polly-python 1.4.0__tar.gz → 2.0.0__tar.gz

{polly_python-1.4.0/polly_python.egg-info → polly_python-2.0.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: polly_python
-Version: 1.4.0
+Version: 2.0.0
 Summary: Polly SDK
 Home-page: https://github.com/ElucidataInc/polly-python
 Project-URL: Documentation, https://docs.elucidata.io
@@ -17,6 +17,7 @@ Requires-Dist: mixpanel==4.10.0
 Requires-Dist: Deprecated>=1.2.12
 Requires-Dist: pytest>=6.2.5
 Requires-Dist: cryptography<=38.0.0,>=37.0.1
+Requires-Dist: numpy==1.26.4
 Requires-Dist: plotly<5.0.0,>=4.8.1; python_version > "3.6" and python_version < "3.7"
 Requires-Dist: plotly>=5.0.0; python_version >= "3.7"
 Requires-Dist: pandas<1.2.0,>=1.1.0; python_version > "3.6" and python_version < "3.7"

polly_python-2.0.0/polly/__init__.py

@@ -0,0 +1 @@
+__version__ = "2.0.0"

{polly_python-1.4.0 → polly_python-2.0.0}/polly/helpers.py

@@ -29,6 +29,7 @@ import pandas as pd
 import polly.http_response_codes as http_codes
 from polly.tracking import Track
 import polly.constants as const
+import string
 
 
 def get_platform_value_from_env(
@@ -54,7 +55,9 @@ def make_path(prefix: any, postfix: any) -> str:
     if not prefix:
         raise InvalidParameterException("prefix")
     if not postfix:
-        raise InvalidParameterException("postfix")
+        raise InvalidParameterException(
+            'path can\'t be empty. if u want to push to root then make path as "/"'
+        )
     return os.path.normpath(f"{prefix}/{postfix}")
 
 
@@ -149,22 +152,60 @@ def display_df_from_list(val_list: list, column_name_in_df: str):
 
 def upload_to_S3(cloud_path: str, local_path: str, credentials: dict) -> None:
     """
-    Function to upload file/folder to S3 cloud path
+    Uploads a file or folder to a specified S3 cloud path.
     """
     access_key_id = credentials["AccessKeyId"]
     secret_access_key = credentials["SecretAccessKey"]
     session_token = credentials["SessionToken"]
+
+    # use these extras for all CloudPaths
     client = S3Client(
         aws_access_key_id=access_key_id,
         aws_secret_access_key=secret_access_key,
         aws_session_token=session_token,
         extra_args={"ContentType": "text/html"},
     )
-    # use these extras for all CloudPaths
+
+    # Behavior:
+    # 1. Validation of cloud_path:
+    #    - If any part of the cloud_path contains directory names starting with special characters,
+    #      the function will not consider them as valid unless they already exist in the S3 bucket.
+    #    - The function will iterate through the cloud_path, checking each segment. If it encounters
+    #      a segment starting with a special character that does not already exist in the S3 bucket,
+    #      it will raise an error.
+
+    # Example:
+    # - Given a cloud_path of 'a/b/c/#/d/e':
+    #   1. if path not exists already, The function first checks if the path '{s3}/a/b/c/#/d' exists and that 'e'
+    #      does not start with a special character.
+    #   2. If '{s3}/a/b/c/#/d' not exists, it checks '{s3}/a/b/c/#' and ensures 'd' does not start with a special character.
+    #   3. This process continues until a exist path is found or the directory start with special character.
+    # - If the cloud_path is invalid, the function raises an exception.
+    # - Once a valid path is determined, the file or folder is uploaded to this path in the S3 bucket.
     client.set_as_default_client()
     source_path = client.CloudPath(cloud_path)
+    splt_char = os.path.sep
+    punctuation_chars = set(string.punctuation)
     if not source_path.exists():
+        # if path is S3://UserDataBucket/12345/path
+        # then path split will be [S3, , UserDataBucket, 12345, path]
+        path_split = cloud_path.split("/")
+        # s3 path have length = 4, so ignore that and checking path after that
+        while len(path_split) > 4:
+            cloud_path = splt_char.join(path_split[: len(path_split) - 1])
+            if (
+                path_split[len(path_split) - 1]
+                and path_split[len(path_split) - 1][0] in punctuation_chars
+            ):
+                raise InvalidParameterException(
+                    f"path can't start with {path_split[len(path_split)-1][0]}"
+                )
+            new_source_path = client.CloudPath(cloud_path)
+            if new_source_path.exists():
+                break
+            path_split = cloud_path.split("/")
         source_path.mkdir()
+
     try:
         source_path.upload_from(local_path, force_overwrite_to_cloud=True)
     except ClientError as e:
@@ -172,7 +213,11 @@ def upload_to_S3(cloud_path: str, local_path: str, credentials: dict) -> None:
 
