FIX: Fetch templates during workflow construction

smriprep/interfaces/templateflow.py

@@ -108,34 +108,9 @@ def _run_interface(self, runtime):
         if isdefined(self.inputs.cohort):
             specs['cohort'] = self.inputs.cohort
 
-        name = self.inputs.template.strip(':').split(':', 1)
-        if len(name) > 1:
-            specs.update(
-                {
-                    k: v
-                    for modifier in name[1].split(':')
-                    for k, v in [tuple(modifier.split('-'))]
-                    if k not in specs
-                }
-            )
-
-        if specs['resolution'] and not isinstance(specs['resolution'], list):
-            specs['resolution'] = [specs['resolution']]
-
-        available_resolutions = tf.TF_LAYOUT.get_resolutions(template=name[0])
-        if specs['resolution'] and not set(specs['resolution']) & set(available_resolutions):
-            fallback_res = available_resolutions[0] if available_resolutions else None
-            LOGGER.warning(
-                f"Template {name[0]} does not have resolution(s): {specs['resolution']}."
-                f"Falling back to resolution: {fallback_res}."
-            )
-            specs['resolution'] = fallback_res
-
-        self._results['t1w_file'] = tf.get(name[0], desc=None, suffix='T1w', **specs)
-
-        self._results['brain_mask'] = tf.get(
-            name[0], desc='brain', suffix='mask', **specs
-        ) or tf.get(name[0], label='brain', suffix='mask', **specs)
+        files = fetch_template_files(self.inputs.template, specs)
+        self._results['t1w_file'] = files['t1w']
+        self._results['brain_mask'] = files['mask']
         return runtime
 
 
@@ -186,3 +161,49 @@ def _run_interface(self, runtime):
                 descsplit = desc.split('-')
                 self._results['spec'][descsplit[0]] = descsplit[1]
         return runtime
+
+
+def fetch_template_files(
+    template: str,
+    specs: dict | None = None,
+    sloppy: bool = False,
+) -> dict:
+    if specs is None:
+        specs = {}
+
+    name = template.strip(':').split(':', 1)
+    if len(name) > 1:
+        specs.update(
+            {
+                k: v
+                for modifier in name[1].split(':')
+                for k, v in [tuple(modifier.split('-'))]
+                if k not in specs
+            }
+        )
+
+    if res := specs.pop('res', None):
+        if res != 'native':
+            specs['resolution'] = res
+
+    if not specs.get('resolution'):
+        specs['resolution'] = 2 if sloppy else 1
+
+    if specs.get('resolution') and not isinstance(specs['resolution'], list):
+        specs['resolution'] = [specs['resolution']]
+
+    available_resolutions = tf.TF_LAYOUT.get_resolutions(template=name[0])
+    if specs.get('resolution') and not set(specs['resolution']) & set(available_resolutions):
+        fallback_res = available_resolutions[0] if available_resolutions else None
+        LOGGER.warning(
+            f"Template {name[0]} does not have resolution(s): {specs['resolution']}."
+            f"Falling back to resolution: {fallback_res}."
+        )
+        specs['resolution'] = fallback_res
+
+    files = {}
+    files['t1w'] = tf.get(name[0], desc=None, suffix='T1w', **specs)
+    files['mask'] = tf.get(name[0], desc='brain', suffix='mask', **specs) or tf.get(
+        name[0], label='brain', suffix='mask', **specs
+    )
+    return files

smriprep/workflows/anatomical.py

@@ -277,7 +277,7 @@ def init_anat_preproc_wf(
         omp_nthreads=omp_nthreads,
         skull_strip_fixed_seed=skull_strip_fixed_seed,
     )
-    template_iterator_wf = init_template_iterator_wf(spaces=spaces)
+    template_iterator_wf = init_template_iterator_wf(spaces=spaces, sloppy=sloppy)
     ds_std_volumes_wf = init_ds_anat_volumes_wf(
         bids_root=bids_root,
         output_dir=output_dir,
@@ -725,6 +725,7 @@ def init_anat_fit_wf(
         spaces=spaces,
         freesurfer=freesurfer,
         output_dir=output_dir,
+        sloppy=sloppy,
     )
     # fmt:off
     workflow.connect([

