Add JobResources class

Added a class to represent job resources required/requested for a batch
job. Includes documentation and some unit tests, but not enough.
Incorporated into `_click_batch`.

flepimop/gempyor_pkg/src/gempyor/batch.py

@@ -5,7 +5,7 @@
 metadata and job size calculations for example.
 """
 
-__all__ = ["BatchSystem", "JobSize", "JobTimeLimit", "write_manifest"]
+__all__ = ["BatchSystem", "JobSize", "JobResources", "JobTimeLimit", "write_manifest"]
 
 
 from dataclasses import dataclass
@@ -150,14 +150,90 @@ def size_from_jobs_sims_blocks(
         Returns:
             A job size instance with either the explicit or inferred job sizing.
         """
-        inference_method = (
-            inference_method if inference_method is None else inference_method.lower()
-        )
         if inference_method == "emcee":
             return cls(jobs=jobs, simulations=blocks * simulations, blocks=1)
         return cls(jobs=jobs, simulations=simulations, blocks=blocks)
 
 
+@dataclass(frozen=True, slots=True)
+class JobResources:
+    """
+    A batch submission job resources request.
+
+    Attributes:
+        nodes: The number of nodes to request.
+        cpus: The number of CPUs to request per a node.
+        memory: The amount of memory to request per a node in MB.
+
+    Raises:
+        ValueError: If any of the attributes are less than 1.
+    """
+
+    nodes: int
+    cpus: int
+    memory: int
+
+    def __post_init__(self) -> None:
+        for p in self.__slots__:
+            if (val := getattr(self, p)) < 1:
+                raise ValueError(
+                    (
+                        f"The '{p}' attribute must be greater than 0, "
+                        f"but instead was given '{val}'."
+                    )
+                )
+
+    @classmethod
+    def from_presets(
+        cls, job_size: JobSize, inference_method: Literal["emcee"] | None
+    ) -> "JobResources":
+        """
+        Calculate suggested job resources from presets.
+
+        Args:
+            job_size: The size of the job being ran.
+            inference_method: The inference method being used for this job.
+
+        Returns:
+            A job resources instances scaled to the job size given.
+        """
+        if inference_method == "emcee":
+            return cls(
+                nodes=1, cpus=2 * job_size.jobs, memory=2 * 1024 * job_size.simulations
+            )
+        return cls(nodes=job_size.jobs, cpus=2, memory=2 * 1024)
+
+    @property
+    def total_cpus(self) -> int:
+        """
+        Calculate the total number of CPUs.
+
+        Returns:
+            The total number of CPUs represented by this instance.
+        """
+        return self.nodes * self.cpus
+
+    @property
+    def total_memory(self) -> int:
+        """
+        Calculate the total amount of memory.
+
+        Returns:
+            The total amount of memory represented by this instance.
+        """
+        return self.nodes * self.memory
+
+    def total_resources(self) -> tuple[int, int, int]:
+        """
+        Calculate the total resources.
+
+        Returns:
+            A tuple of the nodes, total CPUs, and total memory represented by
+            this instance.
+        """
+        return (self.nodes, self.total_cpus, self.total_memory)
+
+
 @dataclass(frozen=True, slots=True)
 class JobTimeLimit:
     """
@@ -261,8 +337,6 @@ def from_per_simulation_time(
         Raises:
             ValueError: If `time_per_simulation` is non-positive.
             ValueError: If `initial_time` is non-positive.
-
-        Examples:
         """
         if (total_seconds := time_per_simulation.total_seconds()) <= 0.0:
             raise ValueError(
@@ -672,6 +746,9 @@ def _click_batch(ctx: click.Context = mock_context, **kwargs) -> None:
             "The `flepimop batch` CLI only supports EMCEE inference jobs."
         )
     inference_method = cfg["inference"]["method"].as_str()
+    inference_method = (
+        inference_method if inference_method is None else inference_method.lower()
+    )
 
     # Job name/run id
     name = cfg["name"].get(str) if cfg["name"].exists() else None
@@ -719,6 +796,9 @@ def _click_batch(ctx: click.Context = mock_context, **kwargs) -> None:
     )
     logger.info("Setting a total job time limit of %s minutes", job_time_limit.format())
 
+    # Job resources
+    job_resources = JobResources.from_presets(job_size, inference_method)
+
     # Cluster info
     cluster: Cluster | None = None
     if batch_system == BatchSystem.SLURM:
@@ -756,7 +836,7 @@ def _click_batch(ctx: click.Context = mock_context, **kwargs) -> None:
         "flepi_path": kwargs["flepi_path"].absolute(),
         "job_name": job_name,
         "jobs": job_size.jobs,
-        "nslots": kwargs["nslots"],
+        "nslots": job_size.simulations,
         "nthin": kwargs["nthin"],
         "prefix": kwargs["prefix"],
         "project_path": kwargs["project_path"].absolute(),
@@ -766,12 +846,11 @@ def _click_batch(ctx: click.Context = mock_context, **kwargs) -> None:
     options = {
         "chdir": kwargs["project_path"].absolute(),
         "comment": f"Generated on {now:%c %Z} and submitted by {getuser()}.",
-        "cpus-per-task": job_size.jobs,
+        "cpus-per-task": job_resources.cpus,
         "job-name": job_name,
-        "mem": "100GB",
-        "nodes": 1,  # EMCEE can only run on one node for now.
-        "ntasks": 1,
-        "partition": "general",
+        "mem": job_resources.memory,
+        "nodes": job_resources.nodes,
+        "ntasks-per-node": 1,
         "time": job_time_limit.format("slurm"),
     }
     if kwargs["partition"] is not None:

flepimop/gempyor_pkg/tests/batch/test_job_resources_class.py

@@ -0,0 +1,31 @@
+from typing import Literal
+
+import pytest
+
+from gempyor.batch import JobResources
+
+
+@pytest.mark.parametrize(
+    "kwargs",
+    (
+        {"nodes": 0, "cpus": 1, "memory": 1},
+        {"nodes": 1, "cpus": 0, "memory": 1},
+        {"nodes": 1, "cpus": 1, "memory": 0},
+        {"nodes": 0, "cpus": 0, "memory": 1},
+        {"nodes": 1, "cpus": 0, "memory": 0},
+        {"nodes": 0, "cpus": 1, "memory": 0},
+        {"nodes": 0, "cpus": 0, "memory": 0},
+    ),
+)
+def test_less_than_one_value_error(
+    kwargs: dict[Literal["nodes", "cpus", "memory"], int]
+) -> None:
+    param = next(k for k, v in kwargs.items() if v < 1)
+    with pytest.raises(
+        ValueError,
+        match=(
+            f"^The '{param}' attribute must be greater than 0, "
+            f"but instead was given '{kwargs.get(param)}'.$"
+        ),
+    ):
+        JobResources(**kwargs)