def save_sub_events(self, event_data, sub_event_time_ranges, super_event):
super_event._changed = False
def delete_sub_event(obj):
logger.debug('{} deleted'.format(obj))
obj.deleted = True
obj.save()
sub_event_syncher = ModelSyncher(
super_event.sub_events.filter(deleted=False), lambda o: o.id, delete_func=delete_sub_event)
sub_event_data = deepcopy(event_data)
sub_event_data['super_event'] = super_event
for sub_event_time_range in sub_event_time_ranges:
sub_event_data['start_time'] = sub_event_time_range.start
sub_event_data['end_time'] = sub_event_time_range.end
sub_event_data['origin_id'] = (
event_data['origin_id'] + self.create_sub_event_origin_id_suffix(sub_event_time_range))
sub_event = self.save_event(sub_event_data)
if sub_event._changed:
super_event._changed = True
sub_event_syncher.mark(sub_event)
old_sub_event_count = super_event.sub_events.count()
sub_event_syncher.finish(force=True)
if super_event.sub_events.count() != old_sub_event_count:
super_event._changed = True
if super_event._changed:
super_event.save()
def set_images(self, obj, images_data):
image_syncher = ModelSyncher(obj.images.all(),
lambda image: image.url,
delete_func=partial(
self._remove_image, obj))
for image_data in images_data:
image_url = image_data.get('url', '').strip()
if not image_url:
print('Invalid image url "{}" obj {}'.format(
image_data.get('url'), obj))
continue
new_image = False
image = image_syncher.get(image_url)
if not image:
new_image = True
image = self._get_image(image_url)
image = self._update_image(image, image_data)
if new_image:
obj.images.add(image)
obj._changed = True
obj._changed_fields.append('images')
elif image._changed:
obj._changed = True
obj._changed_fields.append('images')
image_syncher.mark(image)
image_syncher.finish(force=True)
def save_sub_events(self, event_data, sub_event_time_ranges, super_event):
super_event._changed = False
def delete_sub_event(obj):
logger.debug('{} deleted'.format(obj))
obj.deleted = True
obj.save()
sub_event_syncher = ModelSyncher(
super_event.sub_events.filter(deleted=False), lambda o: o.id, delete_func=delete_sub_event)
sub_event_data = deepcopy(event_data)
sub_event_data['super_event'] = super_event
for sub_event_time_range in sub_event_time_ranges:
sub_event_data['start_time'] = sub_event_time_range.start
sub_event_data['end_time'] = sub_event_time_range.end
sub_event_data['origin_id'] = (
event_data['origin_id'] + self.create_sub_event_origin_id_suffix(sub_event_time_range))
sub_event = self.save_event(sub_event_data)
if sub_event._changed:
super_event._changed = True
sub_event_syncher.mark(sub_event)
old_sub_event_count = super_event.sub_events.count()
sub_event_syncher.finish(force=True)
if super_event.sub_events.count() != old_sub_event_count:
super_event._changed = True
if super_event._changed:
super_event.save()
def set_images(self, obj, images_data):
image_syncher = ModelSyncher(obj.images.all(),
lambda image: image.url,
delete_func=partial(self._remove_image, obj))
for image_data in images_data:
image_url = image_data.get('url', '').strip()
if not image_url:
print('Invalid image url "{}" obj {}'.format(image_data.get('url'), obj))
continue
new_image = False
image = image_syncher.get(image_url)
if not image:
new_image = True
image = self._get_image(image_url)
image = self._update_image(image, image_data)
if new_image:
obj.images.add(image)
obj._changed = True
elif image._changed:
obj._changed = True
image_syncher.mark(image)
image_syncher.finish(force=True)
class HarrastushakuImporter(Importer):
name = 'harrastushaku'
supported_languages = ['fi']
def setup(self):
logger.debug('Running Harrastushaku importer setup...')
self.data_source, _ = DataSource.objects.get_or_create(id=self.name, defaults={'name': 'Harrastushaku'})
self.tprek_data_source = DataSource.objects.get(id='tprek')
self.ahjo_data_source, _ = DataSource.objects.get_or_create(id='ahjo', defaults={'name': 'Ahjo'})
self.organization, _ = Organization.objects.get_or_create(origin_id='u48040030',
data_source=self.ahjo_data_source)
self.tprek_ids = {place.origin_id for place in Place.objects.filter(data_source=self.tprek_data_source)}
self.keywords = {keyword.id: keyword for keyword in Keyword.objects.all()}
self.keyword_matcher = KeywordMatcher()
def import_places(self):
"""Import Harrastushaku locations as Places
- If we can find a close-enough match for the location object coming from Harrastushaku in Toimipisterekisteri,
we do not import that location object, as this this will cause duplicate location issue due to
Harrastushaku data being of low quality.
- If, however, we cannot find a match, location object will be imported with data source "harrastushaku".