smriprep/workflows/outputs.py

@@ -33,12 +33,15 @@
 from niworkflows.interfaces.utility import KeySelect
 
 from ..interfaces import DerivativesDataSink
-from ..interfaces.templateflow import TemplateFlowSelect
+from ..interfaces.templateflow import TemplateFlowSelect, fetch_template_files
+
+if ty.TYPE_CHECKING:
+    from niworkflows.utils.spaces import SpatialReferences
 
 BIDS_TISSUE_ORDER = ('GM', 'WM', 'CSF')
 
 
-def init_anat_reports_wf(*, spaces, freesurfer, output_dir, name='anat_reports_wf'):
+def init_anat_reports_wf(*, spaces, freesurfer, output_dir, sloppy=False, name='anat_reports_wf'):
     """
     Set up a battery of datasinks to store reports in the right location.
 
@@ -131,7 +134,7 @@ def init_anat_reports_wf(*, spaces, freesurfer, output_dir, name='anat_reports_w
     # fmt:on
 
     if spaces._cached is not None and spaces.cached.references:
-        template_iterator_wf = init_template_iterator_wf(spaces=spaces)
+        template_iterator_wf = init_template_iterator_wf(spaces=spaces, sloppy=sloppy)
         t1w_std = pe.Node(
             ApplyTransforms(
                 dimension=3,
@@ -1112,7 +1115,9 @@ def init_anat_second_derivatives_wf(
     return workflow
 
 
-def init_template_iterator_wf(*, spaces, name='template_iterator_wf'):
+def init_template_iterator_wf(
+    *, spaces: 'SpatialReferences', sloppy: bool = False, name='template_iterator_wf'
+):
     """Prepare the necessary components to resample an image to a template space
 
     This produces a workflow with an unjoined iterable named "spacesource".
@@ -1122,6 +1127,9 @@ def init_template_iterator_wf(*, spaces, name='template_iterator_wf'):
 
     The fields in `outputnode` can be used as if they come from a single template.
     """
+    for template in spaces.get_spaces(nonstandard=False, dim=(3,)):
+        fetch_template_files(template, specs=None, sloppy=sloppy)
+
     workflow = pe.Workflow(name=name)
 
     inputnode = pe.Node(
@@ -1159,9 +1167,7 @@ def init_template_iterator_wf(*, spaces, name='template_iterator_wf'):
         name='select_xfm',
         run_without_submitting=True,
     )
-    select_tpl = pe.Node(
-        TemplateFlowSelect(resolution=1), name='select_tpl', run_without_submitting=True
-    )
+    select_tpl = pe.Node(TemplateFlowSelect(), name='select_tpl', run_without_submitting=True)
 
     # fmt:off
     workflow.connect([
@@ -1177,7 +1183,7 @@ def init_template_iterator_wf(*, spaces, name='template_iterator_wf'):
         (spacesource, select_tpl, [
             ('space', 'template'),
             ('cohort', 'cohort'),
-            (('resolution', _no_native), 'resolution'),
+            (('resolution', _no_native, sloppy), 'resolution'),
         ]),
         (spacesource, outputnode, [
             ('space', 'space'),
@@ -1243,10 +1249,6 @@ def _pick_cohort(in_template):
     return [_pick_cohort(v) for v in in_template]
 
 
-def _fmt(in_template):
-    return in_template.replace(':', '_')
-
-
 def _empty_report(in_file=None):
     from pathlib import Path
 
@@ -1268,11 +1270,11 @@ def _is_native(value):
     return value == 'native'
 
 
-def _no_native(value):
+def _no_native(value, sloppy=False):
     try:
         return int(value)
     except (TypeError, ValueError):
-        return 1
+        return 2 if sloppy else 1
 
 
 def _drop_path(in_path):