models.py

            ),
            'service': self.service,
            'service_args': self.service_args,
            'service_version': self.service_version,
            'status': self.status.name,
            'title': self.title,
            'url': self.url
        }
        if backrefs:
            json_serializeable['user'] = \
                self.user.to_json_serializeable(backrefs=True)
        if relationships:
            json_serializeable['inputs'] = {
                x.hashid: x.to_json_serializeable(relationships=True)
                for x in self.inputs
            }
            json_serializeable['results'] = {
                x.hashid: x.to_json_serializeable(relationships=True)
                for x in self.results
            }
        return json_serializeable


class CorpusFile(FileMixin, HashidMixin, db.Model):
    __tablename__ = 'corpus_files'
    # Primary key
    id = db.Column(db.Integer, primary_key=True)
    # Foreign keys
    corpus_id = db.Column(db.Integer, db.ForeignKey('corpora.id'))
    # Fields
    author = db.Column(db.String(255))
    description = db.Column(db.String(255))
    publishing_year = db.Column(db.Integer)
    title = db.Column(db.String(255))
    address = db.Column(db.String(255))
    booktitle = db.Column(db.String(255))
    chapter = db.Column(db.String(255))
    editor = db.Column(db.String(255))
    institution = db.Column(db.String(255))
    journal = db.Column(db.String(255))
    pages = db.Column(db.String(255))
    publisher = db.Column(db.String(255))
    school = db.Column(db.String(255))
    # Backrefs: corpus: Corpus

    @property
    def download_url(self):
        return url_for(
            'corpora.download_corpus_file',
            corpus_id=self.corpus_id,
            corpus_file_id=self.id
        )

    @property
    def jsonpatch_path(self):
        return f'{self.corpus.jsonpatch_path}/files/{self.hashid}'

    @property
    def path(self):
        return os.path.join(self.corpus.path, 'files', str(self.id))

    @property
    def url(self):
        return url_for(
            'corpora.corpus_file',
            corpus_id=self.corpus_id,
            corpus_file_id=self.id
        )

    @property
    def user_hashid(self):
        return self.corpus.user.hashid

    @property
    def user_id(self):
        return self.corpus.user_id

    def delete(self):
        try:
            os.remove(self.path)
        except OSError as e:
            current_app.logger.error(e)
        db.session.delete(self)
        self.corpus.status = CorpusStatus.UNPREPARED

    def to_json_serializeable(self, backrefs=False, relationships=False):
        json_serializeable = {
            'id': self.hashid,
            'url': self.url,
            'address': self.address,
            'author': self.author,
            'booktitle': self.booktitle,
            'chapter': self.chapter,
            'editor': self.editor,
            'institution': self.institution,
            'journal': self.journal,
            'pages': self.pages,
            'publisher': self.publisher,
            'publishing_year': self.publishing_year,
            'school': self.school,
            'title': self.title,
            **self.file_mixin_to_json_serializeable(
                backrefs=backrefs,
                relationships=relationships
            )
        }
        if backrefs:
            json_serializeable['corpus'] = \
                self.corpus.to_json_serializeable(backrefs=True)
        return json_serializeable

class Corpus(HashidMixin, db.Model):
    '''
    Class to define a corpus.
    '''
    __tablename__ = 'corpora'
    # Primary key
    id = db.Column(db.Integer, primary_key=True)
    # Foreign keys
    user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
    # Fields
    creation_date = db.Column(db.DateTime(), default=datetime.utcnow)
    description = db.Column(db.String(255))
    status = db.Column(
        IntEnumColumn(CorpusStatus),
        default=CorpusStatus.UNPREPARED
    )
    title = db.Column(db.String(32))
    num_analysis_sessions = db.Column(db.Integer, default=0)
    num_tokens = db.Column(db.Integer, default=0)
    is_public = db.Column(db.Boolean, default=False)
    # Backrefs: user: User
    # Relationships
    files = db.relationship(
        'CorpusFile',
        backref='corpus',
        lazy='dynamic',
        cascade='all, delete-orphan'
    )
    # "static" attributes
    max_num_tokens = 2_147_483_647

    def __repr__(self):
        return f'<Corpus {self.title}>'

    @property
    def analysis_url(self):
        return url_for('corpora.analyse_corpus', corpus_id=self.id)

    @property
    def jsonpatch_path(self):
        return f'{self.user.jsonpatch_path}/corpora/{self.hashid}'

    @property
    def path(self):
        return os.path.join(self.user.path, 'corpora', str(self.id))

    @property
    def url(self):
        return url_for('corpora.corpus', corpus_id=self.id)

    @property
    def user_hashid(self):
        return self.user.hashid

    @staticmethod
    def create(**kwargs):
        corpus = Corpus(**kwargs)
        db.session.add(corpus)
        db.session.flush(objects=[corpus])
        db.session.refresh(corpus)
        try:
            os.mkdir(corpus.path)
            os.mkdir(os.path.join(corpus.path, 'files'))
            os.mkdir(os.path.join(corpus.path, 'cwb'))
            os.mkdir(os.path.join(corpus.path, 'cwb', 'data'))
            os.mkdir(os.path.join(corpus.path, 'cwb', 'registry'))
        except OSError as e:
            current_app.logger.error(e)
            db.session.rollback()
            raise e
        return corpus