 
 def download_from_S3(
-    cloud_path: str, workspace_path: str, credentials: dict, destination_path: str
+    cloud_path: str,
+    workspace_path: str,
+    credentials: dict,
+    destination_path: str,
+    copy_workspace_path: bool,
 ) -> None:
     """
     Function to download file/folder from workspaces
@@ -195,13 +240,15 @@ def download_from_S3(
         except ClientError as e:
             raise OperationFailedException(e)
     else:
-        if not cloud_path.endswith("/"):
-            cloud_path += "/"
         source_path = client.CloudPath(cloud_path)
         if not source_path.is_dir():
             raise InvalidPathException
         try:
-            destination_path = f"{make_path(destination_path,workspace_path)}"
+            # If copy_workspace_path is True, append workspace_path to destination_path to copy the directory structure.
+            # ex- make_path('/home/user/project/', 'folder1/folder2') = '/home/user/project/folder1/folder2'
+            if copy_workspace_path is True:
+                destination_path = f"{make_path(destination_path,workspace_path)}"
+
             source_path.copytree(destination_path, force_overwrite_to_cloud=True)
         except ClientError as e:
             raise OperationFailedException(e)

{polly_python-1.4.0 → polly_python-2.0.0}/polly/pipelines.py

@@ -195,8 +195,9 @@ class Pipelines:
         """
         This function is used to create a Pipeline run.\n
         A run is a collection of jobs, this functions creates an empty run in which the jobs can be added.
+
         Args:
-            pipeline_id (str): pipeline_id for which the run is to be created.
+            pipeline_id (str): pipeline_id for which the run is to be created
             run_name (str): name of the run
             priority (str): priority of the run, can be low | medium | high
             tags (dict): a dict of key-value pair with tag_name -> tag_value mapping

{polly_python-1.4.0 → polly_python-2.0.0}/polly/workspaces.py

@@ -270,6 +270,7 @@ initialize a object that can use all function and methods of Workspaces class.
         if not isExists:
             raise InvalidPathException
         # check for access rights for the workspace_id
+        workspace_path = workspace_path.strip()
         access_workspace = helpers.workspaces_permission_check(self, workspace_id)
         if not access_workspace:
             raise AccessDeniedError(
@@ -284,7 +285,11 @@ initialize a object that can use all function and methods of Workspaces class.
 
