dataproc-spark-connect 0.2.1__tar.gz → 0.6.0__tar.gz

{dataproc_spark_connect-0.2.1 → dataproc_spark_connect-0.6.0}/PKG-INFO

@@ -1,42 +1,34 @@
 Metadata-Version: 2.1
 Name: dataproc-spark-connect
-Version: 0.2.1
+Version: 0.6.0
 Summary: Dataproc client library for Spark Connect
 Home-page: https://github.com/GoogleCloudDataproc/dataproc-spark-connect-python
 Author: Google LLC
 License: Apache 2.0
 License-File: LICENSE
 Requires-Dist: google-api-core>=2.19.1
-Requires-Dist: google-cloud-dataproc>=5.15.1
-Requires-Dist: wheel
+Requires-Dist: google-cloud-dataproc>=5.18.0
 Requires-Dist: websockets
-Requires-Dist: pyspark>=3.5
-Requires-Dist: pandas
-Requires-Dist: pyarrow
+Requires-Dist: pyspark[connect]>=3.5
+Requires-Dist: packaging>=20.0
 
 # Dataproc Spark Connect Client
 
-> ⚠️ **Warning:**  
-The package `dataproc-spark-connect` has been renamed to `google-spark-connect`. `dataproc-spark-connect` will no longer be updated.
-For help using `google-spark-connect`, please see [guide](https://github.com/GoogleCloudDataproc/dataproc-spark-connect-python/blob/main/README.md). 
-
-
 A wrapper of the Apache [Spark Connect](https://spark.apache.org/spark-connect/) client with
 additional functionalities that allow applications to communicate with a remote Dataproc
 Spark cluster using the Spark Connect protocol without requiring additional steps.
 
 ## Install
 
-.. code-block:: console
-
-      pip install dataproc_spark_connect
+```console
+pip install dataproc_spark_connect
+```
 
 ## Uninstall
 
-.. code-block:: console
-
-      pip uninstall dataproc_spark_connect
-
+```console
+pip uninstall dataproc_spark_connect
+```
 
 ## Setup
 This client requires permissions to manage [Dataproc sessions and session templates](https://cloud.google.com/dataproc-serverless/docs/concepts/iam).
@@ -51,36 +43,36 @@ If you are running the client outside of Google Cloud, you must set following en
 
 1. Install the latest version of Dataproc Python client and Dataproc Spark Connect modules:
 
-      .. code-block:: console
-
-            pip install google_cloud_dataproc --force-reinstall
-            pip install dataproc_spark_connect --force-reinstall
+   ```console
+   pip install google_cloud_dataproc --force-reinstall
+   pip install dataproc_spark_connect --force-reinstall
+   ```
 
 2. Add the required import into your PySpark application or notebook:
 
-      .. code-block:: python
-
-            from google.cloud.dataproc_spark_connect import DataprocSparkSession
+   ```python
+   from google.cloud.dataproc_spark_connect import DataprocSparkSession
+   ```
 
 3. There are two ways to create a spark session,
 
    1. Start a Spark session using properties defined in `DATAPROC_SPARK_CONNECT_SESSION_DEFAULT_CONFIG`:
 
-      .. code-block:: python
-
-            spark = DataprocSparkSession.builder.getOrCreate()
+      ```python
+      spark = DataprocSparkSession.builder.getOrCreate()
+      ```
 
    2. Start a Spark session with the following code instead of using a config file:
 
-      .. code-block:: python
-
-            from google.cloud.dataproc_v1 import SparkConnectConfig
-            from google.cloud.dataproc_v1 import Session
-            dataproc_config = Session()
-            dataproc_config.spark_connect_session = SparkConnectConfig()
-            dataproc_config.environment_config.execution_config.subnetwork_uri = "<subnet>"
-            dataproc_config.runtime_config.version = '3.0'
-            spark = DataprocSparkSession.builder.dataprocConfig(dataproc_config).getOrCreate()
+      ```python
+      from google.cloud.dataproc_v1 import SparkConnectConfig
+      from google.cloud.dataproc_v1 import Session
+      dataproc_session_config = Session()
+      dataproc_session_config.spark_connect_session = SparkConnectConfig()
+      dataproc_session_config.environment_config.execution_config.subnetwork_uri = "<subnet>"
+      dataproc_session_config.runtime_config.version = '3.0'
+      spark = DataprocSparkSession.builder.dataprocSessionConfig(dataproc_session_config).getOrCreate()
+      ```
 
 ## Billing
 As this client runs the spark workload on Dataproc, your project will be billed as per [Dataproc Serverless Pricing](https://cloud.google.com/dataproc-serverless/pricing).
@@ -91,29 +83,29 @@ This will happen even if you are running the client from a non-GCE instance.
 
 1. Install the requirements in virtual environment.
 
-      .. code-block:: console
-
-            pip install -r requirements.txt
+      ```console
+      pip install -r requirements-dev.txt
+      ```
 
 2. Build the code.
 
-      .. code-block:: console
-
-            python setup.py sdist bdist_wheel
-
+      ```console
+      python setup.py sdist bdist_wheel
+      ```
 
 3. Copy the generated `.whl` file to Cloud Storage. Use the version specified in the `setup.py` file.
 
-      .. code-block:: console
-
-            VERSION=<version> gsutil cp dist/dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl gs://<your_bucket_name>
+      ```sh
+      VERSION=<version>
+      gsutil cp dist/dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl gs://<your_bucket_name>
+      ```
 
 4. Download the new SDK on Vertex, then uninstall the old version and install the new one.
 
-      .. code-block:: console
-
-            %%bash
-            export VERSION=<version>
-            gsutil cp gs://<your_bucket_name>/dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl .
-            yes | pip uninstall dataproc_spark_connect
-            pip install dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl
+      ```sh
+      %%bash
+      export VERSION=<version>
+      gsutil cp gs://<your_bucket_name>/dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl .
+      yes | pip uninstall dataproc_spark_connect
+      pip install dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl
+      ```

