fractal-server 2.13.1__py3-none-any.whl → 2.14.0a0__py3-none-any.whl

fractal_server/app/runner/v2/runner_functions.py

@@ -1,7 +1,6 @@
 import functools
 import logging
 import traceback
-from concurrent.futures import Executor
 from pathlib import Path
 from typing import Any
 from typing import Callable
@@ -20,6 +19,7 @@ from fractal_server.app.models.v2 import TaskV2
 from fractal_server.app.models.v2 import WorkflowTaskV2
 from fractal_server.app.runner.components import _COMPONENT_KEY_
 from fractal_server.app.runner.components import _index_to_component
+from fractal_server.app.runner.executors.base_runner import BaseRunner
 
 
 __all__ = [
@@ -59,13 +59,7 @@ def _cast_and_validate_InitTaskOutput(
         )
 
 
-def no_op_submit_setup_call(
-    *,
-    wftask: WorkflowTaskV2,
-    workflow_dir_local: Path,
-    workflow_dir_remote: Path,
-    which_type: Literal["non_parallel", "parallel"],
-) -> dict:
+def no_op_submit_setup_call(*args, **kwargs) -> dict:
     """
     Default (no-operation) interface of submit_setup_call in V2.
     """
@@ -84,8 +78,8 @@ def _get_executor_options(
     try:
         options = submit_setup_call(
             wftask=wftask,
-            workflow_dir_local=workflow_dir_local,
-            workflow_dir_remote=workflow_dir_remote,
+            root_dir_local=workflow_dir_local,
+            root_dir_remote=workflow_dir_remote,
             which_type=which_type,
         )
     except Exception as e:
@@ -114,9 +108,10 @@ def run_v2_task_non_parallel(
     wftask: WorkflowTaskV2,
     workflow_dir_local: Path,
     workflow_dir_remote: Optional[Path] = None,
-    executor: Executor,
+    executor: BaseRunner,
     submit_setup_call: Callable = no_op_submit_setup_call,
-) -> tuple[TaskOutput, int]:
+    history_item_id: int,
+) -> tuple[TaskOutput, int, dict[int, BaseException]]:
     """
     This runs server-side (see `executor` argument)
     """
@@ -141,23 +136,29 @@ def run_v2_task_non_parallel(
         zarr_dir=zarr_dir,
         **(wftask.args_non_parallel or {}),
     )
-    future = executor.submit(
+    function_kwargs[_COMPONENT_KEY_] = _index_to_component(0)
+
+    result, exception = executor.submit(
         functools.partial(
             run_single_task,
             wftask=wftask,
             command=task.command_non_parallel,
-            workflow_dir_local=workflow_dir_local,
-            workflow_dir_remote=workflow_dir_remote,
+            root_dir_local=workflow_dir_local,
+            root_dir_remote=workflow_dir_remote,
         ),
-        function_kwargs,
+        parameters=function_kwargs,
+        history_item_id=history_item_id,
         **executor_options,
     )
-    output = future.result()
+
     num_tasks = 1
-    if output is None:
-        return (TaskOutput(), num_tasks)
+    if exception is None:
+        if result is None:
+            return (TaskOutput(), num_tasks, {})
+        else:
+            return (_cast_and_validate_TaskOutput(result), num_tasks, {})
     else:
-        return (_cast_and_validate_TaskOutput(output), num_tasks)
+        return (TaskOutput(), num_tasks, {0: exception})
 
 
 def run_v2_task_parallel(
@@ -165,14 +166,15 @@ def run_v2_task_parallel(
     images: list[dict[str, Any]],
     task: TaskV2,
     wftask: WorkflowTaskV2,
-    executor: Executor,
+    executor: BaseRunner,
     workflow_dir_local: Path,
     workflow_dir_remote: Optional[Path] = None,
     submit_setup_call: Callable = no_op_submit_setup_call,
-) -> tuple[TaskOutput, int]:
+    history_item_id: int,
+) -> tuple[TaskOutput, int, dict[int, BaseException]]:
 
     if len(images) == 0:
-        return (TaskOutput(), 0)
+        return (TaskOutput(), 0, {})
 
     _check_parallelization_list_size(images)
 
