Merge 919b757 into d9ded15

cumulus_lambda_functions/cumulus_wrapper/query_granules.py

@@ -21,6 +21,7 @@ def __init__(self, cumulus_base: str, cumulus_token: str):
         super().__init__(cumulus_base, cumulus_token)
         self._conditions.append('status=completed')
         self._item_transformer = ItemTransformer()
+        self.__collection_id = None
 
     def with_filter(self, filter_key, filter_values: list):
         if len(filter_values) < 1:
@@ -34,6 +35,7 @@ def with_filter(self, filter_key, filter_values: list):
 
     def with_collection_id(self, collection_id: str):
         self._conditions.append(f'{self.__collection_id_key}={collection_id}')
+        self.__collection_id = collection_id
         return self
 
     def with_bbox(self):
@@ -130,6 +132,48 @@ def query_direct_to_private_api(self, private_api_prefix: str, transform=True):
             return {'server_error': f'error while invoking:{str(e)}'}
         return {'results': stac_list}
 
+    def delete_entry(self, private_api_prefix: str, granule_id: str):
+        payload = {
+            'httpMethod': 'DELETE',
+            'resource': '/{proxy+}',
+            'path': f'/{self.__granules_key}/{self.__collection_id}/{granule_id}',
+            'queryStringParameters': {**{k[0]: k[1] for k in [k1.split('=') for k1 in self._conditions]}},
+            # 'queryStringParameters': {'limit': '30'},
+            'headers': {
+                'Content-Type': 'application/json',
+            },
+            # 'body': json.dumps({"action": "removeFromCmr"})
+        }
+        LOGGER.debug(f'payload: {payload}')
+        try:
+            query_result = self._invoke_api(payload, private_api_prefix)
+            """
+        {'statusCode': 200, 'body': '{"meta":{"name":"cumulus-api","stack":"am-uds-dev-cumulus","table":"granule","limit":3,"page":1,"count":0},"results":[]}', 'headers': {'x-powered-by': 'Express', 'access-control-allow-origin': '*', 'strict-transport-security': 'max-age=31536000; includeSubDomains', 'content-type': 'application/json; charset=utf-8', 'content-length': '120', 'etag': 'W/"78-YdHqDNIH4LuOJMR39jGNA/23yOQ"', 'date': 'Tue, 07 Jun 2022 22:30:44 GMT', 'connection': 'close'}, 'isBase64Encoded': False}
+            """
+            if query_result['statusCode'] >= 500:
+                LOGGER.error(f'server error status code: {query_result.statusCode}. details: {query_result}')
+                return {'server_error': query_result}
+            if query_result['statusCode'] >= 400:
+                LOGGER.error(f'client error status code: {query_result.statusCode}. details: {query_result}')
+                return {'client_error': query_result}
+            query_result = json.loads(query_result['body'])
+            LOGGER.info(f'json query_result: {query_result}')
+            """
+            {
+              "detail": "Record deleted"
+            }
+            """
+            if 'detail' not in query_result:
+                LOGGER.error(f'missing key: detail. invalid response json: {query_result}')
+                return {'server_error': f'missing key: detail. invalid response json: {query_result}'}
+            if query_result['detail'] != 'Record deleted':
+                LOGGER.error(f'Wrong Message: {query_result}')
+                return {'server_error': f'Wrong Message: {query_result}'}
+        except Exception as e:
+            LOGGER.exception('error while invoking')
+            return {'server_error': f'error while invoking:{str(e)}'}
+        return {}
+
     def query(self, transform=True):
         conditions_str = '&'.join(self._conditions)
         LOGGER.info(f'cumulus_base: {self.cumulus_base}')

cumulus_lambda_functions/lib/uds_db/granules_db_index.py

@@ -201,6 +201,23 @@ def get_entry(self, tenant: str, tenant_venue: str, doc_id: str, ):
             raise ValueError(f"no such granule: {doc_id}")
         return result
 