{dataproc_spark_connect-0.2.1 → dataproc_spark_connect-0.6.0}/README.md

@@ -1,26 +1,20 @@
 # Dataproc Spark Connect Client
 
-> ⚠️ **Warning:**  
-The package `dataproc-spark-connect` has been renamed to `google-spark-connect`. `dataproc-spark-connect` will no longer be updated.
-For help using `google-spark-connect`, please see [guide](https://github.com/GoogleCloudDataproc/dataproc-spark-connect-python/blob/main/README.md). 
-
-
 A wrapper of the Apache [Spark Connect](https://spark.apache.org/spark-connect/) client with
 additional functionalities that allow applications to communicate with a remote Dataproc
 Spark cluster using the Spark Connect protocol without requiring additional steps.
 
 ## Install
 
-.. code-block:: console
-
-      pip install dataproc_spark_connect
+```console
+pip install dataproc_spark_connect
+```
 
 ## Uninstall
 
-.. code-block:: console
-
-      pip uninstall dataproc_spark_connect
-
+```console
+pip uninstall dataproc_spark_connect
+```
 
 ## Setup
 This client requires permissions to manage [Dataproc sessions and session templates](https://cloud.google.com/dataproc-serverless/docs/concepts/iam).
@@ -35,36 +29,36 @@ If you are running the client outside of Google Cloud, you must set following en
 
 1. Install the latest version of Dataproc Python client and Dataproc Spark Connect modules:
 
-      .. code-block:: console
-
-            pip install google_cloud_dataproc --force-reinstall
-            pip install dataproc_spark_connect --force-reinstall
+   ```console
+   pip install google_cloud_dataproc --force-reinstall
+   pip install dataproc_spark_connect --force-reinstall
+   ```
 
 2. Add the required import into your PySpark application or notebook:
 
-      .. code-block:: python
-
-            from google.cloud.dataproc_spark_connect import DataprocSparkSession
+   ```python
+   from google.cloud.dataproc_spark_connect import DataprocSparkSession
+   ```
 
 3. There are two ways to create a spark session,
 
    1. Start a Spark session using properties defined in `DATAPROC_SPARK_CONNECT_SESSION_DEFAULT_CONFIG`:
 
-      .. code-block:: python
-
-            spark = DataprocSparkSession.builder.getOrCreate()
+      ```python
+      spark = DataprocSparkSession.builder.getOrCreate()
+      ```
 
    2. Start a Spark session with the following code instead of using a config file:
 
-      .. code-block:: python
-
-            from google.cloud.dataproc_v1 import SparkConnectConfig
-            from google.cloud.dataproc_v1 import Session
-            dataproc_config = Session()
-            dataproc_config.spark_connect_session = SparkConnectConfig()
-            dataproc_config.environment_config.execution_config.subnetwork_uri = "<subnet>"
-            dataproc_config.runtime_config.version = '3.0'
-            spark = DataprocSparkSession.builder.dataprocConfig(dataproc_config).getOrCreate()
+      ```python
+      from google.cloud.dataproc_v1 import SparkConnectConfig
+      from google.cloud.dataproc_v1 import Session
+      dataproc_session_config = Session()
+      dataproc_session_config.spark_connect_session = SparkConnectConfig()
+      dataproc_session_config.environment_config.execution_config.subnetwork_uri = "<subnet>"
+      dataproc_session_config.runtime_config.version = '3.0'
+      spark = DataprocSparkSession.builder.dataprocSessionConfig(dataproc_session_config).getOrCreate()
+      ```
 
 ## Billing
 As this client runs the spark workload on Dataproc, your project will be billed as per [Dataproc Serverless Pricing](https://cloud.google.com/dataproc-serverless/pricing).
@@ -75,29 +69,29 @@ This will happen even if you are running the client from a non-GCE instance.
 
 1. Install the requirements in virtual environment.
 
-      .. code-block:: console
-
-            pip install -r requirements.txt
+      ```console
+      pip install -r requirements-dev.txt
+      ```
 
 2. Build the code.
 
-      .. code-block:: console
-
-            python setup.py sdist bdist_wheel
-
+      ```console
+      python setup.py sdist bdist_wheel
+      ```
 
 3. Copy the generated `.whl` file to Cloud Storage. Use the version specified in the `setup.py` file.
 
-      .. code-block:: console
-
-            VERSION=<version> gsutil cp dist/dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl gs://<your_bucket_name>
+      ```sh
+      VERSION=<version>
+      gsutil cp dist/dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl gs://<your_bucket_name>
+      ```
 
 4. Download the new SDK on Vertex, then uninstall the old version and install the new one.
 
-      .. code-block:: console
-
-            %%bash
-            export VERSION=<version>
-            gsutil cp gs://<your_bucket_name>/dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl .
-            yes | pip uninstall dataproc_spark_connect
-            pip install dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl
+      ```sh
+      %%bash
+      export VERSION=<version>
+      gsutil cp gs://<your_bucket_name>/dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl .
+      yes | pip uninstall dataproc_spark_connect
+      pip install dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl
+      ```

{dataproc_spark_connect-0.2.1 → dataproc_spark_connect-0.6.0}/dataproc_spark_connect.egg-info/PKG-INFO

