datamarket 0.9.4__py3-none-any.whl → 0.9.5__py3-none-any.whl

datamarket/interfaces/ftp.py

@@ -12,7 +12,6 @@ from ..utils.main import Config
 
 logger = logging.getLogger(__name__)
 
-
 class FTPInterface:
     def __init__(self, config: Config):
         if "ftp" in config:
@@ -23,7 +22,7 @@ class FTPInterface:
             logger.warning("no ftp section in config")
 
     def get_ftp(self):
-        if self.config["ftps"].lower() == "true":
+        if self.config["ftps"]:
             ftp_conn = FTP_TLS(self.config["server"])
 
         else:

datamarket/interfaces/peerdb.py

@@ -216,26 +216,52 @@ class TransientS3:
             self.config = section
             self.bucket_name = self.config["bucket"]
             self.session = boto3.Session(profile_name=self.config["profile"])
-            self.s3_resource = self.session.resource("s3")
+            self.s3_client = self.session.client("s3")
             self.credentials = self.session.get_credentials()
             self.access_key = self.credentials.access_key
             self.secret_key = self.credentials.secret_key
             self.region_name = self.session.region_name
-            self.endpoint_url = self.s3_resource.meta.endpoint_url
+            self.endpoint_url = self.s3_client.meta.endpoint_url
         else:
             logger.warning("no peerdb.s3 section in config")
 
     def delete_paths_with_schema(self, schema_name):
         logger.info(f"Deleting paths containing '{schema_name}' from S3")
 
-        bucket = self.s3_resource.Bucket(self.bucket_name)
+        paginator = self.s3_client.get_paginator("list_objects_v2")
+        pages = paginator.paginate(Bucket=self.bucket_name, Delimiter="/")
 
-        for prefix in [schema_name, f"clone_{schema_name}"]:
-            objects_to_delete = bucket.objects.filter(Prefix=prefix)
-            objects_to_delete.delete()
+        for page in pages:
+            if "CommonPrefixes" in page:
+                for prefix in page["CommonPrefixes"]:
+                    folder = prefix["Prefix"]
+                    if schema_name in folder:
+                        self._delete_folder_contents(folder)
 
         logger.info(f"Deleted paths containing '{schema_name}' from S3")
 
+    def _delete_folder_contents(self, folder):
+        logger.info(f"Deleting contents of folder: {folder}")
+
+        paginator = self.s3_client.get_paginator("list_objects_v2")
+        pages = paginator.paginate(Bucket=self.bucket_name, Prefix=folder)
+
+        delete_us = dict(Objects=[])
+        for page in pages:
+            if "Contents" in page:
+                for obj in page["Contents"]:
+                    delete_us["Objects"].append(dict(Key=obj["Key"]))
+
+                    # AWS limits to deleting 1000 objects at a time
+                    if len(delete_us["Objects"]) >= 1000:
+                        self.s3_client.delete_objects(Bucket=self.bucket_name, Delete=delete_us)
+                        delete_us = dict(Objects=[])
+
+        if len(delete_us["Objects"]):
+            self.s3_client.delete_objects(Bucket=self.bucket_name, Delete=delete_us)
+
+        logger.info(f"Deleted contents of folder: {folder}")
+
 
 class PeerDBInterface:
     def __init__(self, config):
@@ -282,11 +308,11 @@ class PeerDBInterface:
         if not self.docker_host_mapping or not host:
             return host
 
-        if host in ["localhost", "127.0.0.1"]:
+        if host in ['localhost', '127.0.0.1']:
             logger.debug(f"Mapping host {host} to {self.docker_host_mapping} for Docker environment")
             return self.docker_host_mapping
 
-        url_pattern = r"(localhost|127\.0\.0\.1)"
+        url_pattern = r'(localhost|127\.0\.0\.1)'
         match = re.search(url_pattern, host)
         if match:
             original_host = match.group(1)