@@ -194,30 +196,36 @@ def run_v2_task_parallel(
         )
         list_function_kwargs[-1][_COMPONENT_KEY_] = _index_to_component(ind)
 
-    results_iterator = executor.map(
+    results, exceptions = executor.multisubmit(
         functools.partial(
             run_single_task,
             wftask=wftask,
             command=task.command_parallel,
-            workflow_dir_local=workflow_dir_local,
-            workflow_dir_remote=workflow_dir_remote,
+            root_dir_local=workflow_dir_local,
+            root_dir_remote=workflow_dir_remote,
         ),
-        list_function_kwargs,
+        list_parameters=list_function_kwargs,
+        history_item_id=history_item_id,
         **executor_options,
     )
-    # Explicitly iterate over the whole list, so that all futures are waited
-    outputs = list(results_iterator)
 
-    # Validate all non-None outputs
-    for ind, output in enumerate(outputs):
-        if output is None:
-            outputs[ind] = TaskOutput()
+    outputs = []
+    for ind in range(len(list_function_kwargs)):
+        if ind in results.keys():
+            result = results[ind]
+            if result is None:
+                output = TaskOutput()
+            else:
+                output = _cast_and_validate_TaskOutput(result)
+            outputs.append(output)
+        elif ind in exceptions.keys():
+            print(f"Bad: {exceptions[ind]}")
         else:
-            outputs[ind] = _cast_and_validate_TaskOutput(output)
+            print("VERY BAD - should have not reached this point")
 
     num_tasks = len(images)
     merged_output = merge_outputs(outputs)
-    return (merged_output, num_tasks)
+    return (merged_output, num_tasks, exceptions)
 
 
 def run_v2_task_compound(
@@ -226,11 +234,12 @@ def run_v2_task_compound(
     zarr_dir: str,
     task: TaskV2,
     wftask: WorkflowTaskV2,
-    executor: Executor,
+    executor: BaseRunner,
     workflow_dir_local: Path,
     workflow_dir_remote: Optional[Path] = None,
     submit_setup_call: Callable = no_op_submit_setup_call,
-) -> TaskOutput:
+    history_item_id: int,
+) -> tuple[TaskOutput, int, dict[int, BaseException]]:
 
     executor_options_init = _get_executor_options(
         wftask=wftask,
@@ -253,32 +262,40 @@ def run_v2_task_compound(
         zarr_dir=zarr_dir,
         **(wftask.args_non_parallel or {}),
     )
-    future = executor.submit(
+    function_kwargs[_COMPONENT_KEY_] = f"init_{_index_to_component(0)}"
+    result, exception = executor.submit(
         functools.partial(
             run_single_task,
             wftask=wftask,
             command=task.command_non_parallel,
-            workflow_dir_local=workflow_dir_local,
-            workflow_dir_remote=workflow_dir_remote,
+            root_dir_local=workflow_dir_local,
+            root_dir_remote=workflow_dir_remote,
         ),
-        function_kwargs,
+        parameters=function_kwargs,
+        history_item_id=history_item_id,
+        in_compound_task=True,
         **executor_options_init,
     )
-    output = future.result()
-    if output is None:
-        init_task_output = InitTaskOutput()
+
+    num_tasks = 1
+    if exception is None:
+        if result is None:
+            init_task_output = InitTaskOutput()
+        else:
+            init_task_output = _cast_and_validate_InitTaskOutput(result)
     else:
-        init_task_output = _cast_and_validate_InitTaskOutput(output)
+        return (TaskOutput(), num_tasks, {0: exception})
+
     parallelization_list = init_task_output.parallelization_list
     parallelization_list = deduplicate_list(parallelization_list)
 
-    num_task = 1 + len(parallelization_list)
+    num_tasks = 1 + len(parallelization_list)
 
     # 3/B: parallel part of a compound task
     _check_parallelization_list_size(parallelization_list)
 
     if len(parallelization_list) == 0:
-        return (TaskOutput(), 0)
+        return (TaskOutput(), 0, {})
 
     list_function_kwargs = []
     for ind, parallelization_item in enumerate(parallelization_list):
@@ -289,29 +306,35 @@ def run_v2_task_compound(
                 **(wftask.args_parallel or {}),
             ),
         )
-        list_function_kwargs[-1][_COMPONENT_KEY_] = _index_to_component(ind)
+        list_function_kwargs[-1][
+            _COMPONENT_KEY_
+        ] = f"compute_{_index_to_component(ind)}"
 