@@ -1,42 +1,34 @@
 Metadata-Version: 2.1
 Name: dataproc-spark-connect
-Version: 0.2.1
+Version: 0.6.0
 Summary: Dataproc client library for Spark Connect
 Home-page: https://github.com/GoogleCloudDataproc/dataproc-spark-connect-python
 Author: Google LLC
 License: Apache 2.0
 License-File: LICENSE
 Requires-Dist: google-api-core>=2.19.1
-Requires-Dist: google-cloud-dataproc>=5.15.1
-Requires-Dist: wheel
+Requires-Dist: google-cloud-dataproc>=5.18.0
 Requires-Dist: websockets
-Requires-Dist: pyspark>=3.5
-Requires-Dist: pandas
-Requires-Dist: pyarrow
+Requires-Dist: pyspark[connect]>=3.5
+Requires-Dist: packaging>=20.0
 
 # Dataproc Spark Connect Client
 
-> ⚠️ **Warning:**  
-The package `dataproc-spark-connect` has been renamed to `google-spark-connect`. `dataproc-spark-connect` will no longer be updated.
-For help using `google-spark-connect`, please see [guide](https://github.com/GoogleCloudDataproc/dataproc-spark-connect-python/blob/main/README.md). 
-
-
 A wrapper of the Apache [Spark Connect](https://spark.apache.org/spark-connect/) client with
 additional functionalities that allow applications to communicate with a remote Dataproc
 Spark cluster using the Spark Connect protocol without requiring additional steps.
 
 ## Install
 
-.. code-block:: console
-
-      pip install dataproc_spark_connect
+```console
+pip install dataproc_spark_connect
+```
 
 ## Uninstall
 
-.. code-block:: console
-
-      pip uninstall dataproc_spark_connect
-
+```console
+pip uninstall dataproc_spark_connect
+```
 
 ## Setup
 This client requires permissions to manage [Dataproc sessions and session templates](https://cloud.google.com/dataproc-serverless/docs/concepts/iam).
@@ -51,36 +43,36 @@ If you are running the client outside of Google Cloud, you must set following en
 
 1. Install the latest version of Dataproc Python client and Dataproc Spark Connect modules:
 
-      .. code-block:: console
-
-            pip install google_cloud_dataproc --force-reinstall
-            pip install dataproc_spark_connect --force-reinstall
+   ```console
+   pip install google_cloud_dataproc --force-reinstall
+   pip install dataproc_spark_connect --force-reinstall
+   ```
 
 2. Add the required import into your PySpark application or notebook:
 
-      .. code-block:: python
-
-            from google.cloud.dataproc_spark_connect import DataprocSparkSession
+   ```python
+   from google.cloud.dataproc_spark_connect import DataprocSparkSession
+   ```
 
 3. There are two ways to create a spark session,
 
    1. Start a Spark session using properties defined in `DATAPROC_SPARK_CONNECT_SESSION_DEFAULT_CONFIG`:
 
-      .. code-block:: python
-
-            spark = DataprocSparkSession.builder.getOrCreate()
+      ```python
+      spark = DataprocSparkSession.builder.getOrCreate()
+      ```
 
    2. Start a Spark session with the following code instead of using a config file:
 
-      .. code-block:: python
-
-            from google.cloud.dataproc_v1 import SparkConnectConfig
-            from google.cloud.dataproc_v1 import Session
-            dataproc_config = Session()
-            dataproc_config.spark_connect_session = SparkConnectConfig()
-            dataproc_config.environment_config.execution_config.subnetwork_uri = "<subnet>"
-            dataproc_config.runtime_config.version = '3.0'
-            spark = DataprocSparkSession.builder.dataprocConfig(dataproc_config).getOrCreate()
+      ```python
+      from google.cloud.dataproc_v1 import SparkConnectConfig
+      from google.cloud.dataproc_v1 import Session
+      dataproc_session_config = Session()
+      dataproc_session_config.spark_connect_session = SparkConnectConfig()
+      dataproc_session_config.environment_config.execution_config.subnetwork_uri = "<subnet>"
+      dataproc_session_config.runtime_config.version = '3.0'
+      spark = DataprocSparkSession.builder.dataprocSessionConfig(dataproc_session_config).getOrCreate()
+      ```
 
 ## Billing
 As this client runs the spark workload on Dataproc, your project will be billed as per [Dataproc Serverless Pricing](https://cloud.google.com/dataproc-serverless/pricing).
@@ -91,29 +83,29 @@ This will happen even if you are running the client from a non-GCE instance.
 
 1. Install the requirements in virtual environment.
 
-      .. code-block:: console
-
-            pip install -r requirements.txt
+      ```console
+      pip install -r requirements-dev.txt
+      ```
 
 2. Build the code.
 
-      .. code-block:: console
-
-            python setup.py sdist bdist_wheel
-
+      ```console
+      python setup.py sdist bdist_wheel
+      ```
 
 3. Copy the generated `.whl` file to Cloud Storage. Use the version specified in the `setup.py` file.
 
-      .. code-block:: console
-
-            VERSION=<version> gsutil cp dist/dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl gs://<your_bucket_name>
+      ```sh
+      VERSION=<version>
+      gsutil cp dist/dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl gs://<your_bucket_name>
+      ```
 
 4. Download the new SDK on Vertex, then uninstall the old version and install the new one.
 
-      .. code-block:: console
-
-            %%bash
-            export VERSION=<version>
-            gsutil cp gs://<your_bucket_name>/dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl .
-            yes | pip uninstall dataproc_spark_connect
-            pip install dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl
+      ```sh
+      %%bash
+      export VERSION=<version>
+      gsutil cp gs://<your_bucket_name>/dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl .
+      yes | pip uninstall dataproc_spark_connect
+      pip install dataproc_spark_connect-${VERSION}-py2.py3-none-any.whl
+      ```

{dataproc_spark_connect-0.2.1 → dataproc_spark_connect-0.6.0}/dataproc_spark_connect.egg-info/SOURCES.txt