{datamarket-0.9.4.dist-info → datamarket-0.9.5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: datamarket
-Version: 0.9.4
+Version: 0.9.5
 Summary: Utilities that integrate advanced scraping knowledge into just one library.
 License: GPL-3.0-or-later
 Author: DataMarket
@@ -63,7 +63,7 @@ Provides-Extra: xmltodict
 Requires-Dist: SQLAlchemy (>=2.0.0,<3.0.0) ; extra == "alchemy"
 Requires-Dist: azure-storage-blob (>=12.0.0,<13.0.0) ; extra == "azure-storage-blob"
 Requires-Dist: beautifulsoup4 (>=4.0.0,<5.0.0)
-Requires-Dist: boto3 (>=1.0.0,<2.0.0) ; extra == "boto3" or extra == "aws" or extra == "peerdb"
+Requires-Dist: boto3 (>=1.35.0,<1.36.0) ; extra == "boto3" or extra == "aws" or extra == "peerdb"
 Requires-Dist: chompjs (>=1.0.0,<2.0.0) ; extra == "chompjs"
 Requires-Dist: click (>=8.0.0,<9.0.0) ; extra == "click"
 Requires-Dist: clickhouse-driver (>=0.2.0,<0.3.0) ; extra == "clickhouse-driver" or extra == "peerdb"

{datamarket-0.9.4.dist-info → datamarket-0.9.5.dist-info}/RECORD

@@ -3,9 +3,9 @@ datamarket/interfaces/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3h
 datamarket/interfaces/alchemy.py,sha256=V8E1GtokxUNmrUftKTFkIpNoXaqJME7ACES2BY0znQM,4214
 datamarket/interfaces/aws.py,sha256=R6lYdSCD6a4g9l6aFMtNDt_EX3kroe2untDhgy7XG1k,2384
 datamarket/interfaces/drive.py,sha256=shbV5jpQVe_KPE-8Idx6Z9te5Zu1SmVfrvSAyd9ZIgE,2915
-datamarket/interfaces/ftp.py,sha256=Owk3D7tiF47_ZFT3Dc9h4_BaPsWtcJUbhagjpQB19q8,1900
+datamarket/interfaces/ftp.py,sha256=VZSxISKquMIVbt-Nvb1HgOB9pwkzYunoror-anZNiiQ,1881
 datamarket/interfaces/nominatim.py,sha256=_gFJ04D-ju5xn3wuaGT5Pj5jhf4F5eINpxOpuQL_dIQ,3664
-datamarket/interfaces/peerdb.py,sha256=FhBLJfR2EMT9Rsnj_OJXvC14E5OlXGsMrPUQ1AQlwPY,20717
+datamarket/interfaces/peerdb.py,sha256=rNQ1-THcVvrej8BEPJs9zM4VfH5dlByafOIHYN9sB2A,21833
 datamarket/interfaces/proxy.py,sha256=updoOStKd8-nQBbxWbnD9eOt6HksnYi-5dQ0rEySf5M,3152
 datamarket/interfaces/tinybird.py,sha256=AYrcRGNOCoCt7ojilkWa27POROee9sTCwZ61GGHEPeM,2698
 datamarket/params/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -17,7 +17,7 @@ datamarket/utils/main.py,sha256=0Abt3ww1VSPnX4AVKDcYzqDLAOEV_54iUHMLJfre2bg,6129
 datamarket/utils/selenium.py,sha256=IMKlbLzXABFhACnWzhHmB0l2hhVzNwHGZwbo14nEewQ,2499
 datamarket/utils/soda.py,sha256=eZTXFbI1P3WoMd1MM-YjoVTpdjTcDSWuvBb7ViBMhSQ,941
 datamarket/utils/typer.py,sha256=FDF3l6gh3UlAFPsHCtesnekvct2rKz0oFn3uKARBQvE,814
-datamarket-0.9.4.dist-info/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
-datamarket-0.9.4.dist-info/METADATA,sha256=YYuxN--M3y9MX62_hG5Y1piS2TBBo_fl6MfJeLyyOZA,6360
-datamarket-0.9.4.dist-info/WHEEL,sha256=XbeZDeTWKc1w7CSIyre5aMDU_-PohRwTQceYnisIYYY,88
-datamarket-0.9.4.dist-info/RECORD,,
+datamarket-0.9.5.dist-info/LICENSE,sha256=OXLcl0T2SZ8Pmy2_dmlvKuetivmyPd5m1q-Gyd-zaYY,35149
+datamarket-0.9.5.dist-info/METADATA,sha256=0OFyrz2YcKfH1HachMVzhD7C_kL9iODaxEFLTB4e2NI,6362
+datamarket-0.9.5.dist-info/WHEEL,sha256=XbeZDeTWKc1w7CSIyre5aMDU_-PohRwTQceYnisIYYY,88
+datamarket-0.9.5.dist-info/RECORD,,