-    results_iterator = executor.map(
+    results, exceptions = executor.multisubmit(
         functools.partial(
             run_single_task,
             wftask=wftask,
             command=task.command_parallel,
-            workflow_dir_local=workflow_dir_local,
-            workflow_dir_remote=workflow_dir_remote,
+            root_dir_local=workflow_dir_local,
+            root_dir_remote=workflow_dir_remote,
         ),
-        list_function_kwargs,
+        list_parameters=list_function_kwargs,
+        history_item_id=history_item_id,
+        in_compound_task=True,
         **executor_options_compute,
     )
-    # Explicitly iterate over the whole list, so that all futures are waited
-    outputs = list(results_iterator)
 
-    # Validate all non-None outputs
-    for ind, output in enumerate(outputs):
-        if output is None:
-            outputs[ind] = TaskOutput()
-        else:
-            validated_output = _cast_and_validate_TaskOutput(output)
-            outputs[ind] = validated_output
+    outputs = []
+    for ind in range(len(list_function_kwargs)):
+        if ind in results.keys():
+            result = results[ind]
+            if result is None:
+                output = TaskOutput()
+            else:
+                output = _cast_and_validate_TaskOutput(result)
+            outputs.append(output)
+        elif ind in exceptions.keys():
+            print(f"Bad: {exceptions[ind]}")
 
     merged_output = merge_outputs(outputs)
-    return (merged_output, num_task)
+    return (merged_output, num_tasks, exceptions)

fractal_server/app/runner/v2/runner_functions_low_level.py

@@ -11,11 +11,11 @@ from ..components import _COMPONENT_KEY_
 from ..exceptions import JobExecutionError
 from ..exceptions import TaskExecutionError
 from fractal_server.app.models.v2 import WorkflowTaskV2
-from fractal_server.app.runner.task_files import get_task_file_paths
+from fractal_server.app.runner.task_files import TaskFiles
 from fractal_server.string_tools import validate_cmd
 
 
-def _call_command_wrapper(cmd: str, log_path: Path) -> None:
+def _call_command_wrapper(cmd: str, log_path: str) -> None:
     """
     Call a command and write its stdout and stderr to files
 
@@ -50,7 +50,7 @@ def _call_command_wrapper(cmd: str, log_path: Path) -> None:
             raise e
 
     if result.returncode > 0:
-        with log_path.open("r") as fp_stderr:
+        with open(log_path, "r") as fp_stderr:
             err = fp_stderr.read()
         raise TaskExecutionError(err)
     elif result.returncode < 0:
@@ -60,49 +60,49 @@ def _call_command_wrapper(cmd: str, log_path: Path) -> None:
 
 
 def run_single_task(
-    args: dict[str, Any],
+    parameters: dict[str, Any],
     command: str,
     wftask: WorkflowTaskV2,
-    workflow_dir_local: Path,
-    workflow_dir_remote: Optional[Path] = None,
+    root_dir_local: Path,
+    root_dir_remote: Optional[Path] = None,
     logger_name: Optional[str] = None,
 ) -> dict[str, Any]:
     """
-    Runs within an executor.
+    Runs within an executor (AKA on the SLURM cluster).
     """
 
     logger = logging.getLogger(logger_name)
     logger.debug(f"Now start running {command=}")
 
-    if not workflow_dir_remote:
-        workflow_dir_remote = workflow_dir_local
+    if not root_dir_remote:
+        root_dir_remote = root_dir_local
 
     task_name = wftask.task.name
 
-    component = args.pop(_COMPONENT_KEY_, None)
-    task_files = get_task_file_paths(
-        workflow_dir_local=workflow_dir_local,
-        workflow_dir_remote=workflow_dir_remote,
-        task_order=wftask.order,
+    component = parameters.pop(_COMPONENT_KEY_)
+    task_files = TaskFiles(
+        root_dir_local=root_dir_local,
+        root_dir_remote=root_dir_remote,
         task_name=task_name,
+        task_order=wftask.order,
         component=component,
     )
 