@@ -9,6 +9,8 @@ dataproc_spark_connect.egg-info/dependency_links.txt
 dataproc_spark_connect.egg-info/requires.txt
 dataproc_spark_connect.egg-info/top_level.txt
 google/cloud/dataproc_spark_connect/__init__.py
+google/cloud/dataproc_spark_connect/exceptions.py
+google/cloud/dataproc_spark_connect/pypi_artifacts.py
 google/cloud/dataproc_spark_connect/session.py
 google/cloud/dataproc_spark_connect/client/__init__.py
 google/cloud/dataproc_spark_connect/client/core.py

dataproc_spark_connect-0.6.0/dataproc_spark_connect.egg-info/requires.txt

@@ -0,0 +1,5 @@
+google-api-core>=2.19.1
+google-cloud-dataproc>=5.18.0
+websockets
+pyspark[connect]>=3.5
+packaging>=20.0

{dataproc_spark_connect-0.2.1 → dataproc_spark_connect-0.6.0}/google/cloud/dataproc_spark_connect/__init__.py

@@ -11,13 +11,19 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-from .session import DataprocSparkSession
+import importlib.metadata
 import warnings
 
-warnings.warn(
-    "The package 'dataproc-spark-connect' has been renamed to 'google-spark-connect'. "
-    "'dataproc-spark-connect' will no longer be updated. "
-    "For help using 'google-spark-connect', "
-    "see https://github.com/GoogleCloudDataproc/dataproc-spark-connect-python/blob/main/README.md. ",
-    DeprecationWarning,
-)
+from .session import DataprocSparkSession
+
+old_package_name = "google-spark-connect"
+current_package_name = "dataproc-spark-connect"
+try:
+    importlib.metadata.distribution(old_package_name)
+    warnings.warn(
+        f"Package '{old_package_name}' is already installed in your environment. "
+        f"This might cause conflicts with '{current_package_name}'. "
+        f"Consider uninstalling '{old_package_name}' and only install '{current_package_name}'."
+    )
+except:
+    pass

{dataproc_spark_connect-0.2.1 → dataproc_spark_connect-0.6.0}/google/cloud/dataproc_spark_connect/client/core.py

@@ -11,12 +11,16 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+import logging
+
 import google
 import grpc
 from pyspark.sql.connect.client import ChannelBuilder
 
 from . import proxy
 
+logger = logging.getLogger(__name__)
+
 
 class DataprocChannelBuilder(ChannelBuilder):
     """
@@ -36,6 +40,10 @@ class DataprocChannelBuilder(ChannelBuilder):
     True
     """
 
+    def __init__(self, url, is_active_callback=None):
+        self._is_active_callback = is_active_callback
+        super().__init__(url)
+
     def toChannel(self) -> grpc.Channel:
         """
         Applies the parameters of the connection string and creates a new
@@ -51,7 +59,7 @@ class DataprocChannelBuilder(ChannelBuilder):
         return self._proxied_channel()
 
     def _proxied_channel(self) -> grpc.Channel:
-        return ProxiedChannel(self.host)
+        return ProxiedChannel(self.host, self._is_active_callback)
 
     def _direct_channel(self) -> grpc.Channel:
         destination = f"{self.host}:{self.port}"
@@ -75,7 +83,8 @@ class DataprocChannelBuilder(ChannelBuilder):
 
 class ProxiedChannel(grpc.Channel):
 
-    def __init__(self, target_host):
+    def __init__(self, target_host, is_active_callback):
+        self._is_active_callback = is_active_callback
         self._proxy = proxy.DataprocSessionProxy(0, target_host)
         self._proxy.start()
         self._proxied_connect_url = f"sc://localhost:{self._proxy.port}"
@@ -94,20 +103,37 @@ class ProxiedChannel(grpc.Channel):
         self._proxy.stop()
         return ret
 
+    def _wrap_method(self, wrapped_method):
+        if self._is_active_callback is None:
+            return wrapped_method
+
+        def checked_method(*margs, **mkwargs):
+            if (
+                self._is_active_callback is not None
+                and not self._is_active_callback()
+            ):
+                logger.warning(f"Session is no longer active")
+                raise RuntimeError(
+                    "Session not active. Please create a new session"
+                )
+            return wrapped_method(*margs, **mkwargs)
+
+        return checked_method
+
     def stream_stream(self, *args, **kwargs):
-        return self._wrapped.stream_stream(*args, **kwargs)
+        return self._wrap_method(self._wrapped.stream_stream(*args, **kwargs))
 
     def stream_unary(self, *args, **kwargs):
-        return self._wrapped.stream_unary(*args, **kwargs)
+        return self._wrap_method(self._wrapped.stream_unary(*args, **kwargs))
 
     def subscribe(self, *args, **kwargs):
-        return self._wrapped.subscribe(*args, **kwargs)
+        return self._wrap_method(self._wrapped.subscribe(*args, **kwargs))
 
     def unary_stream(self, *args, **kwargs):
-        return self._wrapped.unary_stream(*args, **kwargs)
+        return self._wrap_method(self._wrapped.unary_stream(*args, **kwargs))
 
     def unary_unary(self, *args, **kwargs):
-        return self._wrapped.unary_unary(*args, **kwargs)
+        return self._wrap_method(self._wrapped.unary_unary(*args, **kwargs))
 
     def unsubscribe(self, *args, **kwargs):
-        return self._wrapped.unsubscribe(*args, **kwargs)
+        return self._wrap_method(self._wrapped.unsubscribe(*args, **kwargs))

{dataproc_spark_connect-0.2.1 → dataproc_spark_connect-0.6.0}/google/cloud/dataproc_spark_connect/client/proxy.py

@@ -81,6 +81,7 @@ def connect_tcp_bridge(hostname):
     return websocketclient.connect(
         f"wss://{hostname}/{path}",
         additional_headers={"Authorization": f"Bearer {creds.token}"},
+        open_timeout=30,
     )
 
