Merge pull request #1044 from dmwelch/csvReader

ENH: Added csvReader() utility

Author: satra

CHANGES

@@ -8,45 +8,46 @@ Next release
 * FIX: Fix split_filename behaviour when path has no file component (https://github.com/nipy/nipype/pull/1035)
 * ENH: Updated FSL dtifit to include option for grad non-linearities (https://github.com/nipy/nipype/pull/1032)
 * ENH: Updated Camino tracking interfaces, which can now use FSL bedpostx output.
-       New options also include choice of tracker, interpolator, stepsize and 
+       New options also include choice of tracker, interpolator, stepsize and
        curveinterval for angle threshold (https://github.com/nipy/nipype/pull/1029)
 * FIX: Interfaces redirecting X crashed if $DISPLAY not defined (https://github.com/nipy/nipype/pull/1027)
 * FIX: Bug crashed 'make api' (https://github.com/nipy/nipype/pull/1026)
 * ENH: Updated antsIntroduction to handle RA and RI registrations (https://github.com/nipy/nipype/pull/1009)
-* ENH: Updated N4BiasCorrection input spec to include weight image and spline order. Made 
+* ENH: Updated N4BiasCorrection input spec to include weight image and spline order. Made
 	   argument formatting consistent. Cleaned ants.segmentation according to PEP8.
 	   (https://github.com/nipy/nipype/pull/990/files)
 * ENH: SPM12 Normalize interface (https://github.com/nipy/nipype/pull/986)
 * FIX: Utility interface test dir (https://github.com/nipy/nipype/pull/986)
 * FIX: IPython engine directory reset after crash (https://github.com/nipy/nipype/pull/987)
 * ENH: Resting state fMRI example with NiPy realignment and no SPM (https://github.com/nipy/nipype/pull/992)
-* FIX: Corrected Freesurfer SegStats _list_outputs to avoid error if summary_file is 
+* FIX: Corrected Freesurfer SegStats _list_outputs to avoid error if summary_file is
 	   undefined (issue #994)(https://https://github.com/nipy/nipype/pull/996)
 * FIX: OpenfMRI support and FSL 5.0.7 changes (https://github.com/nipy/nipype/pull/1006)
 * FIX: Output prefix in SPM Normalize with modulation (https://github.com/nipy/nipype/pull/1023)
 * ENH: Usability improvements in cluster environments (https://github.com/nipy/nipype/pull/1025)
 * ENH: ANTs JointFusion() (https://github.com/nipy/nipype/pull/1042)
+* ENH: Added csvReader() utility (https://github.com/nipy/nipype/pull/1044)
 
 Release 0.10.0 (October 10, 2014)
 ============
 
 * ENH: New miscelaneous interfaces: SplitROIs (mapper), MergeROIs (reducer)
        to enable parallel processing of very large images.
 * ENH: Updated FSL interfaces: BEDPOSTX and XFibres, former interfaces are still
-       available with the version suffix: BEDPOSTX4 and XFibres4. Added gpu 
+       available with the version suffix: BEDPOSTX4 and XFibres4. Added gpu
        versions of BEDPOSTX: BEDPOSTXGPU, BEDPOSTX5GPU, and BEDPOSTX4GPU
 * ENH: Added experimental support for MIPAV algorithms thorugh JIST plugins
 * ENH: New dipy interfaces: Denoise, Resample
 * ENH: New Freesurfer interfaces: Tkregister2 (for conversion of fsl style matrices to freesurfer format), MRIPretess
 * ENH: New FSL interfaces: WarpPoints, WarpPointsToStd, EpiReg, ProbTrackX2, WarpUtils, ConvertWarp
 * ENH: New miscelaneous interfaces: AddCSVRow, NormalizeProbabilityMapSet, AddNoise
 * ENH: New AFNI interfaces:  Eval, Means, SVMTest, SVMTrain
-* ENH: FUGUE interface has been refactored to use the name_template system, 3 examples 
+* ENH: FUGUE interface has been refactored to use the name_template system, 3 examples
        added to doctests, some bugs solved.
 * API: Interfaces to external packages are no longer available in the top-level
        ``nipype`` namespace, and must be imported directly (e.g.
        ``from nipype.interfaces import fsl``).
-* ENH: Support for elastix via a set of new interfaces: Registration, ApplyWarp, 
+* ENH: Support for elastix via a set of new interfaces: Registration, ApplyWarp,
        AnalyzeWarp, PointsWarp, and EditTransform
 * ENH: New ANTs interface: ApplyTransformsToPoints, LaplacianThickness
 * ENH: New Diffusion Toolkit interface: TrackMerge
@@ -62,7 +63,7 @@ Release 0.10.0 (October 10, 2014)
 * ENH: New color mode for write_graph
 * ENH: You can now force MapNodes to be run serially
 * ENH: Added ANTS based openfmri workflow
-* ENH: MapNode now supports flattening of nested lists 
+* ENH: MapNode now supports flattening of nested lists
 * ENH: Support for headless mode using Xvfb
 * ENH: nipype_display_crash has a debugging mode
 * FIX: MRTrix tracking algorithms were ignoring mask parameters.

