renamed output_dir to cache_dir

Author: tclose

docs/source/tutorial/2-advanced-execution.ipynb

@@ -238,7 +238,7 @@
     "    result2 = submitter(mrgrid_varying_vox2)\n",
     "\n",
     "# Check that the output directory is the same for both runs\n",
-    "assert result2.output_dir == result1.output_dir\n",
+    "assert result2.cache_dir == result1.cache_dir\n",
     "\n",
     "# Change the voxel sizes to resample the NIfTI files to for one of the files\n",
     "mrgrid_varying_vox2.voxel[2] = [0.25]\n",
@@ -248,7 +248,7 @@
     "    result3 = submitter(mrgrid_varying_vox2)\n",
     "\n",
     "# The output directory will be different as the inputs are now different\n",
-    "assert result3.output_dir != result1.output_dir"
+    "assert result3.cache_dir != result1.cache_dir"
    ]
   },
   {
@@ -285,7 +285,7 @@
     "with submitter:\n",
     "    result4 = submitter(mrgrid_varying_vox3)\n",
     "\n",
-    "assert result4.output_dir == result1.output_dir\n",
+    "assert result4.cache_dir == result1.cache_dir\n",
     "\n",
     "# Replace the first NIfTI file with a new file\n",
     "nifti_files[0] = Nifti1.sample(nifti_dir, seed=100)\n",
@@ -302,7 +302,7 @@
     "    result4 = submitter(mrgrid_varying_vox4)\n",
     "\n",
     "# The cache directory for the new run is different\n",
-    "assert result4.output_dir != result1.output_dir"
+    "assert result4.cache_dir != result1.cache_dir"
    ]
   },
   {
@@ -380,7 +380,7 @@
     "\n",
     "def notify_task_completion(task: Job, result: Result):\n",
     "    # Print a message to the terminal\n",
-    "    print(f\"Job completed! Results are stored in {str(task.output_dir)!r}\")\n",
+    "    print(f\"Job completed! Results are stored in {str(task.cache_dir)!r}\")\n",
     "\n",
     "    # Platform-specific notifications\n",
     "    if platform.system() == \"Darwin\":  # macOS\n",

docs/source/tutorial/3-troubleshooting.ipynb

@@ -205,7 +205,7 @@
     "    with Submitter(worker=\"cf\") as sub:\n",
     "        result = sub(wf)\n",
     "\n",
-    "print(f\"Workflow completed successfully, results saved in: {result.output_dir}\")"
+    "print(f\"Workflow completed successfully, results saved in: {result.cache_dir}\")"
    ]
   },
   {

pydra/compose/base/task.py

@@ -70,7 +70,7 @@ def _from_task(cls, job: "Job[TaskType]") -> Self:
                 default = output.default
             defaults[output.name] = default
         outputs = cls(**defaults)
-        outputs._output_dir = job.output_dir
+        outputs._output_dir = job.cache_dir
         return outputs
 
     @property
@@ -256,7 +256,7 @@ def __call__(
                 f"Job {self} failed @ {time_of_crash} with the "
                 f"following errors:\n{error_message}\n"
                 "To inspect, please load the pickled job object from here: "
-                f"{result.output_dir}/_job.pklz"
+                f"{result.cache_dir}/_job.pklz"
             )
         return result.outputs
 

pydra/compose/shell/task.py

@@ -90,7 +90,7 @@ def _from_task(cls, job: "Job[Task]") -> ty.Self:
             ):
                 resolved_value = job.inputs[fld.name]
             elif is_set(fld.default):
-                resolved_value = cls._resolve_default_value(fld, job.output_dir)
+                resolved_value = cls._resolve_default_value(fld, job.cache_dir)
             else:
                 resolved_value = cls._resolve_value(fld, job)
             # Set the resolved value
@@ -174,7 +174,7 @@ def _resolve_value(
             return template_update_single(
                 fld,
                 task=job.task,
-                output_dir=job.output_dir,
+                output_dir=job.cache_dir,
                 spec_type="output",
             )
         assert fld.callable, (
@@ -192,7 +192,7 @@ def _resolve_value(
             if argnm == "field":
                 call_args_val[argnm] = fld
             elif argnm == "output_dir":
-                call_args_val[argnm] = job.output_dir
+                call_args_val[argnm] = job.cache_dir
             elif argnm == "executable":
                 call_args_val[argnm] = job.task.executable
             elif argnm == "inputs":

pydra/compose/tests/test_workflow_run.py

@@ -2251,7 +2251,7 @@ def Worky2(x, y):
     assert 2 == results2.outputs.out
 
     # checking if both output_dirs are created
-    assert results1.output_dir != results2.output_dir
+    assert results1.cache_dir != results2.cache_dir
 
