def test_column_uniqueness_when_truncated(self):
problem_spec = {
"display_name": "practicing_lessons",
"property_name": "long_column",
"choices": [
"duplicate_choice_1",
"duplicate_choice_2",
],
"select_style": "multiple",
"column_id": "a_very_long_base_selection_column_name_with_limited_room",
"type": "choice_list",
}
data_source_config = DataSourceConfiguration(
domain='test',
display_name='foo',
referenced_doc_type='CommCareCase',
table_id=uuid.uuid4().hex,
configured_filter={},
configured_indicators=[problem_spec],
)
adapter = IndicatorSqlAdapter(data_source_config)
adapter.rebuild_table()
# ensure we can save data to the table.
adapter.save({
'_id': uuid.uuid4().hex,
'domain': 'test',
'doc_type': 'CommCareCase',
'long_column': 'duplicate_choice_1',
})
# and query it back
q = adapter.get_query_object()
self.assertEqual(1, q.count())
def test_column_uniqueness_when_truncated(self):
problem_spec = {
"display_name": "practicing_lessons",
"property_name": "long_column",
"choices": [
"duplicate_choice_1",
"duplicate_choice_2",
],
"select_style": "multiple",
"column_id":
"a_very_long_base_selection_column_name_with_limited_room",
"type": "choice_list",
}
data_source_config = DataSourceConfiguration(
domain='test',
display_name='foo',
referenced_doc_type='CommCareCase',
table_id=uuid.uuid4().hex,
configured_filter={},
configured_indicators=[problem_spec],
)
adapter = IndicatorSqlAdapter(data_source_config)
adapter.rebuild_table()
# ensure we can save data to the table.
adapter.save({
'_id': uuid.uuid4().hex,
'domain': 'test',
'doc_type': 'CommCareCase',
'long_column': 'duplicate_choice_1',
})
# and query it back
q = adapter.get_query_object()
self.assertEqual(1, q.count())
def test_table_population(self):
adapter = IndicatorSqlAdapter(self.config)
# Delete and create table
adapter.rebuild_table()
# Create a doc
now = datetime.datetime.now()
one_hour = datetime.timedelta(hours=1)
logs = [
{"start_time": now, "end_time": now + one_hour, "person": "al"},
{"start_time": now + one_hour, "end_time": now + (one_hour * 2), "person": "chris"},
{"start_time": now + (one_hour * 2), "end_time": now + (one_hour * 3), "person": "katie"},
]
doc = _test_doc(form={"time_logs": logs})
# Save this document into the table
adapter.save(doc)
# Get rows from the table
rows = adapter.get_query_object()
retrieved_logs = [{"start_time": r.start_time, "end_time": r.end_time, "person": r.person} for r in rows]
# Check those rows against the expected result
self.assertItemsEqual(
retrieved_logs, logs, "The repeat data saved in the data source table did not match the expected data!"
)
def rebuild_indicators(indicator_config_id):
is_static = indicator_config_id.startswith(CustomDataSourceConfiguration._datasource_id_prefix)
if is_static:
config = CustomDataSourceConfiguration.by_id(indicator_config_id)
else:
config = DataSourceConfiguration.get(indicator_config_id)
# Save the start time now in case anything goes wrong. This way we'll be
# able to see if the rebuild started a long time ago without finishing.
config.meta.build.initiated = datetime.datetime.utcnow()
config.save()
adapter = IndicatorSqlAdapter(config)
adapter.rebuild_table()
couchdb = _get_db(config.referenced_doc_type)
relevant_ids = get_doc_ids(config.domain, config.referenced_doc_type,
database=couchdb)
for doc in iter_docs(couchdb, relevant_ids, chunksize=500):
try:
# save is a noop if the filter doesn't match
adapter.save(doc)
except DataError as e:
logging.exception('problem saving document {} to table. {}'.format(doc['_id'], e))
if not is_static:
config.meta.build.finished = True
config.save()
def rebuild_indicators(indicator_config_id):
is_static = indicator_config_id.startswith(
CustomDataSourceConfiguration._datasource_id_prefix)
if is_static:
config = CustomDataSourceConfiguration.by_id(indicator_config_id)
else:
config = DataSourceConfiguration.get(indicator_config_id)
