snakemake/settings/types.py (1)

205-205: LGTM! Consider adding a docstring.

The new boolean flag is well-placed within the DAGSettings class and aligns with the PR's objective of optimizing temporary file handling in checkpoint jobs. However, adding a docstring would improve code maintainability.

Consider adding a docstring to document the flag's purpose:

 class DAGSettings(SettingsBase):
     targets: AnySet[str] = frozenset()
     target_jobs: AnySet[str] = frozenset()
     target_files_omit_workdir_adjustment: bool = False
     batch: Optional[Batch] = None
     forcetargets: bool = False
     forceall: bool = False
     forcerun: AnySet[str] = frozenset()
     until: AnySet[str] = frozenset()
     omit_from: AnySet[str] = frozenset()
     force_incomplete: bool = False
     allowed_rules: AnySet[str] = frozenset()
     rerun_triggers: AnySet[RerunTrigger] = RerunTrigger.all()
     max_inventory_wait_time: int = 20
-    assume_checkpoint_safe_temp_files: bool = False
+    assume_checkpoint_safe_temp_files: bool = False
+    """When True, allows Snakemake to remove temporary files that are not required by current jobs,
+    even if there are pending checkpoint jobs. This can significantly reduce storage requirements
+    in workflows with checkpoint rules."""

docs/snakefiles/rules.rst (1)

1659-1666: Documentation enhancement looks good, with a minor suggestion.

The added documentation clearly explains the behavior of temporary files with checkpoint rules and introduces the new --assume-checkpoint-safe-temp-files option. The warning about potential premature deletion is well-articulated.

Consider adding a small code example to illustrate the proper way to include temporary files as inputs to checkpoint jobs or their dependencies. This would make the documentation more actionable for users.

Example addition:

# Example of properly including temp files in checkpoint dependencies
rule create_temp:
    output:
        temp("temp_file.txt")
    shell:
        "..."

checkpoint process_data:
    input:
        "temp_file.txt"  # temp file properly included as input
    output:
        "checkpoint_output.txt"
    shell:
        "..."

snakemake/dag.py (1)

1979-1982: Implement Conditional Temp File Deletion and Address TODO

The condition now allows temp files to be deleted when assume_checkpoint_safe_temp_files is True, which is correct. There's a TODO comment suggesting a more accurate method to determine dependencies on temp files in relation to checkpoints. I can assist with this enhancement.

Would you like me to work on a solution or open a new GitHub issue to address this TODO?

Configuration used: CodeRabbit UI
Review profile: CHILL

docs/snakefiles/rules.rst (1)

1659-1666: Documentation enhancement looks good with a minor formatting suggestion.

The added documentation clearly explains:

• The limitation that temporary files cannot be deleted until all checkpoint jobs complete
• The new --assume-checkpoint-safe-temp-files option to override this behavior
• Important safety considerations when using this option

Consider applying this formatting improvement to make the documentation more readable:

-When using checkpoint rules, Snakemake cannot determine whether temporary files are no longer needed until 
-all checkpoint jobs have been completed. This delay in removal may result in unexpectedly high storage usage.
-This behavior can be disabled by using the option ``--assume-checkpoint-safe-temp-files``. 
-However, if any temporary files might still be needed after future checkpoints, you must ensure they are not 
-deleted prematurely. To do so, include them as input of the checkpoint job (or any jobs dependent on the 
-checkpoint job)."
+When using checkpoint rules, Snakemake cannot determine whether temporary files are no longer needed until 
+all checkpoint jobs have been completed. This delay in removal may result in unexpectedly high storage usage.
+
+This behavior can be disabled by using the option ``--assume-checkpoint-safe-temp-files``. 
+However, if any temporary files might still be needed after future checkpoints, you must ensure they are not 
+deleted prematurely. To do so, include them as input of the checkpoint job (or any jobs dependent on the 
+checkpoint job).

Configuration used: CodeRabbit UI
Review profile: CHILL

docs/snakefiles/rules.rst (1)

1827-1833: Documentation Clarity and Formatting Improvement

The new paragraph effectively explains that when using checkpoint rules, temporary files cannot be removed until all checkpoint jobs have completed—which might lead to excessive storage usage—and that this behavior can be disabled with the new --assume-checkpoint-safe-temp-files option. However, there is a minor formatting issue: a stray closing double-quote at the end of the paragraph. Additionally, consider a slight rephrasing for enhanced clarity and consistency.

Suggested diff for polishing the text:

-When using checkpoint rules, Snakemake cannot determine whether temporary files are no longer needed until 
-all checkpoint jobs have been completed. This delay in removal may result in unexpectedly high storage usage.
-This behavior can be disabled by using the option ``--assume-checkpoint-safe-temp-files``. 
-However, if any temporary files might still be needed after future checkpoints, you must ensure they are not 
-deleted prematurely. To do so, include them as input of the checkpoint job (or any jobs dependent on the 
-checkpoint job)."
+When using checkpoint rules, Snakemake cannot determine whether temporary files are no longer needed until all checkpoint jobs have been completed, which may delay their removal and result in unexpectedly high storage usage. This behavior can be disabled by using the ``--assume-checkpoint-safe-temp-files`` option. However, if any temporary files might still be required after future checkpoints, ensure they are not deleted prematurely by including them as input to the checkpoint job (or to any jobs that depend on the checkpoint).

Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro