun', u'adv', u'verb']
    for pos_tag in poses:
        pos = POS(tag=pos_tag)
        pos.save()

def import_statuses():
    statuses = [(10, u'do obróbki'), (20, u'w obróbce'), (25, u'do usunięcia'), (30, u'gotowe'), (35, u'zalążkowe'), (40, u'sprawdzone'), (50, u'(F) w obróbce'), (60, u'(F) gotowe'), (70, u'(F) sprawdzone'), (80, u'(S) w obróbce'), (90, u'(S) gotowe'), (100, u'(S) sprawdzone')]
    for pri, name in statuses:
        status = Status(key=name, priority=pri)
        status.save()

def import_schema_opinions():
    opinions = [(60, u'vul'), (50, u'col'), (40, u'dat'), (30, u'bad'), (20, u'unc'), (10, u'cer')]
    for pri, short in opinions:
        opinion = SchemaOpinion(key=short, priority=pri)
        opinion.save()

def import_frame_opinions():
    opinions = [(70, u'met'), (60, u'vul'), (50, u'col'), (40, u'dat'), (30, u'bad'), (20, u'unc'), (10, u'cer'), (80, u'dom'), (90, u'rar'), (100, u'unk')]
    for pri, short in opinions:
        opinion = FrameOpinion(key=short, priority=pri)
        opinion.save()

def import_aspects():
    aspects = [(10, u'imperf'), (20, u'perf'), (32, u'_'), (42, u'')]
    for pri, name in aspects:
        aspect = Aspect(name=name, priority=pri)
        aspect.save()

def import_inherent_sies():
    sies = [(10, u'false'), (20, u'true')]
    for pri, name in sies:
        sie = InherentSie(name=name, priority=pri)
        sie.save()

def import_negativities():
    negativities = [(20, u'aff'), (10, u'neg'), (31, u'_'), (41, u'')]
    for pri, name in negativities:
        neg = Negativity(name=name, priority=pri)
        neg.save()

def import_predicativities():
    predicativities = [(20, u'false'), (10, u'true')]
    for pri, name in predicativities:
        pred = Predicativity(name=name, priority=pri)
        pred.save()

def import_syntactic_functions():
    functions = [(0, u'subj'), (20, u'head'), (10, u'obj')]
    for pri, name in functions:
        sf = SyntacticFunction(name=name, priority=pri)
        sf.save()

def import_control_tags():
    controls = [(10, u'controller'), (20, u'controllee'), (30, u'controller2'), (40, u'controllee2')]
    for pri, name in controls:
        cont = Control(name=name, priority=pri)
        cont.save()
    controls = [(10, u'pred_controller'), (20, 'pred_controllee')]
    for pri, name in controls:
        cont = PredicativeControl(name=name, priority=pri)
        cont.save()

def import_semantic_roles():
    roles = [
        (10, u'Initiator', u'91,106,217', None),
        (20, u'Stimulus', u'62,173,226', None),
        (30, u'Condition', u'127,199,195', None),
        (40, u'Factor', u'82,150,87', None),
        (50, u'Experiencer', u'149,195,86', None),
        (60, u'Theme', u'90,179,69', None),
        (70, u'Recipient', u'203,77,141', None),
        (80, u'Result', u'231,155,159', None),
        (90, u'Instrument', u'199,221,60', None),
        (100, u'Manner', u'191,48,44', None),
        (110, u'Purpose', u'171,85,186', None),
        (120, u'Attribute', u'220,53,47', None),
        (130, u'Location', u'187,129,45', None),
        (140, u'Path', u'224,121,44', None),
        (150, u'Time', u'242,236,54', None),
        (160, u'Duration', u'233,192,6', None),
        (170, u'Measure', u'238,72,154', None),
        (180, u'Lemma', u'256,256,256', None)
    ]
    # priorities set so that, when role and attribute priorities are added,
    # Role_Source < Role_Foreground < Role_Background < Role_Goal
    # and Role can be inserted anywhere into that hierarchy
    attributes = [(1, u'Source', None, u'left'), (3, u'Foreground', None, u'top'), (5, u'Background', None, u'bottom'), (7, u'Goal', None, u'right')]
    for pri, role, color, gradient in roles:
        role = SemanticRole(role=role, color=color, priority=pri)
        role.save()
    for pri, role, color, gradient in attributes:
        role = RoleAttribute(attribute=role, gradient=gradient, priority=pri)
        role.save()

    for role in SemanticRole.objects.all():
        r = ArgumentRole(role=role, attribute=None)
        r.save()
        for attribute in RoleAttribute.objects.all():
            r = ArgumentRole(role=role, attribute=attribute)
            r.save()

def import_predefined_preferences():
    predefs = [u'ALL', u'LUDZIE', u'ISTOTY', u'PODMIOTY', u'KOMUNIKAT', u'KONCEPCJA', u'WYTWÓR', u'JADŁO', u'CZAS', u'OBIEKTY', u'CECHA', u'CZYNNOŚĆ', u'KIEDY', u'CZEMU', u'ILOŚĆ', u'POŁOŻENIE', u'DOBRA', u'MIEJSCE', u'SYTUACJA', u'OTOCZENIE']
    for name in predefs:
        predef = PredefinedSelectionalPreference(key=name)
        predef.save()