 
@@ -101,8 +102,11 @@ def forward_bytes(name, from_sock, to_sock):
         try:
             bs = from_sock.recv(1024)
             if not bs:
+                to_sock.close()
                 return
-            while bs:
+            attempt = 0
+            while bs and (attempt < 10):
+                attempt += 1
                 try:
                     to_sock.send(bs)
                     bs = None
@@ -110,6 +114,8 @@ def forward_bytes(name, from_sock, to_sock):
                     # On timeouts during a send, we retry just the send
                     # to make sure we don't lose any bytes.
                     pass
+            if bs:
+                raise Exception(f"Failed to forward bytes for {name}")
         except TimeoutError:
             # On timeouts during a receive, we retry the entire flow.
             pass
@@ -163,6 +169,11 @@ def forward_connection(conn_number, conn, addr, target_host):
     with conn:
         with connect_tcp_bridge(target_host) as websocket_conn:
             backend_socket = bridged_socket(websocket_conn)
+            # Set a timeout on how long we will allow send/recv calls to block
+            #
+            # The code that reads and writes to this connection will retry
+            # on timeouts, so this is a safe change.
+            conn.settimeout(10)
             connect_sockets(conn_number, conn, backend_socket)
 
 
@@ -210,14 +221,6 @@ class DataprocSessionProxy(object):
             s.release()
             while not self._killed:
                 conn, addr = frontend_socket.accept()
-                # Set a timeout on how long we will allow send/recv calls to block
-                #
-                # The code that reads and writes to this connection will retry
-                # on timeouts, so this is a safe change.
-                #
-                # The chosen timeout is a very short one because it allows us
-                # to more quickly detect when a connection has been closed.
-                conn.settimeout(1)
                 logger.debug(f"Accepted a connection from {addr}...")
                 self._conn_number += 1
                 threading.Thread(

dataproc_spark_connect-0.6.0/google/cloud/dataproc_spark_connect/exceptions.py

@@ -0,0 +1,27 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+
+class DataprocSparkConnectException(Exception):
+    """A custom exception class to only print the error messages.
+    This would be used for exceptions where the stack trace
+    doesn't provide any additional information.h
+    """
+
+    def __init__(self, message):
+        self.message = message
+        super().__init__(message)
+
+    def _render_traceback_(self):
+        return self.message

dataproc_spark_connect-0.6.0/google/cloud/dataproc_spark_connect/pypi_artifacts.py

@@ -0,0 +1,48 @@
+import json
+import logging
+import os
+import tempfile
+
+from packaging.requirements import Requirement
+
+logger = logging.getLogger(__name__)
+
+
+class PyPiArtifacts:
+    """
+    This is a helper class to serialize the PYPI package installation request with a "magic" file name
+    that Spark Connect server understands
+    """
+
+    @staticmethod
+    def __try_parsing_package(packages: set[str]) -> list[Requirement]:
+        reqs = [Requirement(p) for p in packages]
+        if 0 in [len(req.specifier) for req in reqs]:
+            logger.info("It is recommended to pin the version of the package")
+        return reqs
+
+    def __init__(self, packages: set[str]):
+        self.requirements = PyPiArtifacts.__try_parsing_package(packages)
+
+    def write_packages_config(self, s8s_session_uuid: str) -> str:
+        """
+        Can't use the same file-name as Spark throws exception that file already exists
+        Keep the filename/format in sync with server
+        """
+        dependencies = {
+            "version": "0.5",
+            "packageType": "PYPI",
+            "packages": [str(req) for req in self.requirements],
+        }
+
+        file_path = os.path.join(
+            tempfile.gettempdir(),
+            s8s_session_uuid,
+            "add-artifacts-1729-" + self.__str__() + ".json",
+        )
+
+        os.makedirs(os.path.dirname(file_path), exist_ok=True)
+        with open(file_path, "w") as json_file:
+            json.dump(dependencies, json_file, indent=4)
+        logger.debug("Dumping dependencies request in file: " + file_path)
+        return file_path

{dataproc_spark_connect-0.2.1 → dataproc_spark_connect-0.6.0}/google/cloud/dataproc_spark_connect/session.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+import atexit
 import json
 import logging
 import os
@@ -24,9 +25,10 @@ from typing import Any, cast, ClassVar, Dict, Optional
 from google.api_core import retry
 from google.api_core.future.polling import POLLING_PREDICATE
 from google.api_core.client_options import ClientOptions
-from google.api_core.exceptions import FailedPrecondition, InvalidArgument, NotFound
+from google.api_core.exceptions import Aborted, FailedPrecondition, InvalidArgument, NotFound, PermissionDenied
 from google.cloud.dataproc_v1.types import sessions
 
+from google.cloud.dataproc_spark_connect.pypi_artifacts import PyPiArtifacts
 from google.cloud.dataproc_spark_connect.client import DataprocChannelBuilder
 from google.cloud.dataproc_v1 import (
     CreateSessionRequest,
@@ -41,6 +43,7 @@ from google.protobuf.text_format import ParseError
 from pyspark.sql.connect.session import SparkSession
 from pyspark.sql.utils import to_str
 
+from google.cloud.dataproc_spark_connect.exceptions import DataprocSparkConnectException
 
 # Set up logging
 logging.basicConfig(level=logging.INFO)
@@ -61,7 +64,7 @@ class DataprocSparkSession(SparkSession):
     >>> spark = (
     ...     DataprocSparkSession.builder
     ...         .appName("Word Count")
-    ...         .dataprocConfig(Session())
+    ...         .dataprocSessionConfig(Session())
     ...         .getOrCreate()
     ... ) # doctest: +SKIP
     """
@@ -112,15 +115,15 @@ class DataprocSparkSession(SparkSession):
             self._project_id = project_id
             return self
 
