executorlib 1.5.3__tar.gz → 1.6.0__tar.gz

{executorlib-1.5.3 → executorlib-1.6.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: executorlib
-Version: 1.5.3
+Version: 1.6.0
 Summary: Up-scale python functions for high performance computing (HPC) with executorlib.
 Project-URL: Homepage, https://github.com/pyiron/executorlib
 Project-URL: Documentation, https://executorlib.readthedocs.io
@@ -41,13 +41,12 @@ Classifier: Development Status :: 5 - Production/Stable
 Classifier: Intended Audience :: Science/Research
 Classifier: License :: OSI Approved :: BSD License
 Classifier: Operating System :: OS Independent
-Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
 Classifier: Topic :: Scientific/Engineering :: Physics
-Requires-Python: <3.14,>=3.9
+Requires-Python: <3.14,>3.9
 Requires-Dist: cloudpickle<=3.1.1,>=2.0.0
 Requires-Dist: pyzmq<=27.0.0,>=25.0.0
 Provides-Extra: all

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/__init__.py

@@ -36,4 +36,11 @@ __all__: list[str] = [
     "SlurmClusterExecutor",
 ]
 
+try:
+    from executorlib.task_scheduler.file.queue_spawner import terminate_tasks_in_cache
+
+    __all__ += ["terminate_tasks_in_cache"]
+except ImportError:
+    pass
+
 __version__ = _version.__version__

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/_version.py

@@ -17,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '1.5.3'
-__version_tuple__ = version_tuple = (1, 5, 3)
+__version__ = version = '1.6.0'
+__version_tuple__ = version_tuple = (1, 6, 0)

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/backend/cache_parallel.py

@@ -4,6 +4,7 @@ import time
 
 import cloudpickle
 
+from executorlib.standalone.error import backend_write_error_file
 from executorlib.task_scheduler.file.backend import (
     backend_load_file,
     backend_write_file,
@@ -53,6 +54,10 @@ def main() -> None:
                 output={"error": error},
                 runtime=time.time() - time_start,
             )
+            backend_write_error_file(
+                error=error,
+                apply_dict=apply_dict,
+            )
     else:
         if mpi_rank_zero:
             backend_write_file(

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/backend/interactive_parallel.py

@@ -6,6 +6,7 @@ from typing import Optional
 import cloudpickle
 import zmq
 
+from executorlib.standalone.error import backend_write_error_file
 from executorlib.standalone.interactive.backend import call_funct, parse_arguments
 from executorlib.standalone.interactive.communication import (
     interface_connect,
@@ -82,6 +83,10 @@ def main() -> None:
                         socket=socket,
                         result_dict={"error": error},
                     )
+                    backend_write_error_file(
+                        error=error,
+                        apply_dict=input_dict,
+                    )
             else:
                 # Send output
                 if mpi_rank_zero:

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/backend/interactive_serial.py

@@ -2,6 +2,7 @@ import sys
 from os.path import abspath
 from typing import Optional
 
+from executorlib.standalone.error import backend_write_error_file
 from executorlib.standalone.interactive.backend import call_funct, parse_arguments
 from executorlib.standalone.interactive.communication import (
     interface_connect,
@@ -58,6 +59,10 @@ def main(argument_lst: Optional[list[str]] = None):
                     socket=socket,
                     result_dict={"error": error},
                 )
+                backend_write_error_file(
+                    error=error,
+                    apply_dict=input_dict,
+                )
             else:
                 # Send output
                 interface_send(socket=socket, result_dict={"result": output})

