giambattista.bloisi
/
lot1-kickoff
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

initial stage

This commit is contained in:
Giambattista Bloisi 2024-03-26 11:12:43 +01:00
parent 26c2e3eaad
commit 185ca78f71
1 changed files with 11 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
airflow/dags/import_EOSC_graph.py
View File

@ -20,7 +20,7 @@ from opensearchpy import OpenSearch, helpers
from EOSC_indexes import mappings
from EOSC_entity_trasform import transform_entities
S3_BUCKET_NAME = os.getenv("S3_BUCKET_NAME", "eosc-portal-import")
S3_CONN_ID = os.getenv("S3_CONN_ID", "s3_conn")
EXECUTION_TIMEOUT = int(os.getenv("EXECUTION_TIMEOUT", 6))
@ -40,6 +40,10 @@ default_args = {
start_date=pendulum.datetime(2021, 1, 1, tz="UTC"),
catchup=False,
default_args=default_args,
params={
"S3_CONN_ID": "s3_conn",
"EOSC_CATALOG_BUCKET": "eosc-portal-import"
},
tags=["lot1"]
)
def import_EOSC_graph():
@ -87,10 +91,10 @@ def import_EOSC_graph():
hook = S3Hook(S3_CONN_ID, transfer_config_args={'use_threads': False})
pieces = []
for entity in ENTITIES:
keys = hook.list_keys(bucket_name=S3_BUCKET_NAME, prefix=f'{entity}/')
keys = hook.list_keys(bucket_name=kwargs["params"]["EOSC_CATALOG_BUCKET"], prefix=f'{entity}/')
to_delete = list(filter(lambda key: key.endswith('.PROCESSED'), keys))
for obj in to_delete:
hook.get_conn().delete_object(Bucket=S3_BUCKET_NAME, Key=obj)
hook.get_conn().delete_object(Bucket=kwargs["params"]["EOSC_CATALOG_BUCKET"], Key=obj)
for key in keys:
if key.endswith('.gz'):
pieces.append((entity, key))
@ -118,7 +122,7 @@ def import_EOSC_graph():
)
)
})
def bulk_load(files: list[(str, str)]):
def bulk_load(files: list[(str, str)], **kwargs):
conn = BaseHook.get_connection('opensearch_default')
client = OpenSearch(
hosts=[{'host': conn.host, 'port': conn.port}],
@ -130,11 +134,11 @@ def import_EOSC_graph():
hook = S3Hook(S3_CONN_ID, transfer_config_args={'use_threads': False})
for (entity, key) in files:
if hook.check_for_key(key=f"{key}.PROCESSED", bucket_name=S3_BUCKET_NAME):
if hook.check_for_key(key=f"{key}.PROCESSED", bucket_name=kwargs["params"]["EOSC_CATALOG_BUCKET"]):
print(f'Skipping {entity}: {key}')
continue
print(f'Processing {entity}: {key}')
s3_obj = hook.get_key(key, bucket_name=S3_BUCKET_NAME)
s3_obj = hook.get_key(key, bucket_name=kwargs["params"]["EOSC_CATALOG_BUCKET"])
with s3_obj.get()["Body"] as body:
with gzip.GzipFile(fileobj=body) as gzipfile:
def _generate_data():
@ -169,7 +173,7 @@ def import_EOSC_graph():
hook.load_string(
"",
f"{key}.PROCESSED",
bucket_name=S3_BUCKET_NAME,
bucket_name=kwargs["params"]["EOSC_CATALOG_BUCKET"],
replace=False
)