def import_preference_relations():
    relations = [(14, u'meronimia'), (15, u'holonimia'), (20, u'meronimia (typu część)'), (21, u'meronimia (typu porcja)'), (22, u'meronimia (typu miejsce)'), (23, u'meronimia (typu element)'), (24, u'meronimia (typu materiał)'), (25, u'holonimia (typu część)'), (26, u'holonimia (typu porcja)'), (27, u'holonimia (typu miejsce)'), (28, u'holonimia (typu element)'), (29, u'holonimia (typu materiał)'), (51, u'nosiciel stanu/cechy'), (52, u'stan/cecha'), (61, u'synonimia międzyparadygmatyczna'), (64, u'meronimia (typu element taksonomiczny)'), (65, u'holonimia (typu element taksonomiczny)'), (108, u'fuzzynimia synsetów'), (-1, u'RELAT')]
    for id, name in relations:
        relat = SelectionalPreferenceRelation(plwn_id=id, key=name)
        relat.save()


def import_examples_sources():
    sources = [(0, u'NKJP0.5M'), (1, u'NKJP1.2M'), (2, u'NKJP30M'), (3, u'NKJP250M'), (4, u'NKJP300M'), (5, u'NKJP500M'), (6, u'NKJP1800M'), (7, u'linguistic_literature'), (8, u'other_literature'), (9, u'own')]
    for pri, name in sources:
        es = ExampleSource(key=name, priority=pri)
        es.save()

def import_examples_opinions():
    opinions = [(0, 'zły'), (1, 'wątpliwy'), (2, 'dobry')]
    for pri, name in opinions:
        eo = ExampleOpinion(key=name, priority=pri)
        eo.save()

def import_phrase_attributes():
    import_cases()
    import_phrase_aspects()
    import_phrase_negativities()
    import_phrase_inherent_sies()
    import_adverbial_categories()
    import_numbers()
    import_genders()
    import_degrees()
    # TODO this is quite terrible... create a dummy position for storing phrases inside a lex
    # the store() method for phrases requires a position
    dummy_position = Position()
    dummy_position.save()
    assert (dummy_position.id == 1)

def import_cases():
    cases = [(0, u'str'), (1, u'nom'), (2, u'gen'), (3, u'dat'), (4, u'acc'), (5, u'inst'), (6, u'loc'), (10, u'pred'), (11, u'part'), (12, u'postp'), (13, u'agr')]
    for pri, name in cases:
        case = Case(name=name, priority=pri)
        case.save()

def import_phrase_aspects():
    aspects = [(10, u'imperf'), (20, u'perf'), (30, u'_')]
    for pri, name in aspects:
        aspect = PhraseAspect(name=name, priority=pri)
        aspect.save()

def import_phrase_negativities():
    negativities = [(10, u'aff'), (20, u'neg'), (30, u'_')]
    for pri, name in negativities:
        negativity = PhraseNegativity(name=name, priority=pri)
        negativity.save()

def import_phrase_inherent_sies():
    sies = [(10, u'się'), (20, u'')]
    for pri, name in sies:
        sie = PhraseInherentSie(name=name, priority=pri)
        sie.save()

def import_adverbial_categories():
    advcats = [(1, u'locat'), (2, u'abl'), (3, u'adl'), (4, u'perl'), (5, u'temp'), (6, u'dur'), (7, 'mod'), (8, 'caus'), (9, 'dest'), (10, 'instr'), (11, 'pron'), (12, 'misc')]
    for pri, name in advcats:
        advcat = AdverbialCategory(name=name, priority=pri)
        advcat.save()

def import_numbers():
    numbers = [(1, u'sg'), (2, u'pl'), (10, u'agr'), (20, u'_')]
    for pri, name in numbers:
        number = Number(name=name, priority=pri)
        number.save()

def import_genders():
    genders = [(1, u'm1'), (2, u'm2'), (3, u'm3'), (4, u'f'), (5, u'n'), (10, u'agr')]
    for pri, name in genders:
        gender = Gender(name=name, priority=pri)
        gender.save()

def import_degrees():
    degrees = [(1, u'pos'), (2, u'com'), (3, u'sup'), (20, u'_')]
    for pri, name in degrees:
        degree = Degree(name=name, priority=pri)
        degree.save()

def import_lemma_operators():
    operators = [(1, u'xor'), (2, u'or')]
    for pri, name in operators:
        operator = LemmaOperator(name=name, priority=pri)
        operator.save()
    cooccurs = [(1, u'concat'), (2, u'coord')]
    for pri, name in cooccurs:
        cooccur = LemmaCooccur(name=name, priority=pri)
        cooccur.save()

def import_modification_types():
    modtypes = [(1, u'ratr'), (2, u'ratr1'), (3, u'atr'), (4, u'atr1'), (5, u'natr')]
    for pri, name in modtypes:
        modtype = ModificationType(name=name, priority=pri)
        modtype.save()