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/executor/flux.py

@@ -41,6 +41,8 @@ class FluxJobExecutor(BaseExecutor):
                                                            Defaults to None.
                               - exclusive (bool): Whether to exclusively reserve the compute nodes, or allow sharing
                                                   compute notes. Defaults to False.
+                              - error_log_file (str): Name of the error log file to use for storing exceptions raised
+                                                      by the Python functions submitted to the Executor.
         flux_executor (flux.job.FluxExecutor): Flux Python interface to submit the workers to flux
         flux_executor_pmi_mode (str): PMI interface to use (OpenMPI v5 requires pmix) default is None (Flux only)
         flux_executor_nesting (bool): Provide hierarchically nested Flux job scheduler inside the submitted function.
@@ -126,6 +128,8 @@ class FluxJobExecutor(BaseExecutor):
                                                                Defaults to None.
                                   - exclusive (bool): Whether to exclusively reserve the compute nodes, or allow sharing
                                                       compute notes. Defaults to False.
+                                  - error_log_file (str): Name of the error log file to use for storing exceptions
+                                                          raised by the Python functions submitted to the Executor.
             flux_executor (flux.job.FluxExecutor): Flux Python interface to submit the workers to flux
             flux_executor_pmi_mode (str): PMI interface to use (OpenMPI v5 requires pmix) default is None (Flux only)
             flux_executor_nesting (bool): Provide hierarchically nested Flux job scheduler inside the submitted function.
@@ -229,6 +233,8 @@ class FluxClusterExecutor(BaseExecutor):
                               - openmpi_oversubscribe (bool): adds the `--oversubscribe` command line flag (OpenMPI and
                                                               SLURM only) - default False
                               - slurm_cmd_args (list): Additional command line arguments for the srun call (SLURM only)
+                              - error_log_file (str): Name of the error log file to use for storing exceptions raised
+                                                      by the Python functions submitted to the Executor.
         pysqa_config_directory (str, optional): path to the pysqa config directory (only for pysqa based backend).
         hostname_localhost (boolean): use localhost instead of the hostname to establish the zmq connection. In the
                                       context of an HPC cluster this essential to be able to communicate to an
@@ -308,6 +314,8 @@ class FluxClusterExecutor(BaseExecutor):
                                                                   and SLURM only) - default False
                                   - slurm_cmd_args (list): Additional command line arguments for the srun call (SLURM
                                                            only)
+                                  - error_log_file (str): Name of the error log file to use for storing exceptions
+                                                          raised by the Python functions submitted to the Executor.
             pysqa_config_directory (str, optional): path to the pysqa config directory (only for pysqa based backend).
             hostname_localhost (boolean): use localhost instead of the hostname to establish the zmq connection. In the
                                       context of an HPC cluster this essential to be able to communicate to an
@@ -424,6 +432,8 @@ def create_flux_executor(
                                                            Defaults to None.
                               - exclusive (bool): Whether to exclusively reserve the compute nodes, or allow sharing
                                                   compute notes. Defaults to False.
+                              - error_log_file (str): Name of the error log file to use for storing exceptions raised
+                                                      by the Python functions submitted to the Executor.
         flux_executor (flux.job.FluxExecutor): Flux Python interface to submit the workers to flux
         flux_executor_pmi_mode (str): PMI interface to use (OpenMPI v5 requires pmix) default is None (Flux only)
         flux_executor_nesting (bool): Provide hierarchically nested Flux job scheduler inside the submitted function.

