def test_write_to_datastore(self, mock_session_return):
mock_session = mock_session_return.return_value = create_mock_session()
scrape_key = ScrapeKey(REGIONS[0], constants.ScrapeType.BACKGROUND)
ii = ingest_info.IngestInfo()
ii.create_person(full_name=TEST_NAME).create_booking(
booking_id=TEST_ID)
t = Task(
task_type=constants.TaskType.SCRAPE_DATA,
endpoint=TEST_ENDPOINT,
response_type=constants.ResponseType.TEXT,
)
task_hash = hash(json.dumps(t.to_serializable(), sort_keys=True))
expected_batch = BatchIngestInfoData(ingest_info=ii,
task_hash=task_hash)
batch_persistence.write(ii, scrape_key, t)
batch_ingest_info_list = batch_persistence._get_batch_ingest_info_list(
scrape_key.region_code, mock_session.start)
self.assertEqual(len(batch_ingest_info_list), 1)
self.assertEqual(expected_batch, batch_ingest_info_list[0])
def test_persist_to_db(self, mock_write, _mock_region,
mock_session_return):
mock_session = mock_session_return.return_value = create_mock_session()
scrape_key = ScrapeKey(REGIONS[0], constants.ScrapeType.BACKGROUND)
ii = ingest_info.IngestInfo()
ii.create_person(
person_id=TEST_ID,
full_name=TEST_NAME).create_booking(booking_id=TEST_ID)
t = Task(
task_type=constants.TaskType.SCRAPE_DATA,
endpoint=TEST_ENDPOINT,
response_type=constants.ResponseType.TEXT,
)
batch_persistence.write(ii, scrape_key, t)
expected_proto = serialization.convert_ingest_info_to_proto(ii)
batch_persistence.persist_to_database(scrape_key.region_code,
mock_session.start)
result_proto = mock_write.call_args[0][0]
self.assertEqual(result_proto, expected_proto)
# After we persist, there should no longer be ingest infos on Datastore
ingest_infos = datastore_ingest_info.batch_get_ingest_infos_for_region(
REGIONS[0], mock_session.start)
self.assertEqual(len(ingest_infos), 0)
def test_persist_to_db_different_regions(self, mock_write, _mock_region,
mock_session_return):
scrape_key1 = ScrapeKey(REGIONS[0], constants.ScrapeType.BACKGROUND)
scrape_key2 = ScrapeKey(REGIONS[1], constants.ScrapeType.BACKGROUND)
ii = ingest_info.IngestInfo()
ii.create_person(
person_id=TEST_ID,
full_name=TEST_NAME).create_booking(booking_id=TEST_ID)
ii2 = ingest_info.IngestInfo()
ii2.create_person(
person_id=TEST_ID,
full_name=TEST_NAME2).create_booking(booking_id=TEST_ID)
t = Task(
task_type=constants.TaskType.SCRAPE_DATA,
endpoint=TEST_ENDPOINT,
response_type=constants.ResponseType.TEXT,
)
t2 = Task(
task_type=constants.TaskType.SCRAPE_DATA,
endpoint=TEST_ENDPOINT,
response_type=constants.ResponseType.TEXT,
)
mock_session_1 = mock_session_return.return_value = create_mock_session(
)
batch_persistence.write(ii, scrape_key1, t)
expected_proto = serialization.convert_ingest_info_to_proto(ii)
batch_persistence.persist_to_database(scrape_key1.region_code,
mock_session_1.start)
result_proto = mock_write.call_args[0][0]
self.assertEqual(result_proto, expected_proto)
# We expect the region that we persisted to have no more ingest infos.
ingest_infos_1 = datastore_ingest_info.batch_get_ingest_infos_for_region(
REGIONS[0], mock_session_1.start)
self.assertEqual(len(ingest_infos_1), 0)
mock_session_2 = mock_session_return.return_value = create_mock_session(
)
batch_persistence.write(ii2, scrape_key2, t2)
ingest_infos_2 = datastore_ingest_info.batch_get_ingest_infos_for_region(
REGIONS[1], mock_session_2.start)
self.assertEqual(len(ingest_infos_2), 1)
expected_proto = serialization.convert_ingest_info_to_proto(ii2)
batch_persistence.persist_to_database(scrape_key2.region_code,
mock_session_2.start)
result_proto = mock_write.call_args[0][0]
self.assertEqual(result_proto, expected_proto)
self.assertEqual(mock_write.call_count, 2)
def test_persist_to_db_same_task_one_fail_one_pass(self, mock_write,
_mock_region,
mock_session_return):
mock_session = mock_session_return.return_value = create_mock_session()
scrape_key = ScrapeKey(REGIONS[0], constants.ScrapeType.BACKGROUND)
mock_write.return_value = True
ii = ingest_info.IngestInfo()
ii.create_person(
person_id=TEST_ID,
full_name=TEST_NAME).create_booking(booking_id=TEST_ID)
t = Task(
task_type=constants.TaskType.SCRAPE_DATA,
endpoint=TEST_ENDPOINT,
response_type=constants.ResponseType.TEXT,
)