-        def region(self, region):
-            self._region = region
+        def location(self, location):
+            self._region = location
             self._client_options.api_endpoint = os.environ.get(
                 "GOOGLE_CLOUD_DATAPROC_API_ENDPOINT",
                 f"{self._region}-dataproc.googleapis.com",
             )
             return self
 
-        def dataprocConfig(self, dataproc_config: Session):
+        def dataprocSessionConfig(self, dataproc_config: Session):
             with self._lock:
                 self._dataproc_config = dataproc_config
                 for k, v in dataproc_config.runtime_config.properties.items():
@@ -135,14 +138,14 @@ class DataprocSparkSession(SparkSession):
             else:
                 return self
 
-        def create(self) -> "SparkSession":
+        def create(self) -> "DataprocSparkSession":
             raise NotImplemented(
                 "DataprocSparkSession allows session creation only through getOrCreate"
             )
 
         def __create_spark_connect_session_from_s8s(
-            self, session_response
-        ) -> "SparkSession":
+            self, session_response, session_name
+        ) -> "DataprocSparkSession":
             DataprocSparkSession._active_s8s_session_uuid = (
                 session_response.uuid
             )
@@ -153,9 +156,16 @@ class DataprocSparkSession(SparkSession):
                 "Spark Connect Server"
             )
             spark_connect_url = spark_connect_url.replace("https", "sc")
+            if not spark_connect_url.endswith("/"):
+                spark_connect_url += "/"
             url = f"{spark_connect_url.replace('.com/', '.com:443/')};session_id={session_response.uuid};use_ssl=true"
             logger.debug(f"Spark Connect URL: {url}")
-            self._channel_builder = DataprocChannelBuilder(url)
+            self._channel_builder = DataprocChannelBuilder(
+                url,
+                is_active_callback=lambda: is_s8s_session_active(
+                    session_name, self._client_options
+                ),
+            )
 
             assert self._channel_builder is not None
             session = DataprocSparkSession(connection=self._channel_builder)
@@ -164,7 +174,7 @@ class DataprocSparkSession(SparkSession):
             self.__apply_options(session)
             return session
 
-        def __create(self) -> "SparkSession":
+        def __create(self) -> "DataprocSparkSession":
             with self._lock:
 
                 if self._options.get("spark.remote", False):
@@ -185,12 +195,8 @@ class DataprocSparkSession(SparkSession):
                     dataproc_config, session_template
                 )
 
-                spark = self._get_spark(dataproc_config, session_template)
-
                 if not spark_connect_session:
                     dataproc_config.spark_connect_session = {}
-                if not spark:
-                    dataproc_config.spark = {}
                 os.environ["SPARK_CONNECT_MODE_ENABLED"] = "1"
                 session_request = CreateSessionRequest()
                 session_id = self.generate_dataproc_session_id()
@@ -216,14 +222,28 @@ class DataprocSparkSession(SparkSession):
                         multiplier=1.0,
                         timeout=600,  # seconds
                     )
-                    logger.info(
-                        "Creating Spark session. It may take few minutes."
-                    )
+                    print("Creating Spark session. It may take a few minutes.")
+                    if (
+                        "dataproc_spark_connect_SESSION_TERMINATE_AT_EXIT"
+                        in os.environ
+                        and os.getenv(
+                            "dataproc_spark_connect_SESSION_TERMINATE_AT_EXIT"
+                        ).lower()
+                        == "true"
+                    ):
+                        atexit.register(
+                            lambda: terminate_s8s_session(
+                                self._project_id,
+                                self._region,
+                                session_id,
+                                self._client_options,
+                            )
+                        )
                     operation = SessionControllerClient(
                         client_options=self._client_options
                     ).create_session(session_request)
                     print(
-                        f"Interactive Session Detail View:  https://console.cloud.google.com/dataproc/interactive/{self._region}/{session_id}"
+                        f"Interactive Session Detail View:  https://console.cloud.google.com/dataproc/interactive/{self._region}/{session_id}?project={self._project_id}"
                     )
                     session_response: Session = operation.result(
                         polling=session_polling
@@ -249,49 +269,32 @@ class DataprocSparkSession(SparkSession):
                             logger.error(
                                 f"Exception while writing active session to file {file_path} , {e}"
                             )
-                except InvalidArgument as e:
+                except (InvalidArgument, PermissionDenied) as e:
                     DataprocSparkSession._active_s8s_session_id = None
-                    raise RuntimeError(
-                        f"Error while creating serverless session: {e}"
-                    ) from None
+                    raise DataprocSparkConnectException(
+                        f"Error while creating serverless session: {e.message}"
+                    )
                 except Exception as e:
                     DataprocSparkSession._active_s8s_session_id = None
                     raise RuntimeError(
-                        f"Error while creating serverless session https://console.cloud.google.com/dataproc/interactive/{self._region}/{session_id} : {e}"
-                    ) from None
+                        f"Error while creating serverless session"
+                    ) from e
 
                 logger.debug(
                     f"Serverless session created: {session_id}, creation time taken: {int(time.time() - s8s_creation_start_time)} seconds"
                 )
                 return self.__create_spark_connect_session_from_s8s(
-                    session_response
+                    session_response, dataproc_config.name
                 )
 