+    def delete_entry(self, tenant: str, tenant_venue: str, doc_id: str, ):
+        read_alias_name = f'{DBConstants.granules_read_alias_prefix}_{tenant}_{tenant_venue}'.lower().strip()
+        result = self.__es.query({
+            'size': 9999,
+            'query': {'term': {'_id': doc_id}}
+        }, read_alias_name)
+        if result is None:
+            raise ValueError(f"no such granule: {doc_id}")
+        for each_granule in result['hits']['hits']:
+            delete_result = self.__es.delete_by_query({
+                'query': {'term': {'_id': each_granule['_id']}}
+            }, each_granule['_index'])
+            LOGGER.debug(f'delete_result: {delete_result}')
+            if delete_result is None:
+                raise ValueError(f"error deleting {each_granule}")
+        return result
+
     def update_entry(self, tenant: str, tenant_venue: str, json_body: dict, doc_id: str, ):
         write_alias_name = f'{DBConstants.granules_write_alias_prefix}_{tenant}_{tenant_venue}'.lower().strip()
         json_body['event_time'] = TimeUtils.get_current_unix_milli()

cumulus_lambda_functions/uds_api/granules_api.py

@@ -2,6 +2,11 @@
 import os
 from typing import Union
 
+from mdps_ds_lib.lib.aws.aws_s3 import AwsS3
+from pystac import Item
+
+from cumulus_lambda_functions.cumulus_wrapper.query_granules import GranulesQuery
+
 from cumulus_lambda_functions.daac_archiver.daac_archiver_logic import DaacArchiverLogic
 from cumulus_lambda_functions.uds_api.dapa.daac_archive_crud import DaacArchiveCrud, DaacDeleteModel, DaacAddModel, \
     DaacUpdateModel
@@ -239,6 +244,51 @@ async def get_single_granule_dapa(request: Request, collection_id: str, granule_
         raise HTTPException(status_code=500, detail=str(e))
     return granules_result
 
+@router.delete("/{collection_id}/items/{granule_id}")
+@router.delete("/{collection_id}/items/{granule_id}/")
+async def delete_single_granule_dapa(request: Request, collection_id: str, granule_id: str, delete_files: bool=True):
+    authorizer: UDSAuthorizorAbstract = UDSAuthorizerFactory() \
+        .get_instance(UDSAuthorizerFactory.cognito,
+                      es_url=os.getenv('ES_URL'),
+                      es_port=int(os.getenv('ES_PORT', '443'))
+                      )
+    auth_info = FastApiUtils.get_authorization_info(request)
+    collection_identifier = UdsCollections.decode_identifier(collection_id)
+    if not authorizer.is_authorized_for_collection(DBConstants.read, collection_id,
+                                                   auth_info['ldap_groups'],
+                                                   collection_identifier.tenant,
+                                                   collection_identifier.venue):
+        LOGGER.debug(f'user: {auth_info["username"]} is not authorized for {collection_id}')
+        raise HTTPException(status_code=403, detail=json.dumps({
+            'message': 'not authorized to execute this action'
+        }))
+    try:
+        LOGGER.debug(f'deleting granule: {granule_id}')
+        cumulus_lambda_prefix = os.getenv('CUMULUS_LAMBDA_PREFIX')
+        cumulus = GranulesQuery('https://na/dev', 'NA')
+        cumulus.with_collection_id(collection_id)
+        cumulus_delete_result = cumulus.delete_entry(cumulus_lambda_prefix, granule_id)  # TODO not sure it is correct granule ID
+        LOGGER.debug(f'cumulus_delete_result: {cumulus_delete_result}')
+        es_delete_result = GranulesDbIndex().delete_entry(collection_identifier.tenant,
+                                                          collection_identifier.venue,
+                                                          granule_id
+                                                          )
+        LOGGER.debug(f'es_delete_result: {es_delete_result}')
+        es_delete_result = [Item.from_dict(k['_source']) for k in es_delete_result['hits']['hits']]
+        if delete_files is False:
+            LOGGER.debug(f'Not deleting files as it is set to false in the request')
+            return {}
+        s3 = AwsS3()
+        for each_granule in es_delete_result:
+            s3_urls = [v.href for k, v in each_granule.assets.items()]
+            LOGGER.debug(f'deleting S3 for {each_granule.id} - s3_urls: {s3_urls}')
+            delete_result = s3.delete_multiple(s3_urls=s3_urls)
+            LOGGER.debug(f'delete_result for {each_granule.id} - delete_result: {delete_result}')
+    except Exception as e:
+        LOGGER.exception('failed during get_granules_dapa')
+        raise HTTPException(status_code=500, detail=str(e))
+    return {}
+
 
