WIP/ENH option to auto delete full tractogram in export all

AFQ/api/group.py

@@ -690,7 +690,14 @@ def export_up_to(self, attr_name="help"):
         for inputs in calcdata.calcs[idx].inputs:
             self.export(inputs)
 
-    def export_all(self, viz=True, afqbrowser=True, xforms=True, indiv=True):
+    def export_all(
+        self,
+        viz=True,
+        afqbrowser=True,
+        xforms=True,
+        indiv=True,
+        delete_full_tractogram=False,
+    ):
         """Exports all the possible outputs
 
         Parameters
@@ -714,6 +721,14 @@ def export_all(self, viz=True, afqbrowser=True, xforms=True, indiv=True):
             the AFQ segmentation algorithm, individual ROIs are also
             output.
             Default: True
+        delete_full_tractogram : bool
+            Whether to delete the full, unsegmented tractogram (and its
+            associated json) after all other outputs have been generated.
+            This can be useful for saving disk space, as the full
+            tractogram is typically over 90 percent of the total output size.
+            Only the full tractogram and its json are deleted; all other
+            outputs are left in place.
+            Default: False
         """
         start_time = time()
 
@@ -722,8 +737,19 @@ def export_all(self, viz=True, afqbrowser=True, xforms=True, indiv=True):
         self.combine_profiles()
         if afqbrowser:
             self.assemble_AFQ_browser()
+        if delete_full_tractogram:
+            self.delete_full_tractogram()
         self.logger.info(f"Time taken for export all: {str(time() - start_time)}")
 
+    def delete_full_tractogram(self):
+        """
+        Delete the full, unsegmented tractogram (and its associated json)
+        for every subject and session. Only the full tractogram and its
+        json are deleted; all other outputs are left in place.
+        """
+        for pAFQ in self.pAFQ_list:
+            pAFQ.delete_full_tractogram()
+
     def cmd_outputs(
         self, cmd="rm", dependent_on=None, up_to=None, exceptions=None, suffix=""
     ):

AFQ/api/participant.py

@@ -1,5 +1,6 @@
 import logging
 import math
+import os
 import os.path as op
 import tempfile
 from math import radians
@@ -30,7 +31,7 @@
 from AFQ.tasks.utils import get_base_fname
 from AFQ.tasks.viz import get_viz_plan
 from AFQ.utils.bin import pyafq_str_to_val
-from AFQ.utils.path import apply_cmd_to_afq_derivs
+from AFQ.utils.path import apply_cmd_to_afq_derivs, drop_extension
 from AFQ.viz.utils import BEST_BUNDLE_ORIENTATIONS, get_eye, trim
 
 __all__ = ["ParticipantAFQ"]
@@ -240,7 +241,9 @@ def export_up_to(self, attr_name="help"):
         for inputs in calcdata.calcs[idx].inputs:
             self.export(inputs)
 
-    def export_all(self, viz=True, xforms=True, indiv=True):
+    def export_all(
+        self, viz=True, xforms=True, indiv=True, delete_full_tractogram=False
+    ):
         f""" Exports all the possible outputs
         {valid_exports_string}
 
@@ -262,11 +265,33 @@ def export_all(self, viz=True, xforms=True, indiv=True):
             the AFQ segmentation algorithm, individual ROIs are also
             output.
             Default: True
+        delete_full_tractogram : bool
+            Whether to delete the full, unsegmented tractogram (and its
+            associated json) after all other outputs have been generated.
+            This can be useful for saving disk space, as the full
+            tractogram is typically over 90 percent of the total output size.
+            Only the full tractogram and its json are deleted;
+            all other outputs are left in place.
+            Default: False
         """
         start_time = time()
         export_all_helper(self, xforms, indiv, viz)
+        if delete_full_tractogram:
+            self.delete_full_tractogram()
         self.logger.info(f"Time taken for export all: {time() - start_time}")
 
+    def delete_full_tractogram(self):
+        """
+        Delete the full, unsegmented tractogram (and its associated json).
+        Only the full tractogram and its json are deleted; all other
+        outputs are left in place.
+        """
+        tractogram_file = self.export("streamlines")
+        for fname in (tractogram_file, drop_extension(tractogram_file) + ".json"):
+            if op.exists(fname):
+                os.remove(fname)
+                self.logger.info(f"Deleted {fname}")
+
     def participant_montage(self, images_per_row=3, anatomy=True, bundle_names=None):
         """
         Generate montage of all bundles for a given subject.

AFQ/data/fetch.py

@@ -2307,6 +2307,9 @@ def fetch_hcp(
         data_files[
             op.join(sess_dir, "anat", f"sub-{subject}_aparc+aseg_seg.nii.gz")
         ] = f"{study}/{subject}/T1w/aparc+aseg.nii.gz"
+        data_files[
+            op.join(sess_dir, "anat", f"sub-{subject}_desc-PVE_probseg.nii.gz")
+        ] = f"{study}/{subject}/T1w/wmparc.nii.gz"
 
     download_files = {}
     for k in data_files.keys():