def __init__(self, processor, pillow_name, topics, num_processes, process_num, retry_errors=False, processor_chunk_size=0): change_feed = KafkaChangeFeed( topics, client_id=pillow_name, num_processes=num_processes, process_num=process_num ) checkpoint = KafkaPillowCheckpoint(pillow_name, topics) event_handler = KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=1000, change_feed=change_feed, checkpoint_callback=processor ) super(ConfigurableReportKafkaPillow, self).__init__( name=pillow_name, change_feed=change_feed, processor=processor, checkpoint=checkpoint, change_processed_event_handler=event_handler, processor_chunk_size=processor_chunk_size ) # set by the superclass constructor assert self.processors is not None assert len(self.processors) == 1 self._processor = self.processors[0] assert self._processor.bootstrapped is not None # retry errors defaults to False because there is not a solution to # distinguish between doc save errors and data source config errors self.retry_errors = retry_errors
def get_location_pillow(pillow_id='location-ucr-pillow', include_ucrs=None, num_processes=1, process_num=0, ucr_configs=None, **kwargs): # Todo; is ucr_division needed? change_feed = KafkaChangeFeed([LOCATION_TOPIC], client_id=pillow_id, num_processes=num_processes, process_num=process_num) ucr_processor = ConfigurableReportPillowProcessor( data_source_providers=[ DynamicDataSourceProvider('Location'), StaticDataSourceProvider('Location') ], include_ucrs=include_ucrs, ) if ucr_configs: ucr_processor.bootstrap(ucr_configs) checkpoint = KafkaPillowCheckpoint(pillow_id, [LOCATION_TOPIC]) event_handler = KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=1000, change_feed=change_feed, checkpoint_callback=ucr_processor) return ConstructedPillow(name=pillow_id, change_feed=change_feed, checkpoint=checkpoint, change_processed_event_handler=event_handler, processor=[ucr_processor])
def get_group_pillow(pillow_id='group-pillow', num_processes=1, process_num=0, **kwargs): """Group pillow Processors: - :py:class:`corehq.pillows.groups_to_user.GroupsToUsersProcessor` - :py:func:`corehq.pillows.group.get_group_to_elasticsearch_processor` """ assert pillow_id == 'group-pillow', 'Pillow ID is not allowed to change' to_user_es_processor = GroupsToUsersProcessor() to_group_es_processor = get_group_to_elasticsearch_processor() change_feed = KafkaChangeFeed(topics=[topics.GROUP], client_id='groups-to-users', num_processes=num_processes, process_num=process_num) checkpoint_id = "{}-{}-{}".format(pillow_id, USER_INDEX, to_group_es_processor.index_info.index) checkpoint = KafkaPillowCheckpoint(checkpoint_id, [topics.GROUP]) return ConstructedPillow( name=pillow_id, checkpoint=checkpoint, change_feed=change_feed, processor=[to_user_es_processor, to_group_es_processor], change_processed_event_handler=KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=10, change_feed=change_feed), )
def get_case_messaging_sync_pillow( pillow_id='case_messaging_sync_pillow', topics=None, num_processes=1, process_num=0, processor_chunk_size=DEFAULT_PROCESSOR_CHUNK_SIZE, **kwargs): if topics: assert set(topics).issubset( CASE_TOPICS), set(topics) - set(CASE_TOPICS) topics = topics or CASE_TOPICS change_feed = KafkaChangeFeed(topics, client_id=pillow_id, num_processes=num_processes, process_num=process_num) checkpoint = KafkaPillowCheckpoint(pillow_id, topics) event_handler = KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=1000, change_feed=change_feed, ) return ConstructedPillow(name=pillow_id, change_feed=change_feed, checkpoint=checkpoint, change_processed_event_handler=event_handler, processor=[CaseMessagingSyncProcessor()], processor_chunk_size=processor_chunk_size)
