Interface Merge for large files

Author: oesteban

nipype/algorithms/misc.py

@@ -1249,10 +1249,24 @@ def merge_rois(in_files, in_idxs, in_ref,
     import nibabel as nb
     import numpy as np
     import os.path as op
+    import subprocess as sp
 
     if out_file is None:
         out_file = op.abspath('merged.nii.gz')
 
+    if dtype is None:
+        dtype = np.float32
+
+    # if file is compressed, uncompress using os
+    # to avoid memory errors
+    if op.splitext(in_ref)[1] == '.gz':
+        try:
+            iflogger.info('uncompress %i' % in_ref)
+            sp.check_call(['gunzip', in_ref], stdout=sp.PIPE, shell=True)
+            in_ref = op.splitext(in_ref)[0]
+        except:
+            pass
+
     ref = nb.load(in_ref)
     aff = ref.get_affine()
     hdr = ref.get_header().copy()
@@ -1261,23 +1275,47 @@ def merge_rois(in_files, in_idxs, in_ref,
     npix = rsh[0] * rsh[1] * rsh[2]
     ndirs = nb.load(in_files[0]).get_shape()[-1]
     newshape = (rsh[0], rsh[1], rsh[2], ndirs)
-    data = np.zeros((npix, ndirs), dtype=dtype)
-    for cname, iname in zip(in_files, in_idxs):
-        with np.load(iname) as f:
-            idxs = np.squeeze(f['arr_0'])
-        cdata = nb.load(cname).get_data().reshape(-1, ndirs)
-        nels = len(idxs)
-        idata = (idxs, )
-        data[idata, ...] = cdata[0:nels, ...]
-
-    if dtype is None:
-        dtype = np.float32
-
     hdr.set_data_dtype(dtype)
     hdr.set_xyzt_units('mm', 'sec')
-    hdr.set_data_shape(newshape)
-    nb.Nifti1Image(data.reshape(newshape).astype(dtype),
-                   aff, hdr).to_filename(out_file)
+ 
+    if ndirs < 300:
+        data = np.zeros((npix, ndirs))
+        for cname, iname in zip(in_files, in_idxs):
+            with np.load(iname) as f:
+                idxs = np.squeeze(f['arr_0'])
+            cdata = nb.load(cname).get_data().reshape(-1, ndirs)
+            nels = len(idxs)
+            idata = (idxs, )
+            data[idata, ...] = cdata[0:nels, ...]
+        hdr.set_data_shape(newshape)
+
+        nb.Nifti1Image(data.reshape(newshape).astype(dtype),
+                       aff, hdr).to_filename(out_file)
+
+
+    else:
+        hdr.set_data_shape(rsh[:3])
+        nii = []
+        for d in xrange(ndirs):
+            fname = op.abspath('vol%06d.nii' % d)
+            nb.Nifti1Image(np.zeros(rsh[:3]), aff, hdr).to_filename(fname)
+            nii.append(fname)
+
+        for cname, iname in zip(in_files, in_idxs):
+            with np.load(iname) as f:
+                idxs = np.squeeze(f['arr_0'])
+
+            for d, fname in enumerate(nii):
+                data = nb.load(fname).get_data().reshape(-1)
+                cdata = nb.load(cname).get_data().reshape(-1, ndirs)[:, d]
+                nels = len(idxs)
+                idata = (idxs, )
+                data[idata] = cdata[0:nels]
+                nb.Nifti1Image(data.reshape(rsh[:3]), aff, hdr).to_filename(fname)
+
+        imgs = [nb.load(im) for im in nii]
+        allim = nb.concat_images(imgs)
+        allim.to_filename(out_file)
 
     return out_file