"""
logger.info('Importing places...')
locations = self.fetch_locations()
logger.debug('Handling {} locations...'.format(len(locations)))
self.location_id_to_place_id = self.map_harrastushaku_location_ids_to_tprek_ids(locations)
for location in locations:
try:
self.handle_location(location)
except Exception as e: # noqa
message = e if isinstance(e, HarrastushakuException) else traceback.format_exc()
logger.error('Error handling location {}: {}'.format(location.get('id'), message))
def map_harrastushaku_location_ids_to_tprek_ids(self, harrastushaku_locations):
'''
Example mapped dictionary result:
{
'95': 'harrastushaku:95',
'953': 'harrastushaku:953',
'968': 'tprek:20479',
'97': 'tprek:8062',
'972': 'tprek:9079',
'987': 'harrastushaku:987',
'99': 'tprek:8064',
}
'''
result = dict()
for harrastushaku_location in harrastushaku_locations:
harrastushaku_location_id = harrastushaku_location['id']
strict_filters = {
'id__startswith': self.tprek_data_source,
'name': harrastushaku_location['name'],
'address_locality': harrastushaku_location['city'],
'postal_code': harrastushaku_location['zip'],
'street_address': harrastushaku_location['address'],
}
flexible_filters = {
'id__startswith': self.tprek_data_source,
'address_locality': harrastushaku_location['city'],
'postal_code': harrastushaku_location['zip'],
'street_address': harrastushaku_location['address'],
}
tprek_place = (Place.objects.filter(**strict_filters).first() or
Place.objects.filter(**flexible_filters).first())
if tprek_place:
result[harrastushaku_location_id] = tprek_place.id
else:
result[harrastushaku_location_id] = '{}:{}'.format(self.data_source.id, harrastushaku_location_id)
return result
def import_courses(self):
"""Import Harrastushaku activities as Courses
Activities having "active" anything else than "1" or "K" will be
ignored.
When importing and an existing course isn't present in imported data:
- If the course's end time is in the past, the course will be left as
it is.
- If the course's end time is not in the past, the course will be soft
deleted alongside its sub events.
If an activity has something in field "timetables", it will be imported
as a recurring event, otherwise as a one-time event.
A recurring course will have a super event which includes the course's
whole time period, and sub events which will represent individual course
occurrences. Other than start and end times, a super event and its sub
events will all contain the same data.
A recurring course's sub event start and end datetimes will be build using
the activity's "timetables". The time tables contain info out weekday,
times, and repetition which means number of days there is between
occurrences (basically a multiple of 7).
A recurring course's sub events will be given an ID that has the
activity's ID and start and end times of the sub event in a compressed
form. This also means that between imports only sub events that are
happening exactly at the same time are considered to be the same instance,
so if a sub event's begin or end time changes at all, a new sub event will
be created instead of updating an old one (because there is no unambiguous
way to determine which old sub event the new one corresponds to).
A course's keywords will come from both of the following:
- The activity's main category. There are hardcoded keywords for every
main category.
- The activity's sub category's "searchwords". Those are manually
entered words, which are mapped to keywords using KeywordMatcher
(from events.keywords).
A course's audience will come from both of the following:
- The activity's "audience_max_age" and "audience_min_age" using
hardcoded keywords for certain age ranges.
- The course's keywords, adding the ones that are present in
KEYWORDS_TO_ADD_TO_AUDIENCE (from events.importer.yso).
"""
logger.info('Importing courses...')
locations = self.fetch_locations()
if not locations:
logger.warning('No location data fetched, aborting course import.')
return
self.location_id_to_place_id = self.map_harrastushaku_location_ids_to_tprek_ids(locations)
activities = self.fetch_courses()
if not activities:
logger.info('No activity data fetched.')
return
def event_delete(event):
if event.end_time < now():
return
event.soft_delete()
for sub_event in event.sub_events.all():
sub_event.soft_delete()
self.event_syncher = ModelSyncher(
Event.objects.filter(data_source=self.data_source, super_event=None),
lambda event: event.id,
event_delete,
)
num_of_activities = len(activities)
logger.debug('Handling {} activities...'.format(num_of_activities))
for i, activity in enumerate(activities, 1):
try:
self.handle_activity(activity)
except Exception as e: # noqa
message = e if isinstance(e, HarrastushakuException) else traceback.format_exc()
logger.error('Error handling activity {}: {}'.format(activity.get('id'), message))
if not i % 10:
logger.debug('{} / {} activities handled.'.format(i, num_of_activities))
self.event_syncher.finish(force=True)
logger.info('Course import finished.')
def fetch_locations(self):
logger.debug('Fetching locations...')
try:
url = '{}location/'.format(HARRASTUSHAKU_API_BASE_URL)
response = requests.get(url)
response.raise_for_status()
return response.json()
except requests.RequestException as e:
logger.error('Cannot fetch locations: {}'.format(e))
return []
def fetch_courses(self):
logger.debug('Fetching courses...')