 
 @pytest.mark.flaky(reruns=3)
@@ -2311,7 +2311,7 @@ def Worky2(x, y):
         assert t2 < max(1, t1 - 1)
 
     # checking if the second worky didn't run again
-    assert results1.output_dir != results2.output_dir
+    assert results1.cache_dir != results2.cache_dir
 
 
 @pytest.mark.flaky(reruns=3)
@@ -2371,7 +2371,7 @@ def Worky2(x, y):
         assert t2 < max(1, t1 - 1)
 
     # both worky output_dirs should be created
-    assert results1.output_dir != results2.output_dir
+    assert results1.cache_dir != results2.cache_dir
 
 
 @pytest.mark.flaky(reruns=3)
@@ -2427,7 +2427,7 @@ def Worky2(x, y):
         assert t2 < max(1, t1 - 1)
 
     # both worky output_dirs should be created
-    assert results1.output_dir != results2.output_dir
+    assert results1.cache_dir != results2.cache_dir
 
 
 @pytest.mark.flaky(reruns=3)
@@ -2484,7 +2484,7 @@ def Worky2(x, y):
         assert t2 > 2
 
     # checking if the second worky didn't run again
-    assert results1.output_dir != results2.output_dir
+    assert results1.cache_dir != results2.cache_dir
 
 
 @pytest.mark.flaky(reruns=3)
@@ -2538,7 +2538,7 @@ def Worky2(x, y):
     assert 8 == results2.outputs.out
 
     # checking if the second worky runs again
-    assert results1.output_dir != results2.output_dir
+    assert results1.cache_dir != results2.cache_dir
 
     # everything has to be recomputed
     assert len(list(Path(cache_root1).glob("python-*"))) == 2
@@ -2605,7 +2605,7 @@ def Worky2(x, y):
     assert 8 == results2.outputs.out
 
     # checking if the second worky runs again
-    assert results1.output_dir != results2.output_dir
+    assert results1.cache_dir != results2.cache_dir
 
     # for win and dask/slurm the time for dir creation etc. might take much longer
     if not sys.platform.startswith("win") and worker == "cf":
@@ -2676,7 +2676,7 @@ def Worky2(x, y):
 
     assert 8 == results2.outputs.out
 
-    assert results1.output_dir != results2.output_dir
+    assert results1.cache_dir != results2.cache_dir
     # the second task should be recomputed
     assert len(list(Path(cache_root1).glob("python-*"))) == 2
     assert len(list(Path(cache_root2).glob("python-*"))) == 1
@@ -2734,7 +2734,7 @@ def Worky2(x, y=None):
     assert 12 == results2.outputs.out
 
     # checking if the second worky runs again, but runs only one task
-    assert results1.output_dir != results2.output_dir
+    assert results1.cache_dir != results2.cache_dir
     # the second worky should rerun one task
     assert len(list(Path(cache_root1).glob("python-*"))) == 2
     assert len(list(Path(cache_root2).glob("python-*"))) == 1
@@ -2785,7 +2785,7 @@ def Worky2(x, y=None):
     assert 12 == results2.outputs.out
 
     # checking if the second worky runs again, but runs only one task
-    assert results1.output_dir != results2.output_dir
+    assert results1.cache_dir != results2.cache_dir
     # the second worky should have only one task run
     assert len(list(Path(cache_root1).glob("python-*"))) == 2
     assert len(list(Path(cache_root2).glob("python-*"))) == 1
@@ -2851,7 +2851,7 @@ def Worky2(x, y):
     # checking if the second worky didn't run again
     # checking all directories
 