def get_location_pillow(pillow_id='location-ucr-pillow', include_ucrs=None, num_processes=1, process_num=0, ucr_configs=None, **kwargs): """Processes updates to locations for UCR Note this is only applicable if a domain on the environment has `LOCATIONS_IN_UCR` flag enabled. Processors: - :py:func:`corehq.apps.userreports.pillow.ConfigurableReportPillowProcessor` """ change_feed = KafkaChangeFeed( [LOCATION_TOPIC], client_id=pillow_id, num_processes=num_processes, process_num=process_num ) ucr_processor = ConfigurableReportPillowProcessor( data_source_providers=[DynamicDataSourceProvider('Location'), StaticDataSourceProvider('Location')], include_ucrs=include_ucrs, ) if ucr_configs: ucr_processor.bootstrap(ucr_configs) checkpoint = KafkaPillowCheckpoint(pillow_id, [LOCATION_TOPIC]) event_handler = KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=1000, change_feed=change_feed, checkpoint_callback=ucr_processor ) return ConstructedPillow( name=pillow_id, change_feed=change_feed, checkpoint=checkpoint, change_processed_event_handler=event_handler, processor=[ucr_processor] )
def get_form_submission_metadata_tracker_pillow(pillow_id='FormSubmissionMetadataTrackerPillow', num_processes=1, process_num=0, **kwargs): """ This gets a pillow which iterates through all forms and marks the corresponding app as having submissions. Processors: - :py:class:`pillowtop.processors.form.FormSubmissionMetadataTrackerProcessor` """ # todo; To remove after full rollout of https://github.com/dimagi/commcare-hq/pull/21329/ change_feed = KafkaChangeFeed( topics=topics.FORM_TOPICS, client_id='form-processsor', num_processes=num_processes, process_num=process_num ) checkpoint = KafkaPillowCheckpoint('form-submission-metadata-tracker', topics.FORM_TOPICS) form_processor = FormSubmissionMetadataTrackerProcessor() return ConstructedPillow( name=pillow_id, checkpoint=checkpoint, change_feed=change_feed, processor=form_processor, change_processed_event_handler=KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=100, change_feed=change_feed, ), )
def get_form_submission_metadata_tracker_pillow( pillow_id='FormSubmissionMetadataTrackerProcessor', num_processes=1, process_num=0, **kwargs): """ This gets a pillow which iterates through all forms and marks the corresponding app as having submissions. This could be expanded to be more generic and include other processing that needs to happen on each form """ change_feed = KafkaChangeFeed(topics=topics.FORM_TOPICS, group_id='form-processsor', num_processes=num_processes, process_num=process_num) checkpoint = KafkaPillowCheckpoint('form-submission-metadata-tracker', topics.FORM_TOPICS) form_processor = FormSubmissionMetadataTrackerProcessor() return ConstructedPillow( name=pillow_id, checkpoint=checkpoint, change_feed=change_feed, processor=form_processor, change_processed_event_handler=KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=100, change_feed=change_feed, ), )
def get_case_pillow( pillow_id='case-pillow', ucr_division=None, include_ucrs=None, exclude_ucrs=None, num_processes=1, process_num=0, ucr_configs=None, skip_ucr=False, processor_chunk_size=DEFAULT_PROCESSOR_CHUNK_SIZE, topics=None, **kwargs): """Return a pillow that processes cases. The processors include, UCR and elastic processors Processors: - :py:class:`corehq.apps.userreports.pillow.ConfigurableReportPillowProcessor` (disabled when skip_ucr=True) - :py:class:`pillowtop.processors.elastic.BulkElasticProcessor` - :py:function:`corehq.pillows.case_search.get_case_search_processor` - :py:class:`corehq.messaging.pillow.CaseMessagingSyncProcessor` """ if topics: assert set(topics).issubset(CASE_TOPICS), "This is a pillow to process cases only" topics = topics or CASE_TOPICS change_feed = KafkaChangeFeed( topics, client_id=pillow_id, num_processes=num_processes, process_num=process_num ) ucr_processor = ConfigurableReportPillowProcessor( data_source_providers=[DynamicDataSourceProvider('CommCareCase'), StaticDataSourceProvider('CommCareCase')], ucr_division=ucr_division, include_ucrs=include_ucrs, exclude_ucrs=exclude_ucrs, run_migrations=(process_num == 0), # only first process runs migrations ) if ucr_configs: ucr_processor.bootstrap(ucr_configs) case_to_es_processor = BulkElasticProcessor( elasticsearch=get_es_new(), index_info=CASE_INDEX_INFO, doc_prep_fn=transform_case_for_elasticsearch ) case_search_processor = get_case_search_processor() checkpoint_id = "{}-{}-{}-{}".format( pillow_id, CASE_INDEX_INFO.index, case_search_processor.index_info.index, 'messaging-sync') checkpoint = KafkaPillowCheckpoint(checkpoint_id, topics) event_handler = KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=1000, change_feed=change_feed, checkpoint_callback=ucr_processor ) processors = [case_to_es_processor, CaseMessagingSyncProcessor()] if settings.RUN_CASE_SEARCH_PILLOW: processors.append(case_search_processor) if not settings.ENTERPRISE_MODE: processors.append(get_case_to_report_es_processor()) if not skip_ucr: # this option is useful in tests to avoid extra UCR setup where unneccessary processors = [ucr_processor] + processors return ConstructedPillow( name=pillow_id, change_feed=change_feed, checkpoint=checkpoint, change_processed_event_handler=event_handler, processor=processors, processor_chunk_size=processor_chunk_size )
def get_case_pillow( pillow_id='case-pillow', ucr_division=None, include_ucrs=None, exclude_ucrs=None, num_processes=1, process_num=0, ucr_configs=None, skip_ucr=False, processor_chunk_size=UCR_PROCESSING_CHUNK_SIZE, topics=None, **kwargs): """ Return a pillow that processes cases. The processors include, UCR and elastic processors Args: skip_ucr: Can be set to True to avoid passing UCR processor, useful for tests """ if topics: assert set(topics).issubset(CASE_TOPICS), "This is a pillow to process cases only" topics = topics or CASE_TOPICS change_feed = KafkaChangeFeed( topics, client_id=pillow_id, num_processes=num_processes, process_num=process_num ) ucr_processor = ConfigurableReportPillowProcessor( data_source_providers=[DynamicDataSourceProvider(), StaticDataSourceProvider()], ucr_division=ucr_division, include_ucrs=include_ucrs, exclude_ucrs=exclude_ucrs, ) if ucr_configs: ucr_processor.bootstrap(ucr_configs) case_to_es_processor = ElasticProcessor( elasticsearch=get_es_new(), index_info=CASE_INDEX_INFO, doc_prep_fn=transform_case_for_elasticsearch ) case_search_processor = get_case_search_processor() checkpoint_id = "{}-{}-{}".format( pillow_id, CASE_INDEX_INFO.index, case_search_processor.index_info.index) checkpoint = KafkaPillowCheckpoint(checkpoint_id, topics) event_handler = KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=1000, change_feed=change_feed, checkpoint_callback=ucr_processor ) processors = [case_to_es_processor, case_search_processor] if not settings.ENTERPRISE_MODE: processors.append(get_case_to_report_es_processor()) if not skip_ucr: # this option is useful in tests to avoid extra UCR setup where unneccessary processors = [ucr_processor] + processors return ConstructedPillow( name=pillow_id, change_feed=change_feed, checkpoint=checkpoint, change_processed_event_handler=event_handler, processor=processors, processor_chunk_size=processor_chunk_size )