# Because the tasks are the same, we expect that to be counted as a
# pass.
t2 = Task(
task_type=constants.TaskType.SCRAPE_DATA,
endpoint=TEST_ENDPOINT,
response_type=constants.ResponseType.TEXT,
)
batch_persistence.write(ii, scrape_key, t)
batch_persistence.write_error(TEST_ERROR, TEST_TRACE, t2, scrape_key)
expected_proto = serialization.convert_ingest_info_to_proto(ii)
self.assertTrue(
batch_persistence.persist_to_database(scrape_key.region_code,
mock_session.start))
result_proto = mock_write.call_args[0][0]
self.assertEqual(result_proto, expected_proto)
ingest_infos = datastore_ingest_info.batch_get_ingest_infos_for_region(
REGIONS[0], mock_session.start)
self.assertEqual(len(ingest_infos), 0)
def test_persist_duplicates_to_db(self, mock_write, _mock_region,
mock_session_return):
"""Tests that duplicate ingest_info.Person objects are merged before
write."""
mock_session = mock_session_return.return_value = create_mock_session()
scrape_key = ScrapeKey(REGIONS[0], constants.ScrapeType.BACKGROUND)
# Arrange
ii = IngestInfo()
ii.create_person(person_id=TEST_ID,
full_name=TEST_NAME) \
.create_booking(booking_id=TEST_ID)
ii_2 = IngestInfo()
ii.create_person(person_id=TEST_ID2, full_name=TEST_NAME2)
ii_1_dup = copy.deepcopy(ii)
t1, t2, t3 = (Task(task_type=constants.TaskType.SCRAPE_DATA,
endpoint=TEST_ENDPOINT + str(i),
response_type=constants.ResponseType.TEXT)
for i in range(3))
batch_persistence.write(ii, scrape_key, t1)
batch_persistence.write(ii_2, scrape_key, t2)
batch_persistence.write(ii_1_dup, scrape_key, t3)
batch_persistence.persist_to_database(scrape_key.region_code,
mock_session.start)
expected_ii = IngestInfo(people=ii.people + ii_2.people)
expected_proto = ingest_utils.convert_ingest_info_to_proto(expected_ii)
result_proto = mock_write.call_args[0][0]
self.assertEqual(result_proto, expected_proto)
def test_persist_to_db_failed_no_write(self, mock_write, _mock_region,
mock_session_return):
mock_session = mock_session_return.return_value = create_mock_session()
scrape_key = ScrapeKey(REGIONS[0], constants.ScrapeType.BACKGROUND)
ii = IngestInfo()
ii.create_person(person_id=TEST_ID,
full_name=TEST_NAME) \
.create_booking(booking_id=TEST_ID)
t = Task(
task_type=constants.TaskType.SCRAPE_DATA,
endpoint=TEST_ENDPOINT,
response_type=constants.ResponseType.TEXT,
)
# Because the tasks are different, we should fail.
t2 = Task(
task_type=constants.TaskType.SCRAPE_DATA,
endpoint=TEST_ENDPOINT,
response_type=constants.ResponseType.TEXT,
params=TEST_PARAMS,
)
batch_persistence.write(ii, scrape_key, t)
batch_persistence.write_error(TEST_ERROR, TEST_TRACE, t2, scrape_key)
self.assertFalse(
batch_persistence.persist_to_database(scrape_key.region_code,
mock_session.start))
self.assertEqual(mock_write.call_count, 0)
# We should still have both items still on Datastore because they
# weren't persisted.
batch_ingest_info_data_list = batch_persistence \
._get_batch_ingest_info_list(scrape_key.region_code,
mock_session.start)
self.assertEqual(len(batch_ingest_info_data_list), 2)
def _generic_scrape(self, request: QueueRequest):
"""
General handler for all scrape tasks. This function is a generic entry
point into all types of scrapes. It decides what to call based on
params.