     @Track.track_decorator
     def download_from_workspaces(
-        self, workspace_id: int, workspace_path: str, local_path: str
+        self,
+        workspace_id: int,
+        workspace_path: str,
+        local_path: str,
+        copy_workspace_path: bool = True,
     ) -> None:
         """
         Function to download files/folders from workspaces.
@@ -292,6 +297,7 @@ initialize a object that can use all function and methods of Workspaces class.
         Args:
               workspace_id (int) : Id of the workspace where file needs to uploaded
               workspace_path (str) : Downloaded file on workspace. The workspace path should be prefixed with "polly://"
+              copy_workspace_path (bool) : Flag indicating whether the workspace path needs to copied in the working directory
         Returns:
               None
         Raises:
@@ -315,7 +321,9 @@ initialize a object that can use all function and methods of Workspaces class.
         if workspace_path.startswith("polly://"):
             workspace_path = workspace_path.split("polly://")[1]
         s3_path, credentials = self._s3_util(workspace_id, workspace_path)
-        helpers.download_from_S3(s3_path, workspace_path, credentials, local_path)
+        helpers.download_from_S3(
+            s3_path, workspace_path, credentials, local_path, copy_workspace_path
+        )
         logging.basicConfig(level=logging.INFO)
         logging.info(f"Download successful to path={local_path}")
 

{polly_python-1.4.0 → polly_python-2.0.0/polly_python.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: polly_python
-Version: 1.4.0
+Version: 2.0.0
 Summary: Polly SDK
 Home-page: https://github.com/ElucidataInc/polly-python
 Project-URL: Documentation, https://docs.elucidata.io
@@ -17,6 +17,7 @@ Requires-Dist: mixpanel==4.10.0
 Requires-Dist: Deprecated>=1.2.12
 Requires-Dist: pytest>=6.2.5
 Requires-Dist: cryptography<=38.0.0,>=37.0.1
+Requires-Dist: numpy==1.26.4
 Requires-Dist: plotly<5.0.0,>=4.8.1; python_version > "3.6" and python_version < "3.7"
 Requires-Dist: plotly>=5.0.0; python_version >= "3.7"
 Requires-Dist: pandas<1.2.0,>=1.1.0; python_version > "3.6" and python_version < "3.7"

{polly_python-1.4.0 → polly_python-2.0.0}/polly_python.egg-info/requires.txt

@@ -7,6 +7,7 @@ mixpanel==4.10.0
 Deprecated>=1.2.12
 pytest>=6.2.5
 cryptography<=38.0.0,>=37.0.1
+numpy==1.26.4
 requests==2.25.1
 
 [:python_version > "3.6" and python_version < "3.7"]

{polly_python-1.4.0 → polly_python-2.0.0}/setup.cfg

@@ -24,6 +24,7 @@ install_requires =
 	Deprecated >= 1.2.12
 	pytest >= 6.2.5
 	cryptography >= 37.0.1, <= 38.0.0
+	numpy == 1.26.4
 	plotly <5.0.0, >=4.8.1 ; python_version > '3.6' and python_version < '3.7'
 	plotly >= 5.0.0 ; python_version >= '3.7'
 	pandas >= 1.1.0, < 1.2.0 ; python_version > '3.6' and python_version < '3.7'

{polly_python-1.4.0 → polly_python-2.0.0}/tests/test_workspaces.py

@@ -194,6 +194,31 @@ def test_download_from_workspaces(mocker, mock_workspaces_permission_check_fixtu
             valid_workspace_id, invalid_workspace_path, local_path
         )
 
+    exclude_path_result = obj.download_from_workspaces(
+        valid_workspace_id, valid_workspace_path, local_path, copy_workspace_path=False
+    )
+    assert exclude_path_result is None
+    with pytest.raises(
+        InvalidParameterException,
+        match=r".* Invalid Parameters .*",
+    ):
+        obj.download_from_workspaces(
+            invalid_workspace_id,
+            valid_workspace_path,
+            local_path,
+            copy_workspace_path=False,
+        )
+    with pytest.raises(
+        InvalidParameterException,
+        match=r".* Invalid Parameters .*",
+    ):
+        obj.download_from_workspaces(
+            valid_workspace_id,
+            invalid_workspace_path,
+            local_path,
+            copy_workspace_path=False,
+        )
+
 
 def test_sync_data(mocker, mock_workspaces_permission_check_fixture):
     obj = workspaces.Workspaces(token)

polly_python-1.4.0/polly/__init__.py

@@ -1 +0,0 @@
-__version__ = "1.4.0"