def get_user_sync_history_pillow( pillow_id='UpdateUserSyncHistoryPillow', num_processes=1, process_num=0, **kwargs): """ This gets a pillow which iterates through all synclogs """ change_feed = KafkaChangeFeed( topics=[topics.SYNCLOG_SQL], client_id=SYNCLOG_SQL_USER_SYNC_GROUP_ID, num_processes=num_processes, process_num=process_num) checkpoint = KafkaPillowCheckpoint(pillow_id, [topics.SYNCLOG_SQL]) return ConstructedPillow( name=pillow_id, checkpoint=checkpoint, change_feed=change_feed, processor=UserSyncHistoryProcessor(), change_processed_event_handler=KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=100, change_feed=change_feed ), )
def get_user_sync_history_pillow( pillow_id='UpdateUserSyncHistoryPillow', num_processes=1, process_num=0, **kwargs): """Synclog pillow Processors: - :py:func:`corehq.pillows.synclog.UserSyncHistoryProcessor` """ change_feed = KafkaChangeFeed( topics=[topics.SYNCLOG_SQL], client_id=SYNCLOG_SQL_USER_SYNC_GROUP_ID, num_processes=num_processes, process_num=process_num) checkpoint = KafkaPillowCheckpoint(pillow_id, [topics.SYNCLOG_SQL]) return ConstructedPillow( name=pillow_id, checkpoint=checkpoint, change_feed=change_feed, processor=UserSyncHistoryProcessor(), change_processed_event_handler=KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=100, change_feed=change_feed ), )
def __init__(self, processor, pillow_name, topics, num_processes, process_num): change_feed = PartitionedKafkaChangeFeed(topics, group_id=pillow_name, num_processes=num_processes, process_num=process_num) checkpoint = KafkaPillowCheckpoint(pillow_name, topics) event_handler = KafkaCheckpointEventHandler(checkpoint=checkpoint, checkpoint_frequency=1000, change_feed=change_feed) super(ConfigurableReportKafkaPillow, self).__init__(name=pillow_name, change_feed=change_feed, processor=processor, checkpoint=checkpoint, change_processed_event_handler=event_handler) # set by the superclass constructor assert self.processors is not None assert len(self.processors) == 1 self._processor = self.processors[0] assert self._processor.bootstrapped is not None
def get_xform_pillow(pillow_id='xform-pillow', ucr_division=None, include_ucrs=None, exclude_ucrs=None, num_processes=1, process_num=0, ucr_configs=None, skip_ucr=False, processor_chunk_size=DEFAULT_PROCESSOR_CHUNK_SIZE, topics=None, **kwargs): # avoid circular dependency from corehq.pillows.reportxform import transform_xform_for_report_forms_index, report_xform_filter from corehq.pillows.mappings.user_mapping import USER_INDEX if topics: assert set(topics).issubset( FORM_TOPICS), "This is a pillow to process cases only" topics = topics or FORM_TOPICS change_feed = KafkaChangeFeed(topics, client_id=pillow_id, num_processes=num_processes, process_num=process_num) ucr_processor = ConfigurableReportPillowProcessor( data_source_providers=[ DynamicDataSourceProvider('XFormInstance'), StaticDataSourceProvider('XFormInstance') ], ucr_division=ucr_division, include_ucrs=include_ucrs, exclude_ucrs=exclude_ucrs, run_migrations=( process_num == 0), # only first process runs migrations ) xform_to_es_processor = ElasticProcessor( elasticsearch=get_es_new(), index_info=XFORM_INDEX_INFO, doc_prep_fn=transform_xform_for_elasticsearch, doc_filter_fn=xform_pillow_filter, ) unknown_user_form_processor = UnknownUsersProcessor() form_meta_processor = FormSubmissionMetadataTrackerProcessor() checkpoint_id = "{}-{}-{}-{}".format(pillow_id, XFORM_INDEX_INFO.index, REPORT_XFORM_INDEX_INFO.index, USER_INDEX) checkpoint = KafkaPillowCheckpoint(checkpoint_id, topics) event_handler = KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=1000, change_feed=change_feed, checkpoint_callback=ucr_processor) if ucr_configs: ucr_processor.bootstrap(ucr_configs) processors = [xform_to_es_processor] if settings.RUN_UNKNOWN_USER_PILLOW: processors.append(unknown_user_form_processor) if settings.RUN_FORM_META_PILLOW: processors.append(form_meta_processor) if not settings.ENTERPRISE_MODE: xform_to_report_es_processor = ElasticProcessor( elasticsearch=get_es_new(), index_info=REPORT_XFORM_INDEX_INFO, doc_prep_fn=transform_xform_for_report_forms_index, doc_filter_fn=report_xform_filter) processors.append(xform_to_report_es_processor) if not skip_ucr: processors.append(ucr_processor) return ConstructedPillow(name=pillow_id, change_feed=change_feed, checkpoint=checkpoint, change_processed_event_handler=event_handler, processor=processors, processor_chunk_size=processor_chunk_size)