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/executor/single.py

@@ -39,6 +39,8 @@ class SingleNodeExecutor(BaseExecutor):
                               - openmpi_oversubscribe (bool): adds the `--oversubscribe` command line flag (OpenMPI and
                                                               SLURM only) - default False
                               - slurm_cmd_args (list): Additional command line arguments for the srun call (SLURM only)
+                              - error_log_file (str): Name of the error log file to use for storing exceptions raised
+                                                      by the Python functions submitted to the Executor.
         hostname_localhost (boolean): use localhost instead of the hostname to establish the zmq connection. In the
                                       context of an HPC cluster this essential to be able to communicate to an
                                       Executor running on a different compute node within the same allocation. And
@@ -116,6 +118,8 @@ class SingleNodeExecutor(BaseExecutor):
                                                                   and SLURM only) - default False
                                   - slurm_cmd_args (list): Additional command line arguments for the srun call (SLURM
                                                            only)
+                                  - error_log_file (str): Name of the error log file to use for storing exceptions
+                                                          raised by the Python functions submitted to the Executor.
             hostname_localhost (boolean): use localhost instead of the hostname to establish the zmq connection. In the
                                       context of an HPC cluster this essential to be able to communicate to an
                                       Executor running on a different compute node within the same allocation. And
@@ -202,6 +206,8 @@ class TestClusterExecutor(BaseExecutor):
                               - threads_per_core (int): number of OpenMP threads to be used for each function call
                               - gpus_per_core (int): number of GPUs per worker - defaults to 0
                               - cwd (str/None): current working directory where the parallel python task is executed
+                              - error_log_file (str): Name of the error log file to use for storing exceptions raised
+                                                      by the Python functions submitted to the Executor.
         hostname_localhost (boolean): use localhost instead of the hostname to establish the zmq connection. In the
                                       context of an HPC cluster this essential to be able to communicate to an
                                       Executor running on a different compute node within the same allocation. And
@@ -273,6 +279,8 @@ class TestClusterExecutor(BaseExecutor):
                                   - threads_per_core (int): number of OpenMP threads to be used for each function call
                                   - gpus_per_core (int): number of GPUs per worker - defaults to 0
                                   - cwd (str/None): current working directory where the parallel python task is executed
+                                  - error_log_file (str): Name of the error log file to use for storing exceptions
+                                                          raised by the Python functions submitted to the Executor.
             hostname_localhost (boolean): use localhost instead of the hostname to establish the zmq connection. In the
                                       context of an HPC cluster this essential to be able to communicate to an
                                       Executor running on a different compute node within the same allocation. And
@@ -381,6 +389,8 @@ def create_single_node_executor(
                                                               and SLURM only) - default False
                               - slurm_cmd_args (list): Additional command line arguments for the srun call (SLURM
                                                        only)
+                              - error_log_file (str): Name of the error log file to use for storing exceptions raised
+                                                      by the Python functions submitted to the Executor.
         hostname_localhost (boolean): use localhost instead of the hostname to establish the zmq connection. In the
                                   context of an HPC cluster this essential to be able to communicate to an
                                   Executor running on a different compute node within the same allocation. And

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/executor/slurm.py

@@ -41,6 +41,8 @@ class SlurmClusterExecutor(BaseExecutor):
                               - openmpi_oversubscribe (bool): adds the `--oversubscribe` command line flag (OpenMPI and
                                                               SLURM only) - default False
                               - slurm_cmd_args (list): Additional command line arguments for the srun call (SLURM only)
+                              - error_log_file (str): Name of the error log file to use for storing exceptions raised
+                                                      by the Python functions submitted to the Executor.
         pysqa_config_directory (str, optional): path to the pysqa config directory (only for pysqa based backend).
         hostname_localhost (boolean): use localhost instead of the hostname to establish the zmq connection. In the
                                       context of an HPC cluster this essential to be able to communicate to an
@@ -120,6 +122,8 @@ class SlurmClusterExecutor(BaseExecutor):
                                                                   and SLURM only) - default False
                                   - slurm_cmd_args (list): Additional command line arguments for the srun call (SLURM
                                                            only)
+                                  - error_log_file (str): Name of the error log file to use for storing exceptions
+                                                          raised by the Python functions submitted to the Executor.
             pysqa_config_directory (str, optional): path to the pysqa config directory (only for pysqa based backend).
             hostname_localhost (boolean): use localhost instead of the hostname to establish the zmq connection. In the
                                       context of an HPC cluster this essential to be able to communicate to an
@@ -226,6 +230,8 @@ class SlurmJobExecutor(BaseExecutor):
                                                            Defaults to None.
                               - exclusive (bool): Whether to exclusively reserve the compute nodes, or allow sharing
                                                   compute notes. Defaults to False.
+                              - error_log_file (str): Name of the error log file to use for storing exceptions raised
+                                                      by the Python functions submitted to the Executor.
         hostname_localhost (boolean): use localhost instead of the hostname to establish the zmq connection. In the
                                       context of an HPC cluster this essential to be able to communicate to an
                                       Executor running on a different compute node within the same allocation. And
@@ -307,6 +313,8 @@ class SlurmJobExecutor(BaseExecutor):
                                                            Defaults to None.
                                   - exclusive (bool): Whether to exclusively reserve the compute nodes, or allow sharing
                                                       compute notes. Defaults to False.
+                                  - error_log_file (str): Name of the error log file to use for storing exceptions
+                                                          raised by the Python functions submitted to the Executor.
             hostname_localhost (boolean): use localhost instead of the hostname to establish the zmq connection. In the
                                       context of an HPC cluster this essential to be able to communicate to an
                                       Executor running on a different compute node within the same allocation. And
@@ -408,6 +416,8 @@ def create_slurm_executor(
                                                            Defaults to None.
                               - exclusive (bool): Whether to exclusively reserve the compute nodes, or allow sharing
                                                   compute notes. Defaults to False.
+                              - error_log_file (str): Name of the error log file to use for storing exceptions raised
+                                                      by the Python functions submitted to the Executor.
         hostname_localhost (boolean): use localhost instead of the hostname to establish the zmq connection. In the
                                   context of an HPC cluster this essential to be able to communicate to an
                                   Executor running on a different compute node within the same allocation. And