# Save the start time now in case anything goes wrong. This way we'll be
# able to see if the rebuild started a long time ago without finishing.
config.meta.build.initiated = datetime.datetime.utcnow()
config.save()
adapter = IndicatorSqlAdapter(config)
adapter.rebuild_table()
couchdb = _get_db(config.referenced_doc_type)
relevant_ids = get_doc_ids(config.domain,
config.referenced_doc_type,
database=couchdb)
for doc in iter_docs(couchdb, relevant_ids, chunksize=500):
try:
# save is a noop if the filter doesn't match
adapter.save(doc)
except DataError as e:
logging.exception('problem saving document {} to table. {}'.format(
doc['_id'], e))
if not is_static:
config.meta.build.finished = True
config.save()
def rebuild_indicators(indicator_config_id):
config = _get_config_by_id(indicator_config_id)
adapter = IndicatorSqlAdapter(config)
couchdb = _get_db(config.referenced_doc_type)
redis_client = get_redis_client().client.get_client()
redis_key = _get_redis_key_for_config(config)
if not is_static(indicator_config_id):
# Save the start time now in case anything goes wrong. This way we'll be
# able to see if the rebuild started a long time ago without finishing.
config.meta.build.initiated = datetime.datetime.utcnow()
config.meta.build.finished = False
config.save()
redis_key = _get_redis_key_for_config(config)
adapter.rebuild_table()
relevant_ids_chunk = []
for relevant_id in iterate_doc_ids_in_domain_by_type(
config.domain,
config.referenced_doc_type,
chunk_size=CHUNK_SIZE,
database=couchdb):
relevant_ids_chunk.append(relevant_id)
if len(relevant_ids_chunk) >= CHUNK_SIZE:
redis_client.sadd(redis_key, *relevant_ids_chunk)
_build_indicators(indicator_config_id, relevant_ids_chunk)
relevant_ids_chunk = []
if relevant_ids_chunk:
redis_client.sadd(redis_key, *relevant_ids_chunk)
_build_indicators(indicator_config_id, relevant_ids_chunk)
def rebuild_indicators(indicator_config_id):
config = _get_config_by_id(indicator_config_id)
adapter = IndicatorSqlAdapter(config)
couchdb = _get_db(config.referenced_doc_type)
redis_client = get_redis_client().client.get_client()
redis_key = _get_redis_key_for_config(config)
if not is_static(indicator_config_id):
# Save the start time now in case anything goes wrong. This way we'll be
# able to see if the rebuild started a long time ago without finishing.
config.meta.build.initiated = datetime.datetime.utcnow()
config.meta.build.finished = False
config.save()
redis_key = _get_redis_key_for_config(config)
adapter.rebuild_table()
relevant_ids = get_doc_ids_in_domain_by_type(
config.domain,
config.referenced_doc_type,
database=couchdb,
)
for docs in chunked(relevant_ids, 1000):
redis_client.sadd(redis_key, *docs)
_build_indicators(indicator_config_id, relevant_ids)
def rebuild_aggregate_ucr(request, domain, table_id):
table_definition = get_object_or_404(
AggregateTableDefinition, domain=domain, table_id=table_id
)
aggregate_table_adapter = IndicatorSqlAdapter(table_definition)
aggregate_table_adapter.rebuild_table()
populate_aggregate_table_data_task.delay(table_definition.id)
messages.success(request, 'Table rebuild successfully started.')
return HttpResponseRedirect(reverse(AggregateUCRView.urlname, args=[domain, table_id]))
def test_weekly_aggregation(self):
# generate our table
aggregate_table_adapter = IndicatorSqlAdapter(self.weekly_aggregate_table_definition)
aggregate_table_adapter.rebuild_table()
populate_aggregate_table_data(aggregate_table_adapter)
self._check_weekly_results()
# confirm it's also idempotent
populate_aggregate_table_data(aggregate_table_adapter)
self._check_weekly_results()
def test_weekly_aggregation(self):
# generate our table
aggregate_table_adapter = IndicatorSqlAdapter(
self.weekly_aggregate_table_definition)
aggregate_table_adapter.rebuild_table()
populate_aggregate_table_data(aggregate_table_adapter)
self._check_weekly_results()
# confirm it's also idempotent
populate_aggregate_table_data(aggregate_table_adapter)
self._check_weekly_results()
def rebuild_indicators(indicator_config_id):
config = _get_config_by_id(indicator_config_id)
adapter = IndicatorSqlAdapter(config)
if not is_static(indicator_config_id):