def get_xform_pillow(pillow_id='xform-pillow', ucr_division=None, include_ucrs=None, exclude_ucrs=None, num_processes=1, process_num=0, ucr_configs=None, skip_ucr=False, processor_chunk_size=DEFAULT_PROCESSOR_CHUNK_SIZE, topics=None, dedicated_migration_process=False, **kwargs): """Generic XForm change processor Processors: - :py:class:`corehq.apps.userreports.pillow.ConfigurableReportPillowProcessor` (disabled when skip_ucr=True) - :py:class:`pillowtop.processors.elastic.BulkElasticProcessor` - :py:class:`corehq.pillows.user.UnknownUsersProcessor` (disabled when RUN_UNKNOWN_USER_PILLOW=False) - :py:class:`pillowtop.form.FormSubmissionMetadataTrackerProcessor` (disabled when RUN_FORM_META_PILLOW=False) - :py:class:`corehq.apps.data_interfaces.pillow.CaseDeduplicationPillow`` """ # avoid circular dependency from corehq.pillows.reportxform import transform_xform_for_report_forms_index, report_xform_filter from corehq.pillows.mappings.user_mapping import USER_INDEX if topics: assert set(topics).issubset( FORM_TOPICS), "This is a pillow to process cases only" topics = topics or FORM_TOPICS change_feed = KafkaChangeFeed( topics, client_id=pillow_id, num_processes=num_processes, process_num=process_num, dedicated_migration_process=dedicated_migration_process) ucr_processor = get_ucr_processor( data_source_providers=[ DynamicDataSourceProvider('XFormInstance'), StaticDataSourceProvider('XFormInstance') ], ucr_division=ucr_division, include_ucrs=include_ucrs, exclude_ucrs=exclude_ucrs, run_migrations=( process_num == 0), # only first process runs migrations ucr_configs=ucr_configs) xform_to_es_processor = BulkElasticProcessor( elasticsearch=get_es_new(), index_info=XFORM_INDEX_INFO, doc_prep_fn=transform_xform_for_elasticsearch, doc_filter_fn=xform_pillow_filter, change_filter_fn=is_couch_change_for_sql_domain) unknown_user_form_processor = UnknownUsersProcessor() form_meta_processor = FormSubmissionMetadataTrackerProcessor() checkpoint_id = "{}-{}-{}-{}".format(pillow_id, XFORM_INDEX_INFO.index, REPORT_XFORM_INDEX_INFO.index, USER_INDEX) checkpoint = KafkaPillowCheckpoint(checkpoint_id, topics) event_handler = KafkaCheckpointEventHandler( checkpoint=checkpoint, checkpoint_frequency=1000, change_feed=change_feed, checkpoint_callback=ucr_processor) processors = [xform_to_es_processor] if settings.RUN_UNKNOWN_USER_PILLOW: processors.append(unknown_user_form_processor) if settings.RUN_FORM_META_PILLOW: processors.append(form_meta_processor) if settings.RUN_DEDUPLICATION_PILLOW: processors.append(CaseDeduplicationProcessor()) if not settings.ENTERPRISE_MODE: xform_to_report_es_processor = BulkElasticProcessor( elasticsearch=get_es_new(), index_info=REPORT_XFORM_INDEX_INFO, doc_prep_fn=transform_xform_for_report_forms_index, doc_filter_fn=report_xform_filter, change_filter_fn=is_couch_change_for_sql_domain) processors.append(xform_to_report_es_processor) if not skip_ucr: processors.append(ucr_processor) return ConstructedPillow( name=pillow_id, change_feed=change_feed, checkpoint=checkpoint, change_processed_event_handler=event_handler, processor=processors, processor_chunk_size=processor_chunk_size, process_num=process_num, is_dedicated_migration_process=dedicated_migration_process and (process_num == 0))