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/standalone/cache.py

@@ -10,6 +10,7 @@ group_dict = {
     "error": "error",
     "runtime": "runtime",
     "queue_id": "queue_id",
+    "error_log_file": "error_log_file",
 }
 
 

executorlib-1.6.0/executorlib/standalone/error.py

@@ -0,0 +1,21 @@
+import traceback
+
+
+def backend_write_error_file(error: Exception, apply_dict: dict) -> None:
+    """
+    Write an error to a file if specified in the apply_dict.
+
+    Args:
+        error (Exception): The error to be written.
+        apply_dict (dict): Dictionary containing additional parameters.
+
+    Returns:
+        None
+    """
+    error_log_file = apply_dict.get("error_log_file")
+    if error_log_file is not None:
+        with open(error_log_file, "a") as f:
+            f.write("function: " + str(apply_dict["fn"]) + "\n")
+            f.write("args: " + str(apply_dict["args"]) + "\n")
+            f.write("kwargs: " + str(apply_dict["kwargs"]) + "\n")
+            traceback.print_exception(error, file=f)

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/task_scheduler/file/backend.py

@@ -2,6 +2,7 @@ import os
 import time
 from typing import Any
 
+from executorlib.standalone.error import backend_write_error_file
 from executorlib.task_scheduler.file.hdf import dump, load
 from executorlib.task_scheduler.file.shared import FutureItem
 
@@ -77,6 +78,10 @@ def backend_execute_task_in_file(file_name: str) -> None:
         }
     except Exception as error:
         result = {"error": error}
+        backend_write_error_file(
+            error=error,
+            apply_dict=apply_dict,
+        )
 
     backend_write_file(
         file_name=file_name,

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/task_scheduler/file/hdf.py

@@ -52,6 +52,10 @@ def load(file_name: str) -> dict:
             data_dict["kwargs"] = cloudpickle.loads(np.void(hdf["/input_kwargs"]))
         else:
             data_dict["kwargs"] = {}
+        if "error_log_file" in hdf:
+            data_dict["error_log_file"] = cloudpickle.loads(
+                np.void(hdf["/error_log_file"])
+            )
         return data_dict
 
 

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/task_scheduler/file/queue_spawner.py

@@ -1,3 +1,4 @@
+import contextlib
 import os
 import subprocess
 from typing import Optional, Union
@@ -32,7 +33,7 @@ def execute_with_pysqa(
                                   cwd: None,
                               }
         config_directory (str, optional): path to the config directory.
-        backend (str, optional): name of the backend used to spawn tasks.
+        backend (str, optional): name of the backend used to spawn tasks ["slurm", "flux"].
 
