def connect_to_ceph(ceph_bucket_prefix: str, environment: str, bucket: Optional[str] = None) -> CephStore:
"""Connect to Ceph to store SLI metrics for Thoth."""
prefix = _get_sli_metrics_prefix(ceph_bucket_prefix=ceph_bucket_prefix, environment=environment)
ceph = CephStore(prefix=prefix, bucket=bucket)
ceph.connect()
return ceph
def test_init_kwargs(self):
"""Test initialization of Ceph based on arguments."""
adapter = CephStore(_BUCKET_PREFIX, **CEPH_INIT_KWARGS)
for key, value in CEPH_INIT_KWARGS.items():
assert getattr(adapter, key) == value, \
f"Ceph attribute {key!r} has value {getattr(adapter, key)!r} but expected is {value!r}"
assert adapter.prefix == _BUCKET_PREFIX
assert not adapter.is_connected()
def store_thoth_sli_on_ceph(
ceph_sli: CephStore, metric_class: str, metrics_df: pd.DataFrame, ceph_path: str, is_public: bool = False,
) -> None:
"""Store Thoth SLI on Ceph."""
metrics_csv = metrics_df.to_csv(index=False, sep="`", header=False)
if is_public:
_LOGGER.info(f"Storing on public bucket... {ceph_path}")
else:
_LOGGER.info(f"Storing on private bucket... {ceph_path}")
ceph_sli.store_blob(blob=metrics_csv, object_key=ceph_path)
_LOGGER.info(f"Succesfully stored Thoth weekly SLI metrics for {metric_class} at {ceph_path}")
def _fixture_adapter():
"""Retrieve an adapter to Ceph."""
mock_s3().start()
try:
yield CephStore(_BUCKET_PREFIX, **CEPH_INIT_KWARGS)
finally:
mock_s3().stop()
def create_s3_adapter(ceph_bucket_prefix: str, repo: str) -> CephStore:
"""Create Ceph adapter for deployment metrics."""
deployment_name = "aicoe-ci"
prefix = f"{ceph_bucket_prefix}/{deployment_name}/deployment-metrics/{repo}"
ceph = CephStore(prefix=prefix)
return ceph
def connect_to_ceph(
cls,
ceph_bucket_prefix: str,
processed_data_name: str,
environment: str,
bucket: Optional[str] = None,
) -> CephStore:
"""Connect to Ceph to store processed data."""
prefix = cls._get_processed_data_prefix(
ceph_bucket_prefix=ceph_bucket_prefix,
processed_data_name=processed_data_name,
environment=environment,
)
ceph = CephStore(prefix=prefix, bucket=bucket)
ceph.connect()
return ceph
def store_csv_from_dataframe(
csv_from_df: str,
ceph_sli: CephStore,
file_name: str,
ceph_path: str,
is_public: bool = False,
) -> None:
"""Store CSV obtained from pd.DataFrame on Ceph.
param: csv_from_df: CSV given from pd.DataFrame.to_csv()
"""
if is_public:
_LOGGER.info(f"Storing on public bucket... {ceph_path}")
else:
_LOGGER.info(f"Storing on private bucket... {ceph_path}")
ceph_sli.store_blob(blob=csv_from_df, object_key=ceph_path)
_LOGGER.info(f"Succesfully stored {file_name} at {ceph_path}")
def test_init_env(self):
"""Test initialization of Ceph adapter based on env variables."""
adapter = CephStore(_BUCKET_PREFIX)
assert adapter.prefix == _BUCKET_PREFIX
for key, value in CEPH_INIT_ENV.items():
attribute = CEPH_ENV_MAP[key]
assert getattr(adapter, attribute) == value, \
f"Ceph attribute {attribute!r} has value {getattr(adapter, attribute)!r} but expected is " \
f"{value!r} (env: {key!r})"
def retrieve_thoth_sli_from_ceph(ceph_sli: CephStore, ceph_path: str, total_columns: List[str]) -> pd.DataFrame:
"""Retrieve Thoth SLI from Ceph."""
_LOGGER.info(f"Retrieving... \n{ceph_path}")
try:
retrieved_data = ceph_sli.retrieve_blob(object_key=ceph_path).decode('utf-8')
data = StringIO(retrieved_data)
_LOGGER.debug(f"retrieved data:\n {data}")
last_data = pd.read_csv(data, names=total_columns, sep="`")
except Exception as e:
_LOGGER.warning(f"No file could be retrieved from Ceph: {e}")
last_data = pd.DataFrame(columns=total_columns)
return last_data
def _fixture_connected_adapter():
"""Retrieve a connected adapter to Ceph."""
adapter = CephStore(_BUCKET_PREFIX, **CEPH_INIT_KWARGS)
with connected_ceph_adapter(adapter, raw_ceph=True) as connected_adapter:
yield connected_adapter
import os
from typing import List
from thoth.storages import CephStore
store = CephStore(prefix='data/thoth/ash-api/ETM/words')
store.connect()
def retrieve_n_most_similar(token, n) -> List:
if not store.document_exists(token):
return {'error': f'The token: {token} is not in the vocabulary.'}, 404
token_dict = store.retrieve_document(token)
return [t for t in token_dict['most_similar'][:n]]
def get(token):
""" Retrieves the 25 most similar tokens to {token} path param """
return {
'token': token,
'most-similar-tokens': retrieve_n_most_similar(token, n=25),
}, 200