try:
url = '{}activity/'.format(HARRASTUSHAKU_API_BASE_URL)
response = requests.get(url)
response.raise_for_status()
return response.json()['data']
except requests.RequestException as e:
logger.error('Cannot fetch courses: {}'.format(e))
return []
@transaction.atomic
def handle_location(self, location_data):
harrastushaku_location_id = location_data.get('id')
harrastushaku_location_mapped_id = self.location_id_to_place_id.get(harrastushaku_location_id)
if harrastushaku_location_mapped_id.startswith(self.tprek_data_source.id):
return
else:
self.handle_non_tprek_location(location_data)
def handle_non_tprek_location(self, location_data):
get_string = bind_data_getters(location_data)[0]
place_data = {
'name': get_string('name', localized=True),
'info_url': get_string('url', localized=True),
'street_address': get_string('address', localized=True),
'address_locality': get_string('city', localized=True),
'postal_code': get_string('zip'),
'data_source': self.data_source,
'origin_id': location_data['id'],
'publisher': self.organization,
}
self.save_place(place_data)
@transaction.atomic
def handle_activity(self, activity_data):
if activity_data.get('active') not in ('1', 'K'):
logger.debug('Skipping inactive activity {}'.format(activity_data.get('id')))
return
event_data = self.get_event_data(activity_data)
if event_data['start_time'] > event_data['end_time']:
raise HarrastushakuException('Start time after end time')
time_tables = activity_data.get('timetables', [])
if time_tables:
self.handle_recurring_event(event_data, time_tables)
else:
self.handle_one_time_event(event_data)
def get_event_data(self, activity_data):
get_string, get_int, get_datetime = bind_data_getters(activity_data)
keywords = self.get_event_keywords(activity_data)
audience = self.get_event_audiences_from_ages(activity_data) | self.get_event_audiences_from_keywords(keywords)
keywords |= audience
event_data = {
'name': get_string('name', localized=True),
'description': get_string('description', localized=True),
'audience_max_age': get_int('agemax'),
'audience_min_age': get_int('agemin'),
'start_time': get_datetime('startdate'),
'end_time': get_datetime('enddate'),
'date_published': get_datetime('publishdate'),
'extension_course': {
'enrolment_start_date': get_datetime('regstartdate'),
'enrolment_end_date': get_datetime('regenddate'),
'maximum_attendee_capacity': get_int('maxentries'),
'remaining_attendee_capacity': get_int('regavailable'),
},
'data_source': self.data_source,
'origin_id': activity_data['id'],
'publisher': self.organization,
'location': self.get_event_location(activity_data),
'keywords': keywords,
'in_language': self.get_event_languages(activity_data),
'images': self.get_event_images(activity_data),
'offers': self.get_event_offers(activity_data),
'audience': audience,
}
return event_data
def handle_recurring_event(self, event_data, time_tables):
start_date, end_date = self.get_event_start_and_end_dates(event_data)
if not start_date:
raise HarrastushakuException('No start time')
if not end_date:
raise HarrastushakuException('No end time')
if end_date - start_date > timedelta(days=MAX_RECURRING_EVENT_LENGTH):
raise HarrastushakuException('Too long recurring activity')
sub_event_time_ranges = self.build_sub_event_time_ranges(start_date, end_date, time_tables)
if not sub_event_time_ranges:
raise HarrastushakuException('Erroneous time tables: {}'.format(time_tables))
super_event = self.save_super_event(event_data)
self.save_sub_events(event_data, sub_event_time_ranges, super_event)
def handle_one_time_event(self, event_data):
event_data['has_start_time'] = False
event_data['has_end_time'] = False
event = self.save_event(event_data)
self.event_syncher.mark(event)
def get_event_keywords(self, activity_data):
keywords = (self.get_event_keywords_from_main_categories(activity_data) |
self.get_event_keywords_from_search_words(activity_data))
return keywords
def get_event_keywords_from_main_categories(self, activity_data):
main_category_ids = {c.get('maincategory_id') for c in activity_data.get('categories', [])}
keyword_ids = set()
for main_category_id in main_category_ids:
keyword_ids |= MAIN_CATEGORY_KEYWORDS.get(main_category_id, set())
return {self.keywords.get(kw_id) for kw_id in keyword_ids if kw_id in self.keywords}
def get_event_keywords_from_search_words(self, activity_data):
keywords = set()
search_words = activity_data.get('searchwords', [])
cleaned_search_words = [s.strip().lower() for s in search_words.split(',') if s.strip()]
for kw in cleaned_search_words:
matches = self.match_keyword(kw)
if matches:
keywords |= set(matches)
return keywords
def get_event_languages(self, activity_data):
language_text = activity_data.get('languages', '').lower()
languages = {obj for code, obj in self.languages.items() if obj.name_fi and obj.name_fi in language_text}
return languages
def get_event_start_and_end_dates(self, event_data):