     Returns:
         int: queuing system ID
@@ -101,7 +102,7 @@ def terminate_with_pysqa(
     Args:
         queue_id (int): Queuing system ID of the job to delete.
         config_directory (str, optional): path to the config directory.
-        backend (str, optional): name of the backend used to spawn tasks.
+        backend (str, optional): name of the backend used to spawn tasks ["slurm", "flux"].
     """
     qa = QueueAdapter(
         directory=config_directory,
@@ -110,7 +111,35 @@ def terminate_with_pysqa(
     )
     status = qa.get_status_of_job(process_id=queue_id)
     if status is not None and status not in ["finished", "error"]:
-        qa.delete_job(process_id=queue_id)
+        with contextlib.suppress(subprocess.CalledProcessError):
+            qa.delete_job(process_id=queue_id)
+
+
+def terminate_tasks_in_cache(
+    cache_directory: str,
+    config_directory: Optional[str] = None,
+    backend: Optional[str] = None,
+):
+    """
+    Delete all jobs stored in the cache directory from the queuing system
+
+    Args:
+        cache_directory (str): The directory to store cache files.
+        config_directory (str, optional): path to the config directory.
+        backend (str, optional): name of the backend used to spawn tasks ["slurm", "flux"].
+    """
+    hdf5_file_lst = []
+    for root, _, files in os.walk(cache_directory):
+        hdf5_file_lst += [os.path.join(root, f) for f in files if f[-5:] == "_i.h5"]
+
+    for f in hdf5_file_lst:
+        queue_id = get_queue_id(f)
+        if queue_id is not None:
+            terminate_with_pysqa(
+                queue_id=queue_id,
+                config_directory=config_directory,
+                backend=backend,
+            )
 
 
 def _pysqa_execute_command(

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/task_scheduler/file/shared.py

@@ -126,6 +126,7 @@ def execute_tasks_h5(
             )
             cache_key = task_resource_dict.pop("cache_key", None)
             cache_directory = os.path.abspath(task_resource_dict.pop("cache_directory"))
+            error_log_file = task_resource_dict.pop("error_log_file", None)
             task_key, data_dict = serialize_funct_h5(
                 fn=task_dict["fn"],
                 fn_args=task_args,
@@ -133,6 +134,7 @@ def execute_tasks_h5(
                 resource_dict=task_resource_dict,
                 cache_key=cache_key,
             )
+            data_dict["error_log_file"] = error_log_file
             if task_key not in memory_dict:
                 if os.path.join(
                     cache_directory, task_key + "_o.h5"

{executorlib-1.5.3 → executorlib-1.6.0}/executorlib/task_scheduler/interactive/shared.py

@@ -26,6 +26,7 @@ def execute_tasks(
     cache_key: Optional[str] = None,
     queue_join_on_shutdown: bool = True,
     log_obj_size: bool = False,
+    error_log_file: Optional[str] = None,
     **kwargs,
 ) -> None:
     """
@@ -70,6 +71,8 @@ def execute_tasks(
                 future_queue.join()
             break
         elif "fn" in task_dict and "future" in task_dict:
+            if error_log_file is not None:
+                task_dict["error_log_file"] = error_log_file
             if cache_directory is None:
                 _execute_task_without_cache(
                     interface=interface, task_dict=task_dict, future_queue=future_queue

{executorlib-1.5.3 → executorlib-1.6.0}/pyproject.toml

@@ -16,14 +16,13 @@ authors = [
 readme = "README.md"
 license = { file = "LICENSE" }
 keywords = ["high performance computing", "hpc", "task scheduler", "slurm", "flux-framework", "executor"]
-requires-python = ">=3.9, <3.14"
+requires-python = ">3.9, <3.14"
 classifiers = [
     "Development Status :: 5 - Production/Stable",
     "Topic :: Scientific/Engineering :: Physics",
     "License :: OSI Approved :: BSD License",
     "Intended Audience :: Science/Research",
     "Operating System :: OS Independent",
-    "Programming Language :: Python :: 3.9",
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",