-        def _is_s8s_session_active(
-            self, s8s_session_id: str
-        ) -> Optional[sessions.Session]:
-            session_name = f"projects/{self._project_id}/locations/{self._region}/sessions/{s8s_session_id}"
-            get_session_request = GetSessionRequest()
-            get_session_request.name = session_name
-            state = None
-            try:
-                get_session_response = SessionControllerClient(
-                    client_options=self._client_options
-                ).get_session(get_session_request)
-                state = get_session_response.state
-            except Exception as e:
-                logger.debug(f"{s8s_session_id} deleted: {e}")
-                return None
-
-            if state is not None and (
-                state == Session.State.ACTIVE or state == Session.State.CREATING
-            ):
-                return get_session_response
-            return None
-
-        def _get_exiting_active_session(self) -> Optional["SparkSession"]:
+        def _get_exiting_active_session(
+            self,
+        ) -> Optional["DataprocSparkSession"]:
             s8s_session_id = DataprocSparkSession._active_s8s_session_id
-            session_response = self._is_s8s_session_active(s8s_session_id)
+            session_name = f"projects/{self._project_id}/locations/{self._region}/sessions/{s8s_session_id}"
+            session_response = get_active_s8s_session_response(
+                session_name, self._client_options
+            )
 
             session = DataprocSparkSession.getActiveSession()
             if session is None:
@@ -299,11 +302,11 @@ class DataprocSparkSession(SparkSession):
 
             if session_response is not None:
                 print(
-                    f"Using existing session: https://console.cloud.google.com/dataproc/interactive/{self._region}/{s8s_session_id}, configuration changes may not be applied."
+                    f"Using existing session: https://console.cloud.google.com/dataproc/interactive/{self._region}/{s8s_session_id}?project={self._project_id}, configuration changes may not be applied."
                 )
                 if session is None:
                     session = self.__create_spark_connect_session_from_s8s(
-                        session_response
+                        session_response, session_name
                     )
                 return session
             else:
@@ -315,7 +318,7 @@ class DataprocSparkSession(SparkSession):
 
                 return None
 
-        def getOrCreate(self) -> "SparkSession":
+        def getOrCreate(self) -> "DataprocSparkSession":
             with DataprocSparkSession._lock:
                 session = self._get_exiting_active_session()
                 if session is None:
@@ -413,11 +416,11 @@ class DataprocSparkSession(SparkSession):
             ]
             import importlib.metadata
 
-            dataproc_connect_version = importlib.metadata.version(
+            google_connect_version = importlib.metadata.version(
                 "dataproc-spark-connect"
             )
             client_version = importlib.metadata.version("pyspark")
-            version_message = f"Dataproc Spark Connect: {dataproc_connect_version} (PySpark: {client_version}) Dataproc Session Runtime: {version} (Spark: {server_version})"
+            version_message = f"Spark Connect: {google_connect_version} (PySpark: {client_version}) Session Runtime: {version} (Spark: {server_version})"
             logger.info(version_message)
             if trimmed_version(client_version) != trimmed_version(
                 server_version
@@ -435,14 +438,6 @@ class DataprocSparkSession(SparkSession):
                 spark_connect_session = session_template.spark_connect_session
             return spark_connect_session
 
-        def _get_spark(self, dataproc_config, session_template):
-            spark = None
-            if dataproc_config and dataproc_config.spark:
-                spark = dataproc_config.spark
-            elif session_template and session_template.spark:
-                spark = session_template.spark
-            return spark
-
         def generate_dataproc_session_id(self):
             timestamp = datetime.datetime.now().strftime("%Y%m%d-%H%M%S")
             suffix_length = 6
@@ -466,8 +461,8 @@ class DataprocSparkSession(SparkSession):
         <div>
             <p><b>Spark Connect</b></p>
 
-            <p><a href="{s8s_session}">Dataproc Session</a></p>
-            <p><a href="{ui}">Spark UI</a></p>
+            <p><a href="{s8s_session}?project={self._project_id}">Serverless Session</a></p>
+            <p><a href="{ui}?project={self._project_id}">Spark UI</a></p>
         </div>
         """
 
@@ -484,45 +479,70 @@ class DataprocSparkSession(SparkSession):
                     f"Exception while removing active session in file {file_path} , {e}"
                 )
 
+    def addArtifacts(
+        self,
+        *artifact: str,
+        pyfile: bool = False,
+        archive: bool = False,
+        file: bool = False,
+        pypi: bool = False,
+    ) -> None:
+        """
+        Add artifact(s) to the client session. Currently only local files & pypi installations are supported.
+
+        .. versionadded:: 3.5.0
+
+        Parameters
+        ----------
+        *path : tuple of str
+            Artifact's URIs to add.
+        pyfile : bool
+            Whether to add them as Python dependencies such as .py, .egg, .zip or .jar files.
+            The pyfiles are directly inserted into the path when executing Python functions
+            in executors.
+        archive : bool
+            Whether to add them as archives such as .zip, .jar, .tar.gz, .tgz, or .tar files.
+            The archives are unpacked on the executor side automatically.
+        file : bool
+            Add a file to be downloaded with this Spark job on every node.
+            The ``path`` passed can only be a local file for now.
+        pypi : bool
+            This option is only available with DataprocSparkSession. eg. `spark.addArtifacts("spacy==3.8.4", "torch",  pypi=True)`
+            Installs PyPi package (with its dependencies) in the active Spark session on the driver and executors.
+
+        Notes
+        -----
+        This is an API dedicated to Spark Connect client only. With regular Spark Session, it throws
+        an exception.
+        Regarding pypi: Popular packages are already pre-installed in s8s runtime.
+        https://cloud.google.com/dataproc-serverless/docs/concepts/versions/spark-runtime-2.2#python_libraries
+        If there are conflicts/package doesn't exist, it throws an exception.
+        """
+        if sum([pypi, file, pyfile, archive]) > 1:
+            raise ValueError(
+                "'pyfile', 'archive', 'file' and/or 'pypi' cannot be True together."
+            )
+        if pypi:
+            artifacts = PyPiArtifacts(set(artifact))
+            logger.debug("Making addArtifact call to install packages")
+            self.addArtifact(
+                artifacts.write_packages_config(self._active_s8s_session_uuid),
+                file=True,
+            )
+        else:
+            super().addArtifacts(
+                *artifact, pyfile=pyfile, archive=archive, file=file
+            )
+
     def stop(self) -> None:
         with DataprocSparkSession._lock:
             if DataprocSparkSession._active_s8s_session_id is not None:
-                from google.cloud.dataproc_v1 import SessionControllerClient
-
-                logger.debug(
-                    f"Terminating serverless session: {DataprocSparkSession._active_s8s_session_id}"
+                terminate_s8s_session(
+                    DataprocSparkSession._project_id,
+                    DataprocSparkSession._region,
+                    DataprocSparkSession._active_s8s_session_id,
+                    self._client_options,
                 )
-                terminate_session_request = TerminateSessionRequest()
-                session_name = f"projects/{DataprocSparkSession._project_id}/locations/{DataprocSparkSession._region}/sessions/{DataprocSparkSession._active_s8s_session_id}"
-                terminate_session_request.name = session_name
-                state = None
-                try:
-                    SessionControllerClient(
-                        client_options=self._client_options
-                    ).terminate_session(terminate_session_request)
-                    get_session_request = GetSessionRequest()
-                    get_session_request.name = session_name
-                    state = Session.State.ACTIVE
-                    while (
-                        state != Session.State.TERMINATING
-                        and state != Session.State.TERMINATED
-                        and state != Session.State.FAILED
-                    ):
-                        session = SessionControllerClient(
-                            client_options=self._client_options
-                        ).get_session(get_session_request)
-                        state = session.state
-                        sleep(1)
-                except NotFound:
-                    logger.debug(
-                        f"Session {DataprocSparkSession._active_s8s_session_id} already deleted"
-                    )
-                except FailedPrecondition:
-                    logger.debug(
-                        f"Session {DataprocSparkSession._active_s8s_session_id} already terminated manually or terminated automatically through session ttl limits"
-                    )
-                if state is not None and state == Session.State.FAILED:
-                    raise RuntimeError("Serverless session termination failed")
 