start_datetime = event_data.get('start_time')
start_date = start_datetime.date() if start_datetime else None
end_datetime = event_data.get('end_time')
end_date = end_datetime.date() if end_datetime else None
return start_date, end_date
def build_sub_event_time_ranges(self, start_date, end_date, time_tables):
sub_event_time_ranges = []
for time_table in time_tables:
current_date = start_date
weekday = int(time_table.get('weekday'))
start_time = parse_time(time_table.get('starttime'))
end_time = parse_time(time_table.get('endtime'))
repetition = int(time_table.get('repetition'))
if repetition == 0:
repetition = 7 # assume repetition 0 and 7 mean the same thing
if not (weekday and repetition) or start_time >= end_time:
continue
while current_date.isoweekday() != weekday:
current_date += timedelta(days=1)
while current_date <= end_date:
sub_event_time_ranges.append(SubEventTimeRange(
datetime.combine(current_date, start_time).astimezone(TIMEZONE),
datetime.combine(current_date, end_time).astimezone(TIMEZONE),
))
current_date += timedelta(days=repetition)
return sub_event_time_ranges
def save_super_event(self, event_data):
super_event_data = deepcopy(event_data)
super_event_data['super_event_type'] = Event.SuperEventType.RECURRING
event = self.save_event(super_event_data)
self.event_syncher.mark(event)
return event
def save_sub_events(self, event_data, sub_event_time_ranges, super_event):
super_event._changed = False
def delete_sub_event(obj):
logger.debug('{} deleted'.format(obj))
obj.deleted = True
obj.save()
sub_event_syncher = ModelSyncher(
super_event.sub_events.filter(deleted=False), lambda o: o.id, delete_func=delete_sub_event)
sub_event_data = deepcopy(event_data)
sub_event_data['super_event'] = super_event
for sub_event_time_range in sub_event_time_ranges:
sub_event_data['start_time'] = sub_event_time_range.start
sub_event_data['end_time'] = sub_event_time_range.end
sub_event_data['origin_id'] = (
event_data['origin_id'] + self.create_sub_event_origin_id_suffix(sub_event_time_range))
sub_event = self.save_event(sub_event_data)
if sub_event._changed:
super_event._changed = True
sub_event_syncher.mark(sub_event)
old_sub_event_count = super_event.sub_events.count()
sub_event_syncher.finish(force=True)
if super_event.sub_events.count() != old_sub_event_count:
super_event._changed = True
if super_event._changed:
super_event.save()
def create_sub_event_origin_id_suffix(self, sub_event_time_range):
start, end = sub_event_time_range
assert start.date() == end.date()
date = start.date().strftime('%Y%m%d')
times = '{}{}'.format(*(time.time().strftime('%H%M') for time in (start, end)))
return '_{}{}'.format(date, times)
def get_event_images(self, activity_data):
image_data = activity_data.get('images')
if not isinstance(image_data, dict):
return []
event_image_data = [{
'name': image_datum.get('name', ''),
'url': image_datum.get('filename', ''),
} for image_datum in image_data.values()]
return event_image_data
def get_event_location(self, activity_data):
location_id = activity_data.get('location_id')
if not location_id:
return None
return {'id': self.location_id_to_place_id.get(location_id)}
def get_event_offers(self, activity_data):
offers = []
for price_data in activity_data.get('prices', ()):
get_string = bind_data_getters(price_data)[0]
price = get_string('price', localized=True)
description = get_string('description', localized=True)
is_free = price is not None and price['fi'] == '0'
if not description and len(activity_data['prices']) == 1:
description = get_string('pricedetails', localized=True)
offers.append({
'price': price if not is_free else None,
'is_free': is_free,
'description': description,
})
return offers
def get_event_audiences_from_ages(self, activity_data):
audience_keyword_ids = set()
age_min = get_int_from_data(activity_data, 'agemin') or 0
age_max = get_int_from_data(activity_data, 'agemax') or 200
for age_range, keyword_ids in AUDIENCE_BY_AGE_RANGE:
if ranges_overlap(age_min, age_max, age_range[0], age_range[1]):
audience_keyword_ids |= keyword_ids
return {self.keywords.get(k_id) for k_id in audience_keyword_ids if k_id in self.keywords}
def get_event_audiences_from_keywords(self, keywords):
return {kw for kw in keywords if kw.id in KEYWORDS_TO_ADD_TO_AUDIENCE}
@lru_cache()
def match_keyword(self, text):
return self.keyword_matcher.match(text)
def import_courses(self):
"""Import Harrastushaku activities as Courses
Activities having "active" anything else than "1" or "K" will be
ignored.
When importing and an existing course isn't present in imported data:
- If the course's end time is in the past, the course will be left as
it is.
- If the course's end time is not in the past, the course will be soft
deleted alongside its sub events.
If an activity has something in field "timetables", it will be imported
as a recurring event, otherwise as a one-time event.