-    assert results1.output_dir == results2.output_dir
+    assert results1.cache_dir == results2.cache_dir
 
 
 @pytest.mark.flaky(reruns=3)
@@ -3083,7 +3083,7 @@ def Worky2(x, y):
         assert t2 > 2
 
     # checking if both worky run
-    assert results1.output_dir != results2.output_dir
+    assert results1.cache_dir != results2.cache_dir
 
 
 @pytest.mark.flaky(reruns=3)
@@ -3133,7 +3133,7 @@ def Worky2(x, y):
     assert 8 == results2.outputs.out
 
     # checking if both dir exists
-    assert results1.output_dir != results2.output_dir
+    assert results1.cache_dir != results2.cache_dir
 
     # the second worky should have only one task run
     assert len(list(Path(cache_root1).glob("python-*"))) == 2
@@ -3449,7 +3449,7 @@ def Worky1(x, y):
     assert 602 == results1.outputs.out[1]
 
     # checkoing output_dir after the first run
-    assert results1.output_dir.exists()
+    assert results1.cache_dir.exists()
 
     # saving the content of the cache dit after the first run
     cache_root_content = os.listdir(results1.job.cache_root)

pydra/compose/workflow.py

@@ -320,7 +320,7 @@ def _from_task(cls, job: "Job[WorkflowTask]") -> ty.Self:
                         f"Job {node.name!r} failed @ {time_of_crash} running "
                         f"{node._task} with the following errors:\n{error_message}"
                         "\nTo inspect, please load the pickled job object from here: "
-                        f"{result.output_dir}/_job.pklz"
+                        f"{result.cache_dir}/_job.pklz"
                     )
             raise RuntimeError(
                 f"Workflow {job!r} failed with errors:\n\n" + "\n\n".join(errors)
@@ -338,7 +338,7 @@ def _from_task(cls, job: "Job[WorkflowTask]") -> ty.Self:
         # Set the values in the outputs object
         outputs = super()._from_task(job)
         outputs = attrs.evolve(outputs, **values)
-        outputs._output_dir = job.output_dir
+        outputs._output_dir = job.cache_dir
         return outputs
 
 

pydra/engine/job.py

@@ -204,7 +204,7 @@ def checksum(self):
 
     @property
     def lockfile(self):
-        return self.output_dir.with_suffix(".lock")
+        return self.cache_dir.with_suffix(".lock")
 
     @property
     def uid(self):
@@ -225,7 +225,7 @@ def can_resume(self):
         return self._can_resume
 
     @property
-    def output_dir(self):
+    def cache_dir(self):
         """Get the filesystem path where outputs will be written."""
         return self.cache_root / self.checksum
 
@@ -258,7 +258,7 @@ def inputs(self) -> dict[str, ty.Any]:
             if value and TypeParser.contains_type(FileSet, fld.type):
                 copied_value = copy_nested_files(
                     value=value,
-                    dest_dir=self.output_dir,
+                    dest_dir=self.cache_dir,
                     mode=fld.copy_mode,
                     collation=fld.copy_collation,
                     supported_modes=self.SUPPORTED_COPY_MODES,
@@ -267,7 +267,7 @@ def inputs(self) -> dict[str, ty.Any]:
                     map_copyfiles[name] = copied_value
         self._inputs.update(
             template_update(
-                self.task, output_dir=self.output_dir, map_copyfiles=map_copyfiles
+                self.task, output_dir=self.cache_dir, map_copyfiles=map_copyfiles
             )
         )
         return self._inputs
@@ -285,11 +285,11 @@ def _populate_filesystem(self):
         # and the lockfile has to be removed
         with open(self.cache_root / f"{self.uid}_info.json", "w") as jsonfile:
             json.dump({"checksum": self.checksum}, jsonfile)
-        if not self.can_resume and self.output_dir.exists():
-            shutil.rmtree(self.output_dir)
-        self.output_dir.mkdir(parents=False, exist_ok=self.can_resume)
+        if not self.can_resume and self.cache_dir.exists():
+            shutil.rmtree(self.cache_dir)
+        self.cache_dir.mkdir(parents=False, exist_ok=self.can_resume)
         # Save job pkl into the output directory for future reference
-        save(self.output_dir, job=self)
+        save(self.cache_dir, job=self)
 