 @router.put("/{collection_id}/archive/{granule_id}")
 @router.put("/{collection_id}/archive/{granule_id}/")

requirements.txt

@@ -15,7 +15,7 @@ jsonschema==4.23.0
 jsonschema-specifications==2023.12.1
 lark==0.12.0
 mangum==0.18.0
-mdps-ds-lib==1.1.1
+mdps-ds-lib==1.1.1.dev000200
 pydantic==2.9.2
 pydantic_core==2.23.4
 pygeofilter==0.2.4

tests/integration_tests/test_granules_deletion.py

@@ -0,0 +1,68 @@
+import base64
+import json
+import os
+from unittest import TestCase
+
+import requests
+from dotenv import load_dotenv
+from mdps_ds_lib.lib.cognito_login.cognito_login import CognitoLogin
+
+
+class TestGranulesDeletion(TestCase):
+    def setUp(self) -> None:
+        super().setUp()
+        load_dotenv()
+        self._url_prefix = f'{os.environ.get("UNITY_URL")}/{os.environ.get("UNITY_STAGE", "sbx-uds-dapa")}'
+        self.cognito_login = CognitoLogin() \
+            .with_client_id(os.environ.get('CLIENT_ID', '')) \
+            .with_cognito_url(os.environ.get('COGNITO_URL', '')) \
+            .with_verify_ssl(False) \
+            .start(base64.standard_b64decode(os.environ.get('USERNAME')).decode(),
+                   base64.standard_b64decode(os.environ.get('PASSWORD')).decode())
+        self.bearer_token = self.cognito_login.token
+        self.stage = os.environ.get("UNITY_URL").split('/')[-1]
+        self.uds_url = f'{os.environ.get("UNITY_URL")}/{os.environ.get("UNITY_STAGE", "sbx-uds-dapa")}/'
+        self.custom_metadata_body = {
+            'tag': {'type': 'keyword'},
+            'c_data1': {'type': 'long'},
+            'c_data2': {'type': 'boolean'},
+            'c_data3': {'type': 'keyword'},
+        }
+
+        self.tenant = 'UDS_LOCAL_TEST_3'  # 'uds_local_test'  # 'uds_sandbox'
+        self.tenant_venue = 'DEV'  # 'DEV1'  # 'dev'
+        self.collection_name = 'AAA'  # 'uds_collection'  # 'sbx_collection'
+        self.collection_version = '24.03.20.14.40'.replace('.', '')  # '2402011200'
+        return
+
+    def test_query_all(self):
+        collection_id = f'URN:NASA:UNITY:{self.tenant}:{self.tenant_venue}:{self.collection_name}-01___001'
+        granule_id = f'{collection_id}:test_file01'
+        post_url = f'{self.uds_url}collections/{collection_id}/items/'  # MCP Dev
+        headers = {
+            'Authorization': f'Bearer {self.bearer_token}',
+        }
+        print(post_url)
+        query_result = requests.get(url=post_url,
+                                    headers=headers,
+                                    )
+        self.assertEqual(query_result.status_code, 200, f'wrong status code. {query_result.text}')
+        response_json = json.loads(query_result.text)
+        print(json.dumps(response_json, indent=4))
+        return
+
+    def test_delete_01(self):
+        collection_id = f'URN:NASA:UNITY:{self.tenant}:{self.tenant_venue}:{self.collection_name}-01___001'
+        granule_id = f'{collection_id}:test_file01'
+        post_url = f'{self.uds_url}collections/{collection_id}/items/{granule_id}'  # MCP Dev
+        headers = {
+            'Authorization': f'Bearer {self.bearer_token}',
+        }
+        print(post_url)
+        query_result = requests.delete(url=post_url,
+                                    headers=headers,
+                                    )
+        self.assertEqual(query_result.status_code, 200, f'wrong status code. {query_result.text}')
+        response_json = json.loads(query_result.text)
+        print(json.dumps(response_json, indent=4))
+        return