                 self._remove_stoped_session_from_file()
                 DataprocSparkSession._active_s8s_session_uuid = None
@@ -538,3 +558,66 @@ class DataprocSparkSession(SparkSession):
                 DataprocSparkSession._active_session, "session", None
             ):
                 DataprocSparkSession._active_session.session = None
+
+
+def terminate_s8s_session(
+    project_id, region, active_s8s_session_id, client_options=None
+):
+    from google.cloud.dataproc_v1 import SessionControllerClient
+
+    logger.debug(f"Terminating serverless session: {active_s8s_session_id}")
+    terminate_session_request = TerminateSessionRequest()
+    session_name = f"projects/{project_id}/locations/{region}/sessions/{active_s8s_session_id}"
+    terminate_session_request.name = session_name
+    state = None
+    try:
+        session_client = SessionControllerClient(client_options=client_options)
+        session_client.terminate_session(terminate_session_request)
+        get_session_request = GetSessionRequest()
+        get_session_request.name = session_name
+        state = Session.State.ACTIVE
+        while (
+            state != Session.State.TERMINATING
+            and state != Session.State.TERMINATED
+            and state != Session.State.FAILED
+        ):
+            session = session_client.get_session(get_session_request)
+            state = session.state
+            sleep(1)
+    except NotFound:
+        logger.debug(f"Session {active_s8s_session_id} already deleted")
+    # Client will get 'Aborted' error if session creation is still in progress and
+    # 'FailedPrecondition' if another termination is still in progress.
+    # Both are retryable but we catch it and let TTL take care of cleanups.
+    except (FailedPrecondition, Aborted):
+        logger.debug(
+            f"Session {active_s8s_session_id} already terminated manually or terminated automatically through session ttl limits"
+        )
+    if state is not None and state == Session.State.FAILED:
+        raise RuntimeError("Serverless session termination failed")
+
+
+def get_active_s8s_session_response(
+    session_name, client_options
+) -> Optional[sessions.Session]:
+    get_session_request = GetSessionRequest()
+    get_session_request.name = session_name
+    try:
+        get_session_response = SessionControllerClient(
+            client_options=client_options
+        ).get_session(get_session_request)
+        state = get_session_response.state
+    except Exception as e:
+        logger.info(f"{session_name} deleted: {e}")
+        return None
+    if state is not None and (
+        state == Session.State.ACTIVE or state == Session.State.CREATING
+    ):
+        return get_session_response
+    return None
+
+
+def is_s8s_session_active(session_name, client_options) -> bool:
+    if get_active_s8s_session_response(session_name, client_options) is None:
+        return False
+    return True

{dataproc_spark_connect-0.2.1 → dataproc_spark_connect-0.6.0}/setup.py

@@ -20,7 +20,7 @@ long_description = (this_directory / "README.md").read_text()
 
 setup(
     name="dataproc-spark-connect",
-    version="0.2.1",
+    version="0.6.0",
     description="Dataproc client library for Spark Connect",
     long_description=long_description,
     author="Google LLC",
@@ -29,11 +29,9 @@ setup(
     packages=find_namespace_packages(include=["google.*"]),
     install_requires=[
         "google-api-core>=2.19.1",
-        "google-cloud-dataproc>=5.15.1",
-        "wheel",
+        "google-cloud-dataproc>=5.18.0",
         "websockets",
-        "pyspark>=3.5",
-        "pandas",
-        "pyarrow",
+        "pyspark[connect]>=3.5",
+        "packaging>=20.0",
     ],
 )

dataproc_spark_connect-0.2.1/dataproc_spark_connect.egg-info/requires.txt

@@ -1,7 +0,0 @@
-google-api-core>=2.19.1
-google-cloud-dataproc>=5.15.1
-wheel
-websockets
-pyspark>=3.5
-pandas
-pyarrow