     def run(self, rerun: bool = False):
         """Prepare the job working directory, execute the task, and save the
@@ -316,16 +316,16 @@ def run(self, rerun: bool = False):
                     return result
             cwd = os.getcwd()
             self._populate_filesystem()
-            os.chdir(self.output_dir)
+            os.chdir(self.cache_dir)
             result = Result(
                 outputs=None,
                 runtime=None,
                 errored=False,
-                output_dir=self.output_dir,
+                output_dir=self.cache_dir,
                 task=self.task,
             )
             self.hooks.pre_run_task(self)
-            self.audit.start_audit(odir=self.output_dir)
+            self.audit.start_audit(odir=self.cache_dir)
             if self.audit.audit_check(AuditFlag.PROV):
                 self.audit.audit_task(job=self)
             try:
@@ -335,13 +335,13 @@ def run(self, rerun: bool = False):
             except Exception:
                 etype, eval, etr = sys.exc_info()
                 traceback = format_exception(etype, eval, etr)
-                record_error(self.output_dir, error=traceback)
+                record_error(self.cache_dir, error=traceback)
                 result.errored = True
                 raise
             finally:
                 self.hooks.post_run_task(self, result)
                 self.audit.finalize_audit(result=result)
-                save(self.output_dir, result=result, job=self)
+                save(self.cache_dir, result=result, job=self)
                 # removing the additional file with the checksum
                 (self.cache_root / f"{self.uid}_info.json").unlink()
                 os.chdir(cwd)
@@ -377,26 +377,26 @@ async def run_async(self, rerun: bool = False) -> Result:
                 outputs=None,
                 runtime=None,
                 errored=False,
-                output_dir=self.output_dir,
+                output_dir=self.cache_dir,
                 task=self.task,
             )
             self.hooks.pre_run_task(self)
-            self.audit.start_audit(odir=self.output_dir)
+            self.audit.start_audit(odir=self.cache_dir)
             try:
                 self.audit.monitor()
                 await self.task._run_async(self, rerun)
                 result.outputs = self.task.Outputs._from_task(self)
             except Exception:
                 etype, eval, etr = sys.exc_info()
                 traceback = format_exception(etype, eval, etr)
-                record_error(self.output_dir, error=traceback)
+                record_error(self.cache_dir, error=traceback)
                 result.errored = True
                 self._errored = True
                 raise
             finally:
                 self.hooks.post_run_task(self, result)
                 self.audit.finalize_audit(result=result)
-                save(self.output_dir, result=result, job=self)
+                save(self.cache_dir, result=result, job=self)
                 # removing the additional file with the checksum
                 (self.cache_root / f"{self.uid}_info.json").unlink()
                 os.chdir(cwd)
@@ -483,7 +483,7 @@ def result(self, return_inputs=False):
                 outputs=None,
                 runtime=None,
                 errored=True,
-                output_dir=self.output_dir,
+                output_dir=self.cache_dir,
                 task=self.task,
             )
 
@@ -614,22 +614,22 @@ def load_and_run(job_pkl: Path, rerun: bool = False) -> Path:
             save(job_pkl.parent, result=result)
         raise
 
-    resultfile = job.output_dir / "_result.pklz"
+    resultfile = job.cache_dir / "_result.pklz"
     try:
         if job.is_async:
             job.submitter.submit(job, rerun=rerun)
         else:
             job.run(rerun=rerun)
     except Exception as e:
         # creating result and error files if missing
-        errorfile = job.output_dir / "_error.pklz"
+        errorfile = job.cache_dir / "_error.pklz"
         if not errorfile.exists():  # not sure if this is needed
             etype, eval, etr = sys.exc_info()
             traceback = format_exception(etype, eval, etr)
-            errorfile = record_error(job.output_dir, error=traceback)
+            errorfile = record_error(job.cache_dir, error=traceback)
         if not resultfile.exists():  # not sure if this is needed
             result = Result(output=None, runtime=None, errored=True, task=None)
-            save(job.output_dir, result=result)
+            save(job.cache_dir, result=result)
         e.add_note(f" full crash report is here: {errorfile}")
         raise
     return resultfile