    def build(self):
        corpus_element = ET.fromstring('<corpus>\n</corpus>')
        for corpus_file in self.files:
            normalized_vrt_path = os.path.join(self.path, 'cwb', f'{corpus_file.id}.norm.vrt')
            try:
                normalize_vrt_file(corpus_file.path, normalized_vrt_path)
            except:
                self.status = CorpusStatus.FAILED
                return
            element_tree = ET.parse(normalized_vrt_path)
            text_element = element_tree.getroot()
            text_element.set('author', corpus_file.author)
            text_element.set('title', corpus_file.title)
            text_element.set(
                'publishing_year',
                f'{corpus_file.publishing_year}'
            )
            text_element.set('address', corpus_file.address or 'NULL')
            text_element.set('booktitle', corpus_file.booktitle or 'NULL')
            text_element.set('chapter', corpus_file.chapter or 'NULL')
            text_element.set('editor', corpus_file.editor or 'NULL')
            text_element.set('institution', corpus_file.institution or 'NULL')
            text_element.set('journal', corpus_file.journal or 'NULL')
            text_element.set('pages', f'{corpus_file.pages}' or 'NULL')
            text_element.set('publisher', corpus_file.publisher or 'NULL')
            text_element.set('school', corpus_file.school or 'NULL')
            text_element.tail = '\n'
            # corpus_element.insert(1, text_element)
            corpus_element.append(text_element)
        ET.ElementTree(corpus_element).write(
            os.path.join(self.path, 'cwb', 'corpus.vrt'),
            encoding='utf-8'
        )
        self.status = CorpusStatus.SUBMITTED

    def delete(self):
        shutil.rmtree(self.path, ignore_errors=True)
        db.session.delete(self)

    def to_json_serializeable(self, backrefs=False, relationships=False):
        json_serializeable = {
            'id': self.hashid,
            'creation_date': f'{self.creation_date.isoformat()}Z',
            'description': self.description,
            'max_num_tokens': self.max_num_tokens,
            'num_analysis_sessions': self.num_analysis_sessions,
            'num_tokens': self.num_tokens,
            'status': self.status.name,
            'title': self.title,
            'is_public': self.is_public
        }
        if backrefs:
            json_serializeable['user'] = self.user.to_json_serializeable(backrefs=True)
        if relationships:
            json_serializeable['files'] = {
                x.hashid: x.to_json_serializeable(relationships=True)
                for x in self.files
            }
        return json_serializeable
# endregion models


##############################################################################
# event_handlers                                                             #
##############################################################################
# region event_handlers
@db.event.listens_for(Corpus, 'after_delete')
@db.event.listens_for(CorpusFile, 'after_delete')
@db.event.listens_for(Job, 'after_delete')
@db.event.listens_for(JobInput, 'after_delete')
@db.event.listens_for(JobResult, 'after_delete')
def ressource_after_delete(mapper, connection, ressource):
    jsonpatch = [{'op': 'remove', 'path': ressource.jsonpatch_path}]
    room = f'users.{ressource.user_hashid}'
    socketio.emit('users.patch', jsonpatch, room=room)
    room = f'/users/{ressource.user_hashid}'
    socketio.emit('PATCH', jsonpatch, room=room)


@db.event.listens_for(Corpus, 'after_insert')
@db.event.listens_for(CorpusFile, 'after_insert')
@db.event.listens_for(Job, 'after_insert')
@db.event.listens_for(JobInput, 'after_insert')
@db.event.listens_for(JobResult, 'after_insert')
def ressource_after_insert_handler(mapper, connection, ressource):
    value = ressource.to_json_serializeable()
    for attr in mapper.relationships:
        value[attr.key] = {}
    jsonpatch = [
        {'op': 'add', 'path': ressource.jsonpatch_path, 'value': value}
    ]
    room = f'/users/{ressource.user_hashid}'
    socketio.emit('PATCH', jsonpatch, room=room)


@db.event.listens_for(Corpus, 'after_update')
@db.event.listens_for(CorpusFile, 'after_update')
@db.event.listens_for(Job, 'after_update')
@db.event.listens_for(JobInput, 'after_update')
@db.event.listens_for(JobResult, 'after_update')
def ressource_after_update_handler(mapper, connection, ressource):
    jsonpatch = []
    for attr in db.inspect(ressource).attrs:
        if attr.key in mapper.relationships:
            continue
        if not attr.load_history().has_changes():
            continue
        if isinstance(attr.value, datetime):
            value = f'{attr.value.isoformat()}Z'
        elif isinstance(attr.value, Enum):
            value = attr.value.name
        else:
            value = attr.value
        jsonpatch.append(
            {
                'op': 'replace',
                'path': f'{ressource.jsonpatch_path}/{attr.key}',
                'value': value
            }
        )
    if jsonpatch:
        room = f'/users/{ressource.user_hashid}'
        socketio.emit('PATCH', jsonpatch, room=room)


@db.event.listens_for(Job, 'after_update')
def job_after_update_handler(mapper, connection, job):
    for attr in db.inspect(job).attrs:
        if attr.key != 'status':
            continue
        if not attr.load_history().has_changes():
            return
        if job.user.setting_job_status_mail_notification_level == UserSettingJobStatusMailNotificationLevel.NONE:
            return
        if job.user.setting_job_status_mail_notification_level == UserSettingJobStatusMailNotificationLevel.END:
            if job.status not in [JobStatus.COMPLETED, JobStatus.FAILED]:
                return
        msg = create_message(
            job.user.email,
            f'Status update for your Job "{job.title}"',
            'tasks/email/notification',
            job=job
        )
        mail.send(msg)
# endregion event_handlers


##############################################################################
# misc                                                                       #
##############################################################################
# region misc
@login.user_loader
def load_user(user_id):
    return User.query.get(int(user_id))
# endregion misc