# Save the start time now in case anything goes wrong. This way we'll be
# able to see if the rebuild started a long time ago without finishing.
config.meta.build.initiated = datetime.datetime.utcnow()
config.meta.build.finished = False
config.save()
adapter.rebuild_table()
_iteratively_build_table(config)
def rebuild_indicators(indicator_config_id):
config = DataSourceConfiguration.get(indicator_config_id)
adapter = IndicatorSqlAdapter(get_engine(), config)
adapter.rebuild_table()
couchdb = _get_db(config.referenced_doc_type)
relevant_ids = get_doc_ids(config.domain, config.referenced_doc_type,
database=couchdb)
for doc in iter_docs(couchdb, relevant_ids, chunksize=500):
if config.filter.filter(doc):
try:
adapter.save(doc)
except DataError as e:
logging.exception('problem saving document {} to table. {}'.format(doc['_id'], e))
def rebuild_indicators(indicator_config_id, initiated_by=None):
config = _get_config_by_id(indicator_config_id)
success = _('Your UCR table {} has finished rebuilding').format(config.table_id)
failure = _('There was an error rebuilding Your UCR table {}.').format(config.table_id)
send = toggles.SEND_UCR_REBUILD_INFO.enabled(initiated_by)
with notify_someone(initiated_by, success_message=success, error_message=failure, send=send):
adapter = IndicatorSqlAdapter(config)
if not id_is_static(indicator_config_id):
# Save the start time now in case anything goes wrong. This way we'll be
# able to see if the rebuild started a long time ago without finishing.
config.meta.build.initiated = datetime.datetime.utcnow()
config.meta.build.finished = False
config.save()
adapter.rebuild_table()
_iteratively_build_table(config)
class IndicatorPillowTest(TestCase):
def setUp(self):
folder = os.path.join(os.path.dirname(__file__), 'data', 'configs')
sample_file = os.path.join(folder, 'sample_indicator_config.json')
self.pillow = ConfigurableIndicatorPillow()
self.engine = self.pillow.get_sql_engine()
with open(sample_file) as f:
structure = json.loads(f.read())
self.config = DataSourceConfiguration.wrap(structure)
self.pillow.bootstrap(configs=[self.config])
self.adapter = IndicatorSqlAdapter(self.engine, self.config)
self.adapter.rebuild_table()
def tearDown(self):
self.adapter.drop_table()
self.engine.dispose()
def testFilter(self):
# note: this is a silly test now that python_filter always returns true
not_matching = [
dict(doc_type="NotCommCareCase", domain='user-reports', type='ticket'),
dict(doc_type="CommCareCase", domain='not-user-reports', type='ticket'),
dict(doc_type="CommCareCase", domain='user-reports', type='not-ticket'),
]
for document in not_matching:
self.assertTrue(self.pillow.python_filter(document))
self.assertTrue(self.pillow.python_filter(
dict(doc_type="CommCareCase", domain='user-reports', type='ticket')
))
def testChangeTransport(self):
# indicators
sample_doc, expected_indicators = get_sample_doc_and_indicators()
self.pillow.change_transport(sample_doc)
with self.engine.begin() as connection:
rows = connection.execute(sqlalchemy.select([self.adapter.get_table()]))
self.assertEqual(1, rows.rowcount)
row = rows.fetchone()
for k, v in row.items():
self.assertEqual(expected_indicators[k], v)
def test_table_population(self):
adapter = IndicatorSqlAdapter(self.config)
# Delete and create table
adapter.rebuild_table()
# Create a doc
now = datetime.datetime.now()
one_hour = datetime.timedelta(hours=1)
logs = [
{
"start_time": now,
"end_time": now + one_hour,
"person": "al"
},
{
"start_time": now + one_hour,
"end_time": now + (one_hour * 2),
"person": "chris"
},
{
"start_time": now + (one_hour * 2),
"end_time": now + (one_hour * 3),
"person": "katie"
},
]
doc = _test_doc(form={'time_logs': logs})
# Save this document into the table
adapter.save(doc)
# Get rows from the table
rows = adapter.get_query_object()
retrieved_logs = [{
'start_time': r.start_time,
'end_time': r.end_time,
'person': r.person,
} for r in rows]
# Check those rows against the expected result
self.assertItemsEqual(
retrieved_logs, logs,
"The repeat data saved in the data source table did not match the expected data!"
)
def rebuild_indicators(indicator_config_id):
is_static = indicator_config_id.startswith(CustomDataSourceConfiguration._datasource_id_prefix)
if is_static:
config = CustomDataSourceConfiguration.by_id(indicator_config_id)
else:
config = DataSourceConfiguration.get(indicator_config_id)
adapter = IndicatorSqlAdapter(get_engine(), config)
adapter.rebuild_table()
couchdb = _get_db(config.referenced_doc_type)
relevant_ids = get_doc_ids(config.domain, config.referenced_doc_type,
database=couchdb)
for doc in iter_docs(couchdb, relevant_ids, chunksize=500):
try:
# save is a noop if the filter doesn't match
adapter.save(doc)
except DataError as e:
logging.exception('problem saving document {} to table. {}'.format(doc['_id'], e))
adapter.engine.dispose()
def test_table_population(self):
engine = get_engine()
adapter = IndicatorSqlAdapter(engine, self.config)
# Delete and create table
adapter.rebuild_table()
# Create a doc
now = datetime.datetime.now()
one_hour = datetime.timedelta(hours=1)
logs = [
{"start_time": now, "end_time": now + one_hour, "person": "al"},
{"start_time": now + one_hour, "end_time": now + (one_hour * 2), "person": "chris"},
{"start_time": now + (one_hour * 2), "end_time": now + (one_hour * 3), "person": "katie"},
]
doc = _test_doc(form={'time_logs': logs})
# Save this document into the table
adapter.save(doc)
# Get rows from the table
with engine.connect() as connection:
rows = connection.execute(adapter.get_table().select())
retrieved_logs = [
{
'start_time': r[3],
'end_time': r[4],
'person': r[5],
} for r in rows
]
# Clean up
engine.dispose()
# Check those rows against the expected result
self.assertItemsEqual(
retrieved_logs,
logs,
"The repeat data saved in the data source table did not match the expected data!"
