get_examples.py 6.4 KB
#-*- coding:utf-8 -*-

import codecs
import datetime
import os

from django.core.management.base import BaseCommand

from dictionary.models import Lemma, NKJP_Example, get_ready_statuses, get_checked_statuses
from semantics.models import LexicalUnitExamples
from settings import PROJECT_PATH

BASE_PATH = os.path.join(PROJECT_PATH, 'data')

LABELS = (u'hasło', 
          u'status hasła',
          u'identyfikator schematu',
          u'schemat', 
          u'opinia o schemacie',
          u'przykład',
          u'opinia o przykładzie',
          u'zródło przykładu',
          u'wybór typów fraz')

class Command(BaseCommand):
    help = 'Get pinned examples from Slowal.'

    def handle(self, **options):
        get_examples()
        
def get_examples():
    write_example_sentences('semantyczne-S_sprawdzone-20170811.txt', ['(S) sprawdzone'], True)
    write_example_sentences('wszystkie-S_sprawdzone-20170811.txt', ['(S) sprawdzone'], False)

    checked_names = [checked.status for checked in get_checked_statuses()]
    write_example_sentences('wszystkie-sprawdzone-20170811.txt', checked_names, False)
    # ready_statuses = get_ready_statuses()
    # write_detailed_examples(ready_statuses)
    # write_examples(ready_statuses)
    
def write_detailed_examples(statuses):
    try:
        lemmas = Lemma.objects.filter(old=False)
        lemmas = lemmas.filter(status__in=statuses)
        now = datetime.datetime.now().strftime('%Y%m%d')
        examples_file = codecs.open(os.path.join(BASE_PATH, 'detailed_examples_%s.csv' % now), 'wt', 'utf-8') 
        examples_file.write(u'%s\n' % u'\t'.join(LABELS)) 
        for lemma in lemmas.order_by('entry_obj__name'):
            print lemma
            lemma_entry = lemma.entry_obj.name
            lemma_status = lemma.status.status
            for frame in lemma.frames.order_by('text_rep').all():
                if not lemma.phraseology_ready() and frame.phraseologic:
                    continue
                frame_opinion = lemma.frame_opinions.filter(frame=frame).all()[0].value
                for example in lemma.nkjp_examples.filter(frame=frame):
                    sentence = example.sentence.replace('\n', ' ').replace('\r', '').replace('\t', ' ') 
                    arguments_selection = u'%s' % u' + '.join([u'%s' % selection.__unicode__() for selection in example.arguments.all()])
                    examples_file.write(u'%s\t%s\t%d\t%s\t%s\t%s\t%s\t%s\t%s\n' % (lemma_entry,
                                                                                   lemma_status,
                                                                                   frame.id,
                                                                                   frame.get_position_spaced_text_rep(),
                                                                                   frame_opinion,
                                                                                   sentence,
                                                                                   example.opinion.opinion,
                                                                                   example.source.source,
                                                                                   arguments_selection))
    finally:
        examples_file.close()

def write_examples(statuses):
    try:       
        examples_file = codecs.open(os.path.join(BASE_PATH, 
                                                 'examples_gotowe_plus.txt'), 'wt', 'utf-8')  
        for lemma in Lemma.objects.filter(old=False).filter(status__in=statuses).order_by('entry_obj__name').all():
            print lemma
            examples_file.write(lemma.entry_obj.name+'\n')
            for frame in lemma.frames.order_by('text_rep').all():
                if lemma.frame_opinions.get(frame=frame).value.value != u'zła':
                    examples_file.write('\t%s\n' % frame.text_rep)
                    for example in lemma.nkjp_examples.filter(frame=frame):
                        examples_file.write('\t\t--> %s\n' % example.sentence)
            examples_file.write('\n\n')
    finally:
        examples_file.close()


def write_example_sentences(filename, statuses, semantic):
    try:
        examples_file = codecs.open(os.path.join(BASE_PATH, filename), 'wt', 'utf-8')
        for lemma in Lemma.objects.filter(old=False, entry_obj__pos__tag='verb').filter(status__status__in=statuses).order_by('entry_obj__name'):
            print lemma
            wrong_examples = lemma.nkjp_examples.filter(opinion__opinion=u'zły')
            not_wanted_semantic_examples = get_not_needed_semantic_examples(lemma)
            wanted_semantic_examples = get_wanted_semantic_examples(lemma)

            for example in lemma.nkjp_examples.filter(source__sym_name__in=['NKJP300M', 'NKJP1800M']):
                if (lemma.frame_opinions.filter(frame=example.frame, value__value__in=[u'archaiczny',
                                                                                       u'zły']).exists()):
                    continue

                if semantic:
                    if (wanted_semantic_examples.filter(pk=example.pk).exists() and
                        not wrong_examples.filter(pk=example.pk).exists()):
                        examples_file.write(u'%s\n' % example.sentence)
                else:
                    if (not not_wanted_semantic_examples.filter(pk=example.pk).exists() and
                        not wrong_examples.filter(pk=example.pk).exists()):
                        examples_file.write(u'%s\n' % example.sentence)

    finally:
        examples_file.close()


def get_not_needed_semantic_examples(lemma):
    not_needed_ids = []
    not_needed_frames = lemma.entry_obj.actual_frames().filter(opinion__value__in=[u'archaiczna', u'zła'])
    for frame in not_needed_frames:
        for lu in frame.lexical_units.all():
            for luex in LexicalUnitExamples.objects.filter(lexical_unit=lu):
                not_needed_ids.append(luex.example.id)
    return NKJP_Example.objects.filter(id__in=not_needed_ids)


def get_wanted_semantic_examples(lemma):
    needed_ids = []
    needed_frames = lemma.entry_obj.actual_frames().exclude(opinion__value__in=[u'archaiczna', u'zła'])
    for frame in needed_frames:
        for lu in frame.lexical_units.all():
            for luex in LexicalUnitExamples.objects.filter(lexical_unit=lu):
                needed_ids.append(luex.example.id)
    return NKJP_Example.objects.filter(id__in=needed_ids)