A recurring course will have a super event which includes the course's
whole time period, and sub events which will represent individual course
occurrences. Other than start and end times, a super event and its sub
events will all contain the same data.
A recurring course's sub event start and end datetimes will be build using
the activity's "timetables". The time tables contain info out weekday,
times, and repetition which means number of days there is between
occurrences (basically a multiple of 7).
A recurring course's sub events will be given an ID that has the
activity's ID and start and end times of the sub event in a compressed
form. This also means that between imports only sub events that are
happening exactly at the same time are considered to be the same instance,
so if a sub event's begin or end time changes at all, a new sub event will
be created instead of updating an old one (because there is no unambiguous
way to determine which old sub event the new one corresponds to).
A course's keywords will come from both of the following:
- The activity's main category. There are hardcoded keywords for every
main category.
- The activity's sub category's "searchwords". Those are manually
entered words, which are mapped to keywords using KeywordMatcher
(from events.keywords).
A course's audience will come from both of the following:
- The activity's "audience_max_age" and "audience_min_age" using
hardcoded keywords for certain age ranges.
- The course's keywords, adding the ones that are present in
KEYWORDS_TO_ADD_TO_AUDIENCE (from events.importer.yso).
"""
logger.info('Importing courses...')
locations = self.fetch_locations()
if not locations:
logger.warning('No location data fetched, aborting course import.')
return
self.location_id_to_place_id = self.map_harrastushaku_location_ids_to_tprek_ids(locations)
activities = self.fetch_courses()
if not activities:
logger.info('No activity data fetched.')
return
def event_delete(event):
if event.end_time < now():
return
event.soft_delete()
for sub_event in event.sub_events.all():
sub_event.soft_delete()
self.event_syncher = ModelSyncher(
Event.objects.filter(data_source=self.data_source, super_event=None),
lambda event: event.id,
event_delete,
)
num_of_activities = len(activities)
logger.debug('Handling {} activities...'.format(num_of_activities))
for i, activity in enumerate(activities, 1):
try:
self.handle_activity(activity)
except Exception as e: # noqa
message = e if isinstance(e, HarrastushakuException) else traceback.format_exc()
logger.error('Error handling activity {}: {}'.format(activity.get('id'), message))
if not i % 10:
logger.debug('{} / {} activities handled.'.format(i, num_of_activities))
self.event_syncher.finish(force=True)
logger.info('Course import finished.')
def import_courses(self):
"""Import Harrastushaku activities as Courses
Activities having "active" anything else than "1" or "K" will be
ignored.
When importing and an existing course isn't present in imported data:
- If the course's end time is in the past, the course will be left as
it is.
- If the course's end time is not in the past, the course will be soft
deleted alongside its sub events.
If an activity has something in field "timetables", it will be imported
as a recurring event, otherwise as a one-time event.
A recurring course will have a super event which includes the course's
whole time period, and sub events which will represent individual course
occurrences. Other than start and end times, a super event and its sub
events will all contain the same data.
A recurring course's sub event start and end datetimes will be build using
the activity's "timetables". The time tables contain info out weekday,
times, and repetition which means number of days there is between
occurrences (basically a multiple of 7).
A recurring course's sub events will be given an ID that has the
activity's ID and start and end times of the sub event in a compressed
form. This also means that between imports only sub events that are
happening exactly at the same time are considered to be the same instance,
so if a sub event's begin or end time changes at all, a new sub event will
be created instead of updating an old one (because there is no unambiguous
way to determine which old sub event the new one corresponds to).
A course's keywords will come from both of the following:
- The activity's main category. There are hardcoded keywords for every
main category.
- The activity's sub category's "searchwords". Those are manually
entered words, which are mapped to keywords using KeywordMatcher
(from events.keywords).
A course's audience will come from both of the following:
- The activity's "audience_max_age" and "audience_min_age" using
hardcoded keywords for certain age ranges.
- The course's keywords, adding the ones that are present in
KEYWORDS_TO_ADD_TO_AUDIENCE (from events.importer.yso).
"""
logger.info('Importing courses...')
locations = self.fetch_locations()
if not locations:
logger.warning('No location data fetched, aborting course import.')
return
self.location_id_to_place_id = {
l['id']: '{}:{}'.format(self.tprek_data_source.id, l['tpr_id'])
if l.get('tpr_id')
else '{}:{}'.format(self.data_source.id, l['id'])
for l in locations
}
activities = self.fetch_courses()
if not activities:
logger.info('No activity data fetched.')