Args:
params: dict of parameters passed from the last scrape session.
"""
try:
task = request.next_task
# Here we handle a special case where we weren't really sure
# we were going to get data when we submitted a task, but then
# we ended up with data, so no more requests are required,
# just the content we already have.
# TODO(#680): remove this
if task.content is not None:
content = self._parse_html_content(task.content)
cookies = None
else:
post_data = task.post_data
# Let the child transform the post_data if it wants before
# sending the requests. This hook is in here in case the
# child did something like compress the post_data before
# it put it on the queue.
self.transform_post_data(post_data)
# We always fetch some content before doing anything.
# Note that we use get here for the post_data to return a
# default value of None if this scraper doesn't set it.
try:
content, cookies = self._fetch_content(
task.endpoint,
task.response_type,
headers=task.headers,
cookies=task.cookies,
params=task.params,
post_data=post_data,
json_data=task.json)
except Exception as e:
raise ScraperFetchError(str(e)) from e
scraped_data = None
if self.should_scrape_data(task.task_type):
# If we want to scrape data, we should either create an
# ingest_info object or get the one that already exists.
logging.info("Scraping data for [%s] and endpoint: [%s]",
self.region.region_code, task.endpoint)
try:
scraped_data = self.populate_data(
content, task, request.ingest_info or IngestInfo())
except Exception as e:
raise ScraperPopulateDataError(str(e)) from e
if self.should_get_more_tasks(task.task_type):
logging.info("Getting more tasks for [%s] and endpoint: [%s]",
self.region.region_code, task.endpoint)
# Only send along ingest info if it will not be persisted now.
ingest_info_to_send = None
if scraped_data is not None and not scraped_data.persist:
ingest_info_to_send = scraped_data.ingest_info
try:
# pylint: disable=assignment-from-no-return
next_tasks = self.get_more_tasks(content, task)
except Exception as e:
raise ScraperGetMoreTasksError(str(e)) from e
for next_task in next_tasks:
# Include cookies received from response, if any
if cookies:
cookies.update(next_task.cookies)
next_task = Task.evolve(next_task, cookies=cookies)
self.add_task(
'_generic_scrape',
QueueRequest(
scrape_type=request.scrape_type,
scraper_start_time=request.scraper_start_time,
next_task=next_task,
ingest_info=ingest_info_to_send,
))
if scraped_data is not None and scraped_data.persist:
if scraped_data.ingest_info:
logging.info("Logging at most 4 people (were %d):",
len(scraped_data.ingest_info.people))
loop_count = min(len(scraped_data.ingest_info.people),
constants.MAX_PEOPLE_TO_LOG)
for i in range(loop_count):
logging.info("[%s]",
str(scraped_data.ingest_info.people[i]))
logging.info("Last seen time of person being set as: [%s]",
request.scraper_start_time)
metadata = IngestMetadata(self.region.region_code,
self.region.jurisdiction_id,
request.scraper_start_time,
self.get_enum_overrides())
if self.BATCH_WRITES:
logging.info(
"Queuing ingest_info ([%d] people) to "
"batch_persistence for [%s]",
len(scraped_data.ingest_info.people),
self.region.region_code)
scrape_key = ScrapeKey(self.region.region_code,
request.scrape_type)
batch_persistence.write(
ingest_info=scraped_data.ingest_info,
scrape_key=scrape_key,
task=task,
)
else:
logging.info(
"Writing ingest_info ([%d] people) to the database"
" for [%s]", len(scraped_data.ingest_info.people),
self.region.region_code)
persistence.write(
ingest_utils.convert_ingest_info_to_proto(
scraped_data.ingest_info), metadata)
for sc in scraped_data.single_counts:
if not sc.date:
scrape_key = ScrapeKey(self.region.region_code,
constants.ScrapeType.BACKGROUND)
session = sessions.get_current_session(scrape_key)
if session:
sc = attr.evolve(sc, date=session.start.date())
single_count.store_single_count(
sc, self.region.jurisdiction_id)
except Exception as e:
if self.BATCH_WRITES:
scrape_key = ScrapeKey(self.region.region_code,
request.scrape_type)
batch_persistence.write_error(
error=str(e),
trace_id=get_trace_id_from_flask(),
task=task,
scrape_key=scrape_key,
)
raise e