)
def rebuild_indicators(indicator_config_id):
is_static = indicator_config_id.startswith(StaticDataSourceConfiguration._datasource_id_prefix)
if is_static:
config = StaticDataSourceConfiguration.by_id(indicator_config_id)
rev = 'static'
else:
config = DataSourceConfiguration.get(indicator_config_id)
rev = config._rev
# Save the start time now in case anything goes wrong. This way we'll be
# able to see if the rebuild started a long time ago without finishing.
config.meta.build.initiated = datetime.datetime.utcnow()
config.save()
adapter = IndicatorSqlAdapter(config)
couchdb = _get_db(config.referenced_doc_type)
client = get_redis_client().client.get_client()
redis_key = 'ucr_queue-{}:{}'.format(indicator_config_id, rev)
if len(client.smembers(redis_key)) > 0:
relevant_ids = client.smembers(redis_key)
else:
adapter.rebuild_table()
relevant_ids = get_doc_ids_in_domain_by_type(config.domain, config.referenced_doc_type,
database=couchdb)
if relevant_ids:
client.sadd(redis_key, *relevant_ids)
for doc in iter_docs(couchdb, relevant_ids, chunksize=500):
try:
# save is a noop if the filter doesn't match
adapter.save(doc)
client.srem(redis_key, doc.get('_id'))
except DataError as e:
logging.exception('problem saving document {} to table. {}'.format(doc['_id'], e))
if not is_static:
client.delete(redis_key)
config.meta.build.finished = True
config.save()
class IndicatorPillowTest(TestCase):
def setUp(self):
self.config = get_sample_data_source()
self.pillow = ConfigurableIndicatorPillow()
self.engine = self.pillow.get_sql_engine()
self.pillow.bootstrap(configs=[self.config])
self.adapter = IndicatorSqlAdapter(self.engine, self.config)
self.adapter.rebuild_table()
def tearDown(self):
self.adapter.drop_table()
self.engine.dispose()
def test_filter(self):
# note: this is a silly test now that python_filter always returns true
not_matching = [
dict(doc_type="NotCommCareCase", domain='user-reports', type='ticket'),
dict(doc_type="CommCareCase", domain='not-user-reports', type='ticket'),
dict(doc_type="CommCareCase", domain='user-reports', type='not-ticket'),
]
for document in not_matching:
self.assertTrue(self.pillow.python_filter(document))
self.assertTrue(self.pillow.python_filter(
dict(doc_type="CommCareCase", domain='user-reports', type='ticket')
))
def test_change_transport(self):
sample_doc, _ = get_sample_doc_and_indicators()
self.pillow.change_transport(sample_doc)
self._check_sample_doc_state()
def test_rebuild_indicators(self):
self.config.save()
sample_doc, _ = get_sample_doc_and_indicators()
CommCareCase.get_db().save_doc(sample_doc)
rebuild_indicators(self.config._id)
self._check_sample_doc_state()
def test_bad_integer_datatype(self):
self.config.save()
bad_ints = ['a', '', None]
for bad_value in bad_ints:
self.pillow.change_transport({
'_id': uuid.uuid4().hex,
'doc_type': 'CommCareCase',
'domain': 'user-reports',
'type': 'ticket',
'priority': bad_value
})
# make sure we saved rows to the table for everything
with self.engine.begin() as connection:
rows = connection.execute(sqlalchemy.select([self.adapter.get_table()]))
self.assertEqual(len(bad_ints), rows.rowcount)
def _check_sample_doc_state(self):
_, expected_indicators = get_sample_doc_and_indicators()
with self.engine.begin() as connection:
rows = connection.execute(sqlalchemy.select([self.adapter.get_table()]))
self.assertEqual(1, rows.rowcount)
row = rows.fetchone()
for k, v in row.items():
if isinstance(expected_indicators[k], decimal.Decimal):
self.assertAlmostEqual(expected_indicators[k], v)
else:
self.assertEqual(expected_indicators[k], v)