return
def event_delete(event):
if event.end_time < now():
return
event.soft_delete()
for sub_event in event.sub_events.all():
sub_event.soft_delete()
self.event_syncher = ModelSyncher(
Event.objects.filter(data_source=self.data_source, super_event=None),
lambda event: event.id,
event_delete,
)
num_of_activities = len(activities)
logger.debug('Handling {} activities...'.format(num_of_activities))
for i, activity in enumerate(activities, 1):
try:
self.handle_activity(activity)
except Exception as e: # noqa
message = e if isinstance(e, HarrastushakuException) else traceback.format_exc()
logger.error('Error handling activity {}: {}'.format(activity.get('id'), message))
if not i % 10:
logger.debug('{} / {} activities handled.'.format(i, num_of_activities))
self.event_syncher.finish(force=True)
logger.info('Course import finished.')
class HarrastushakuImporter(Importer):
name = 'harrastushaku'
supported_languages = ['fi']
def setup(self):
logger.debug('Running Harrastushaku importer setup...')
self.data_source, _ = DataSource.objects.get_or_create(id=self.name, defaults={'name': 'Harrastushaku'})
self.tprek_data_source = DataSource.objects.get(id='tprek')
self.ahjo_data_source, _ = DataSource.objects.get_or_create(id='ahjo', defaults={'name': 'Ahjo'})
self.organization, _ = Organization.objects.get_or_create(origin_id='u48040030',
data_source=self.ahjo_data_source)
self.tprek_ids = {place.origin_id for place in Place.objects.filter(data_source=self.tprek_data_source)}
self.keywords = {keyword.id: keyword for keyword in Keyword.objects.all()}
self.keyword_matcher = KeywordMatcher()
def import_places(self):
"""Import Harrastushaku locations as Places
- If a place in Harrastushaku has "tprek_id" set, assume it is a place that
comes from Toimipisterekisteri and don't import it from Harrastushaku.
- If a place in Harrastushaku does not have "tprek_id", it will be
imported having data source "harrastushaku".
"""
logger.info('Importing places...')
locations = self.fetch_locations()
logger.debug('Handling {} locations...'.format(len(locations)))
for location in locations:
try:
self.handle_location(location)
except Exception as e: # noqa
message = e if isinstance(e, HarrastushakuException) else traceback.format_exc()
logger.error('Error handling location {}: {}'.format(location.get('id'), message))
def import_courses(self):
"""Import Harrastushaku activities as Courses
Activities having "active" anything else than "1" or "K" will be
ignored.
When importing and an existing course isn't present in imported data:
- If the course's end time is in the past, the course will be left as
it is.
- If the course's end time is not in the past, the course will be soft
deleted alongside its sub events.
If an activity has something in field "timetables", it will be imported
as a recurring event, otherwise as a one-time event.
A recurring course will have a super event which includes the course's
whole time period, and sub events which will represent individual course
occurrences. Other than start and end times, a super event and its sub
events will all contain the same data.
A recurring course's sub event start and end datetimes will be build using
the activity's "timetables". The time tables contain info out weekday,
times, and repetition which means number of days there is between
occurrences (basically a multiple of 7).
A recurring course's sub events will be given an ID that has the
activity's ID and start and end times of the sub event in a compressed
form. This also means that between imports only sub events that are
happening exactly at the same time are considered to be the same instance,
so if a sub event's begin or end time changes at all, a new sub event will
be created instead of updating an old one (because there is no unambiguous
way to determine which old sub event the new one corresponds to).
A course's keywords will come from both of the following:
- The activity's main category. There are hardcoded keywords for every
main category.
- The activity's sub category's "searchwords". Those are manually
entered words, which are mapped to keywords using KeywordMatcher
(from events.keywords).
A course's audience will come from both of the following:
- The activity's "audience_max_age" and "audience_min_age" using
hardcoded keywords for certain age ranges.
- The course's keywords, adding the ones that are present in
KEYWORDS_TO_ADD_TO_AUDIENCE (from events.importer.yso).
"""
logger.info('Importing courses...')
locations = self.fetch_locations()
if not locations:
logger.warning('No location data fetched, aborting course import.')
return
self.location_id_to_place_id = {
l['id']: '{}:{}'.format(self.tprek_data_source.id, l['tpr_id'])
if l.get('tpr_id')
else '{}:{}'.format(self.data_source.id, l['id'])
for l in locations
}
activities = self.fetch_courses()
if not activities:
logger.info('No activity data fetched.')
return
def event_delete(event):
if event.end_time < now():
return
event.soft_delete()
for sub_event in event.sub_events.all():
sub_event.soft_delete()
self.event_syncher = ModelSyncher(
Event.objects.filter(data_source=self.data_source, super_event=None),
lambda event: event.id,
event_delete,
)
num_of_activities = len(activities)
logger.debug('Handling {} activities...'.format(num_of_activities))
for i, activity in enumerate(activities, 1):
try:
self.handle_activity(activity)
except Exception as e: # noqa
message = e if isinstance(e, HarrastushakuException) else traceback.format_exc()
logger.error('Error handling activity {}: {}'.format(activity.get('id'), message))
if not i % 10:
logger.debug('{} / {} activities handled.'.format(i, num_of_activities))
self.event_syncher.finish(force=True)
logger.info('Course import finished.')