     # Write arguments to args.json file
-    with task_files.args.open("w") as f:
-        json.dump(args, f, indent=2)
+    with open(task_files.args_file_remote, "w") as f:
+        json.dump(parameters, f, indent=2)
 
     # Assemble full command
     full_command = (
         f"{command} "
-        f"--args-json {task_files.args.as_posix()} "
-        f"--out-json {task_files.metadiff.as_posix()}"
+        f"--args-json {task_files.args_file_remote} "
+        f"--out-json {task_files.metadiff_file_remote}"
     )
 
     try:
         _call_command_wrapper(
             full_command,
-            log_path=task_files.log,
+            log_path=task_files.log_file_remote,
         )
     except TaskExecutionError as e:
         e.workflow_task_order = wftask.order
@@ -111,7 +111,7 @@ def run_single_task(
         raise e
 
     try:
-        with task_files.metadiff.open("r") as f:
+        with open(task_files.metadiff_file_remote, "r") as f:
             out_meta = json.load(f)
     except FileNotFoundError as e:
         logger.debug(

fractal_server/app/schemas/v2/dataset.py

@@ -14,27 +14,12 @@ from .._filter_validators import validate_attribute_filters
 from .._filter_validators import validate_type_filters
 from .._validators import root_validate_dict_keys
 from .._validators import valstr
-from .dumps import WorkflowTaskDumpV2
 from .project import ProjectReadV2
-from .workflowtask import WorkflowTaskStatusTypeV2
 from fractal_server.images import SingleImage
 from fractal_server.images.models import AttributeFiltersType
 from fractal_server.urls import normalize_url
 
 
-class _DatasetHistoryItemV2(BaseModel):
-    """
-    Class for an item of `Dataset.history`.
-    """
-
-    workflowtask: WorkflowTaskDumpV2
-    status: WorkflowTaskStatusTypeV2
-    parallelization: Optional[dict] = None
-
-
-# CRUD
-
-
 class DatasetCreateV2(BaseModel):
     model_config = ConfigDict(extra="forbid")
 
@@ -74,8 +59,6 @@ class DatasetReadV2(BaseModel):
     project_id: int
     project: ProjectReadV2
 
-    history: list[_DatasetHistoryItemV2]
-
     timestamp_created: AwareDatetime
 
     zarr_dir: str

fractal_server/app/schemas/v2/history.py

@@ -0,0 +1,23 @@
+from datetime import datetime
+from typing import Any
+
+from pydantic import BaseModel
+from pydantic import field_serializer
+from pydantic.types import AwareDatetime
+
+
+class HistoryItemV2Read(BaseModel):
+    id: int
+    dataset_id: int
+    workflowtask_id: int
+    timestamp_started: AwareDatetime
+    parameters_hash: str
+    num_available_images: int
+    num_current_images: int
+    images: dict[str, str]
+    workflowtask_dump: dict[str, Any]
+    task_group_dump: dict[str, Any]
+
+    @field_serializer("timestamp_started")
+    def serialize_datetime(v: datetime) -> str:
+        return v.isoformat()

fractal_server/config.py

@@ -705,7 +705,7 @@ class Settings(BaseSettings):
 
         info = f"FRACTAL_RUNNER_BACKEND={self.FRACTAL_RUNNER_BACKEND}"
         if self.FRACTAL_RUNNER_BACKEND == "slurm":
-            from fractal_server.app.runner.executors.slurm._slurm_config import (  # noqa: E501
+            from fractal_server.app.runner.executors.slurm_common._slurm_config import (  # noqa: E501
                 load_slurm_config_file,
             )
 
@@ -735,7 +735,7 @@ class Settings(BaseSettings):
                     f"Must set FRACTAL_SLURM_WORKER_PYTHON when {info}"
                 )
 
-            from fractal_server.app.runner.executors.slurm._slurm_config import (  # noqa: E501
+            from fractal_server.app.runner.executors.slurm_common._slurm_config import (  # noqa: E501
                 load_slurm_config_file,
             )
 

fractal_server/migrations/versions/8223fcef886c_image_status.py

@@ -0,0 +1,63 @@
+"""image status
+
+Revision ID: 8223fcef886c
+Revises: 87cd72a537a2
+Create Date: 2025-02-24 11:32:22.267338
+
+"""
+import sqlalchemy as sa
+import sqlmodel
+from alembic import op
+
+
+# revision identifiers, used by Alembic.
+revision = "8223fcef886c"
+down_revision = "87cd72a537a2"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.create_table(
+        "imagestatus",
+        sa.Column(
+            "zarr_url", sqlmodel.sql.sqltypes.AutoString(), nullable=False
+        ),
+        sa.Column("workflowtask_id", sa.Integer(), nullable=False),
+        sa.Column("dataset_id", sa.Integer(), nullable=False),
+        sa.Column(
+            "parameters_hash",
+            sqlmodel.sql.sqltypes.AutoString(),
+            nullable=False,
+        ),
+        sa.Column(
+            "status", sqlmodel.sql.sqltypes.AutoString(), nullable=False
+        ),
+        sa.Column(
+            "logfile", sqlmodel.sql.sqltypes.AutoString(), nullable=False
+        ),
+        sa.ForeignKeyConstraint(
+            ["dataset_id"],
+            ["datasetv2.id"],
+            name=op.f("fk_imagestatus_dataset_id_datasetv2"),
+        ),
+        sa.ForeignKeyConstraint(
+            ["workflowtask_id"],
+            ["workflowtaskv2.id"],
+            name=op.f("fk_imagestatus_workflowtask_id_workflowtaskv2"),
+        ),
+        sa.PrimaryKeyConstraint(
+            "zarr_url",
+            "workflowtask_id",
+            "dataset_id",
+            name=op.f("pk_imagestatus"),
+        ),
+    )
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_table("imagestatus")
+    # ### end Alembic commands ###

fractal_server/migrations/versions/87cd72a537a2_add_historyitem_table.py

@@ -0,0 +1,68 @@
+"""Add HistoryItem table
+
+Revision ID: 87cd72a537a2
+Revises: af1ef1c83c9b
+Create Date: 2025-02-18 10:48:16.401995
+
+"""
+import sqlalchemy as sa
+import sqlmodel
+from alembic import op
+from sqlalchemy.dialects import postgresql
+
+# revision identifiers, used by Alembic.
+revision = "87cd72a537a2"
+down_revision = "af1ef1c83c9b"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.create_table(
+        "historyitemv2",
+        sa.Column("id", sa.Integer(), nullable=False),
+        sa.Column("dataset_id", sa.Integer(), nullable=False),
+        sa.Column("workflowtask_id", sa.Integer(), nullable=True),
+        sa.Column(
+            "timestamp_started", sa.DateTime(timezone=True), nullable=False
+        ),
+        sa.Column(
+            "workflowtask_dump",
+            postgresql.JSONB(astext_type=sa.Text()),
+            nullable=False,
+        ),
+        sa.Column(
+            "task_group_dump",
+            postgresql.JSONB(astext_type=sa.Text()),
+            nullable=False,
+        ),
+        sa.Column(
+            "parameters_hash",
+            sqlmodel.sql.sqltypes.AutoString(),
+            nullable=False,
+        ),
+        sa.Column("num_available_images", sa.Integer(), nullable=False),
+        sa.Column("num_current_images", sa.Integer(), nullable=False),
+        sa.Column(
+            "images", postgresql.JSONB(astext_type=sa.Text()), nullable=False
+        ),
+        sa.ForeignKeyConstraint(
+            ["dataset_id"],
+            ["datasetv2.id"],
+            name=op.f("fk_historyitemv2_dataset_id_datasetv2"),
+        ),
+        sa.ForeignKeyConstraint(
+            ["workflowtask_id"],
+            ["workflowtaskv2.id"],
+            name=op.f("fk_historyitemv2_workflowtask_id_workflowtaskv2"),
+        ),
+        sa.PrimaryKeyConstraint("id", name=op.f("pk_historyitemv2")),
+    )
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.drop_table("historyitemv2")
+    # ### end Alembic commands ###

{fractal_server-2.13.1.dist-info → fractal_server-2.14.0a0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: fractal-server
-Version: 2.13.1
+Version: 2.14.0a0
 Summary: Backend component of the Fractal analytics platform
 Home-page: https://github.com/fractal-analytics-platform/fractal-server
 License: BSD-3-Clause