nipype/interfaces/tests/test_utility.py

@@ -2,7 +2,7 @@
 # vi: set ft=python sts=4 ts=4 sw=4 et:
 import os
 import shutil
-from tempfile import mkdtemp
+from tempfile import mkdtemp, mkstemp
 
 import numpy as np
 from nipype.testing import assert_equal, assert_true, assert_raises
@@ -138,3 +138,30 @@ def test_split():
     finally:
         os.chdir(origdir)
         shutil.rmtree(tempdir)
+
+
+def test_csvReader():
+    header = "files,labels,erosion\n"
+    lines = ["foo,hello,300.1\n",
+             "bar,world,5\n",
+             "baz,goodbye,0.3\n"]
+    for x in range(2):
+        fd, name = mkstemp(suffix=".csv")
+        with open(name, 'w+b') as fid:
+            reader = utility.CSVReader()
+            if x % 2 == 0:
+                fid.write(header)
+                reader.inputs.header = True
+            fid.writelines(lines)
+            fid.flush()
+            reader.inputs.in_file = name
+            out = reader.run()
+            if x % 2 == 0:
+                yield assert_equal, out.outputs.files, ['foo', 'bar', 'baz']
+                yield assert_equal, out.outputs.labels, ['hello', 'world', 'goodbye']
+                yield assert_equal, out.outputs.erosion, ['300.1', '5', '0.3']
+            else:
+                yield assert_equal, out.outputs.column_0, ['foo', 'bar', 'baz']
+                yield assert_equal, out.outputs.column_1, ['hello', 'world', 'goodbye']
+                yield assert_equal, out.outputs.column_2, ['300.1', '5', '0.3']
+        os.unlink(name)

nipype/interfaces/utility.py

@@ -479,3 +479,83 @@ def _run_interface(self, runtime):
         assert_equal(data1, data2)
 
         return runtime
+
+
+class CSVReaderInputSpec(DynamicTraitedSpec, TraitedSpec):
+    in_file = File(exists=True, mandatory=True, desc='Input comma-seperated value (CSV) file')
+    header = traits.Bool(False, usedefault=True, desc='True if the first line is a column header')
+
+
+class CSVReader(BaseInterface):
+    """
+    Examples
+    --------
+
+    >>> reader = CSVReader()  # doctest: +SKIP
+    >>> reader.inputs.in_file = 'noHeader.csv'  # doctest: +SKIP
+    >>> out = reader.run()  # doctest: +SKIP
+    >>> out.outputs.column_0 == ['foo', 'bar', 'baz']  # doctest: +SKIP
+    True
+    >>> out.outputs.column_1 == ['hello', 'world', 'goodbye']  # doctest: +SKIP
+    True
+    >>> out.outputs.column_2 == ['300.1', '5', '0.3']  # doctest: +SKIP
+    True
+
+    >>> reader = CSVReader()  # doctest: +SKIP
+    >>> reader.inputs.in_file = 'header.csv'  # doctest: +SKIP
+    >>> reader.inputs.header = True  # doctest: +SKIP
+    >>> out = reader.run()  # doctest: +SKIP
+    >>> out.outputs.files == ['foo', 'bar', 'baz']  # doctest: +SKIP
+    True
+    >>> out.outputs.labels == ['hello', 'world', 'goodbye']  # doctest: +SKIP
+    True
+    >>> out.outputs.erosion == ['300.1', '5', '0.3']  # doctest: +SKIP
+    True
+
+    """
+    input_spec = CSVReaderInputSpec
+    output_spec = DynamicTraitedSpec
+    _always_run = True
+
+    def _append_entry(self, outputs, entry):
+        for key, value in zip(self._outfields, entry):
+            outputs[key].append(value)
+        return outputs
+
+    def _parse_line(self, line):
+        line = line.replace('\n', '')
+        entry = [x.strip() for x in line.split(',')]
+        return entry
+
+    def _get_outfields(self):
+        with open(self.inputs.in_file, 'r') as fid:
+            entry = self._parse_line(fid.readline())
+            if self.inputs.header:
+                self._outfields = tuple(entry)
+            else:
+                self._outfields = tuple(['column_' + str(x) for x in range(len(entry))])
+        return self._outfields
+
+    def _run_interface(self, runtime):
+        self._get_outfields()
+        return runtime
+
+    def _outputs(self):
+        return self._add_output_traits(super(CSVReader, self)._outputs())
+
+    def _add_output_traits(self, base):
+        return add_traits(base, self._get_outfields())
+
+    def _list_outputs(self):
+        outputs = self.output_spec().get()
+        isHeader = True
+        for key in self._outfields:
+            outputs[key] = []  # initialize outfields
+        with open(self.inputs.in_file, 'r') as fid:
+            for line in fid.readlines():
+                if self.inputs.header and isHeader:  # skip header line
+                    isHeader = False
+                    continue
+                entry = self._parse_line(line)
+                outputs = self._append_entry(outputs, entry)
+        return outputs