def fetch_locations(self):
logger.debug('Fetching locations...')
try:
url = '{}location/'.format(HARRASTUSHAKU_API_BASE_URL)
response = requests.get(url)
response.raise_for_status()
return response.json()
except requests.RequestException as e:
logger.error('Cannot fetch locations: {}'.format(e))
return []
def fetch_courses(self):
logger.debug('Fetching courses...')
try:
url = '{}activity/'.format(HARRASTUSHAKU_API_BASE_URL)
response = requests.get(url)
response.raise_for_status()
return response.json()['data']
except requests.RequestException as e:
logger.error('Cannot fetch courses: {}'.format(e))
return []
@transaction.atomic
def handle_location(self, location_data):
tprek_id = location_data.get('tpr_id')
if tprek_id:
if tprek_id in self.tprek_ids:
return
logger.warning("Place with Toimipisterekisteri ID {} doesn't exist.".format(tprek_id))
self.handle_non_tprek_location(location_data)
def handle_non_tprek_location(self, location_data):
get_string = bind_data_getters(location_data)[0]
place_data = {
'name': get_string('name', localized=True),
'info_url': get_string('url', localized=True),
'street_address': get_string('address', localized=True),
'address_locality': get_string('city', localized=True),
'postal_code': get_string('zip'),
'data_source': self.data_source,
'origin_id': location_data['id'],
'publisher': self.organization,
}
self.save_place(place_data)
@transaction.atomic
def handle_activity(self, activity_data):
if activity_data.get('active') not in ('1', 'K'):
logger.debug('Skipping inactive activity {}'.format(activity_data.get('id')))
return
event_data = self.get_event_data(activity_data)
if event_data['start_time'] > event_data['end_time']:
raise HarrastushakuException('Start time after end time')
time_tables = activity_data.get('timetables', [])
if time_tables:
self.handle_recurring_event(event_data, time_tables)
else:
self.handle_one_time_event(event_data)
def get_event_data(self, activity_data):
get_string, get_int, get_datetime = bind_data_getters(activity_data)
keywords = self.get_event_keywords(activity_data)
audience = self.get_event_audiences_from_ages(activity_data) | self.get_event_audiences_from_keywords(keywords)
keywords |= audience
event_data = {
'name': get_string('name', localized=True),
'description': get_string('description', localized=True),
'audience_max_age': get_int('agemax'),
'audience_min_age': get_int('agemin'),
'start_time': get_datetime('startdate'),
'end_time': get_datetime('enddate'),
'date_published': get_datetime('publishdate'),
'extension_course': {
'enrolment_start_date': get_datetime('regstartdate'),
'enrolment_end_date': get_datetime('regenddate'),
'maximum_attendee_capacity': get_int('maxentries'),
'remaining_attendee_capacity': get_int('regavailable'),
},
'data_source': self.data_source,
'origin_id': activity_data['id'],
'publisher': self.organization,
'location': self.get_event_location(activity_data),
'keywords': keywords,
'in_language': self.get_event_languages(activity_data),
'images': self.get_event_images(activity_data),
'offers': self.get_event_offers(activity_data),
'audience': audience,
}
return event_data
def handle_recurring_event(self, event_data, time_tables):
start_date, end_date = self.get_event_start_and_end_dates(event_data)
if not start_date:
raise HarrastushakuException('No start time')
if not end_date:
raise HarrastushakuException('No end time')
if end_date - start_date > timedelta(days=MAX_RECURRING_EVENT_LENGTH):
raise HarrastushakuException('Too long recurring activity')
sub_event_time_ranges = self.build_sub_event_time_ranges(start_date, end_date, time_tables)
if not sub_event_time_ranges:
raise HarrastushakuException('Erroneous time tables: {}'.format(time_tables))
super_event = self.save_super_event(event_data)
self.save_sub_events(event_data, sub_event_time_ranges, super_event)
def handle_one_time_event(self, event_data):
event_data['has_start_time'] = False
event_data['has_end_time'] = False
event = self.save_event(event_data)
self.event_syncher.mark(event)
def get_event_keywords(self, activity_data):
keywords = (self.get_event_keywords_from_main_categories(activity_data) |
self.get_event_keywords_from_search_words(activity_data))
return keywords
def get_event_keywords_from_main_categories(self, activity_data):
main_category_ids = {c.get('maincategory_id') for c in activity_data.get('categories', [])}
keyword_ids = set()
for main_category_id in main_category_ids:
keyword_ids |= MAIN_CATEGORY_KEYWORDS.get(main_category_id, set())
return {self.keywords.get(kw_id) for kw_id in keyword_ids if kw_id in self.keywords}
def get_event_keywords_from_search_words(self, activity_data):
keywords = set()
search_words = activity_data.get('searchwords', [])
cleaned_search_words = [s.strip().lower() for s in search_words.split(',') if s.strip()]
for kw in cleaned_search_words:
matches = self.match_keyword(kw)
if matches:
keywords |= set(matches)
return keywords
def get_event_languages(self, activity_data):
language_text = activity_data.get('languages', '').lower()
languages = {obj for code, obj in self.languages.items() if obj.name_fi and obj.name_fi in language_text}
return languages
def get_event_start_and_end_dates(self, event_data):
start_datetime = event_data.get('start_time')
start_date = start_datetime.date() if start_datetime else None
end_datetime = event_data.get('end_time')
end_date = end_datetime.date() if end_datetime else None
return start_date, end_date
def build_sub_event_time_ranges(self, start_date, end_date, time_tables):
sub_event_time_ranges = []
for time_table in time_tables:
current_date = start_date
weekday = int(time_table.get('weekday'))
start_time = parse_time(time_table.get('starttime'))
end_time = parse_time(time_table.get('endtime'))
repetition = int(time_table.get('repetition'))
if repetition == 0:
repetition = 7 # assume repetition 0 and 7 mean the same thing
if not (weekday and repetition) or start_time >= end_time:
continue
while current_date.isoweekday() != weekday:
current_date += timedelta(days=1)
while current_date <= end_date:
sub_event_time_ranges.append(SubEventTimeRange(
datetime.combine(current_date, start_time).astimezone(TIMEZONE),
datetime.combine(current_date, end_time).astimezone(TIMEZONE),
))
current_date += timedelta(days=repetition)
return sub_event_time_ranges
def save_super_event(self, event_data):
super_event_data = deepcopy(event_data)
super_event_data['super_event_type'] = Event.SuperEventType.RECURRING
event = self.save_event(super_event_data)
self.event_syncher.mark(event)
return event
def save_sub_events(self, event_data, sub_event_time_ranges, super_event):
super_event._changed = False
def delete_sub_event(obj):
logger.debug('{} deleted'.format(obj))
obj.deleted = True
obj.save()
sub_event_syncher = ModelSyncher(
super_event.sub_events.filter(deleted=False), lambda o: o.id, delete_func=delete_sub_event)
sub_event_data = deepcopy(event_data)
sub_event_data['super_event'] = super_event
for sub_event_time_range in sub_event_time_ranges:
sub_event_data['start_time'] = sub_event_time_range.start
sub_event_data['end_time'] = sub_event_time_range.end
sub_event_data['origin_id'] = (
event_data['origin_id'] + self.create_sub_event_origin_id_suffix(sub_event_time_range))
sub_event = self.save_event(sub_event_data)
if sub_event._changed:
super_event._changed = True
sub_event_syncher.mark(sub_event)
old_sub_event_count = super_event.sub_events.count()
sub_event_syncher.finish(force=True)
if super_event.sub_events.count() != old_sub_event_count:
super_event._changed = True
if super_event._changed:
super_event.save()
def create_sub_event_origin_id_suffix(self, sub_event_time_range):
start, end = sub_event_time_range
assert start.date() == end.date()
date = start.date().strftime('%Y%m%d')
times = '{}{}'.format(*(time.time().strftime('%H%M') for time in (start, end)))
return '_{}{}'.format(date, times)
def get_event_images(self, activity_data):
image_data = activity_data.get('images')
if not isinstance(image_data, dict):
return []
event_image_data = [{
'name': image_datum.get('name', ''),
'url': image_datum.get('filename', ''),
} for image_datum in image_data.values()]
return event_image_data
def get_event_location(self, activity_data):
location_id = activity_data.get('location_id')
if not location_id:
return None
return {'id': self.location_id_to_place_id.get(location_id)}
def get_event_offers(self, activity_data):
offers = []
for price_data in activity_data.get('prices', ()):
get_string = bind_data_getters(price_data)[0]
price = get_string('price', localized=True)
description = get_string('description', localized=True)
is_free = price is not None and price['fi'] == '0'
if not description and len(activity_data['prices']) == 1:
description = get_string('pricedetails', localized=True)
offers.append({
'price': price if not is_free else None,
'is_free': is_free,
'description': description,
})
return offers
def get_event_audiences_from_ages(self, activity_data):
audience_keyword_ids = set()
age_min = get_int_from_data(activity_data, 'agemin') or 0
age_max = get_int_from_data(activity_data, 'agemax') or 200
for age_range, keyword_ids in AUDIENCE_BY_AGE_RANGE:
if ranges_overlap(age_min, age_max, age_range[0], age_range[1]):
audience_keyword_ids |= keyword_ids
return {self.keywords.get(k_id) for k_id in audience_keyword_ids if k_id in self.keywords}
def get_event_audiences_from_keywords(self, keywords):
return {kw for kw in keywords if kw.id in KEYWORDS_TO_ADD_TO_AUDIENCE}
@lru_cache()
def match_keyword(self, text):
return self.keyword_matcher.match(text)