ENH: Make the command-line tools usable in batch runs

scripts/cli_tract_math.py

@@ -33,6 +33,8 @@ def _build_arg_parser():
 
     optional flags:  Always start with "--" and are considered as boolean flags
 
+    The --force flag overwrites an existing output file without asking.
+
     The --use_file_names_as_index col_name1:path_dir_index col_name2:path_dir_index2 .. col_nameN:path_dir_indexN
         Example directory layout:
         =========================
@@ -86,6 +88,9 @@ def main():
 
     parser = _build_arg_parser()
     args = parser.parse_args()
+    # The files are only opened to check that they exist
+    for f in args.tractographies:
+        f.close()
 
     # Load the global modules after the parsing of parameters
     from tract_querier.tractography import tractography_from_files

scripts/cli_tract_querier.py

@@ -117,6 +117,22 @@ def main():
     ):
         parser.error("incorrect number of arguments")
 
+    input_split = os.path.splitext(options.tractography_file_name)
+    output_split = os.path.splitext(options.output_file_name)
+    if len(output_split[-1]) > 0:
+        options.output_file_name = output_split[0]
+        tractography_extension = output_split[1]
+    else:
+        tractography_extension = input_split[1]
+
+    if tractography_extension not in ('.trk', '.vtk', '.vtp'):
+        parser.error(
+            "output format %r not supported, it must be .trk, .vtk or .vtp; "
+            "add it to the output prefix, e.g. -o %s.trk"
+            % (tractography_extension, options.output_file_name +
+               tractography_extension)
+        )
+
     options.threshold = float(options.threshold)
     options.length_threshold = float(options.length_threshold)
 
@@ -208,20 +224,6 @@ def main():
         options.tractography_file_name
     )
 
-    tractography_extension = os.path.splitext(options.tractography_file_name)[-1]
-    if tractography_extension == '.trk':
-        tractography_extra_kwargs = {
-            'affine': tr.affine,
-            'image_dimensions': tr.image_dims
-        }
-    input_split = os.path.splitext(options.tractography_file_name)
-    output_split = os.path.splitext(options.output_file_name)
-    if len(output_split[-1]) > 0:
-        options.output_file_name = output_split[0]
-        tractography_extension = output_split[1]
-    else:
-        tractography_extension = input_split[1]
-
     if tractography_extension == '.trk':
         if input_split[-1] == '.trk':
             tractography_extra_kwargs = {

tract_querier/query_processor.py

@@ -646,7 +646,7 @@ def __init__(self, value):
         self.value = value
 
     def __str__(self):
-        return repr(self.value)
+        return str(self.value)
 
 
 class TractQuerierLabelNotFound(ValueError):
@@ -655,7 +655,7 @@ def __init__(self, value):
         self.value = value
 
     def __str__(self):
-        return repr(self.value)
+        return str(self.value)
 
 
 class RewriteChangeNotInPrescedence(ast.NodeTransformer):

tract_querier/shell.py

@@ -181,7 +181,6 @@ def emptyline(self):
 
     @safe_method
     def default(self, line):
-        print(line)
         if len(line) == 0:
             return False
 

tract_querier/tests/test_tract_math.py

@@ -1,6 +1,8 @@
 #!/usr/bin/env python
 # -*- coding: utf-8 -*-
 
+import csv
+import io
 import re
 
 from tract_querier.tests import datasets
@@ -17,3 +19,29 @@ def test_tract_math_count(script_runner):
     ret = script_runner.run(["tract_math", tract_file, "count"])
     assert re.search('[^0-9]6783[^0-9]', ret.stdout) is not None
     assert ret.success
+
+
+def test_tract_math_existing_output(script_runner, tmp_path):
+    test_data = datasets.TestDataSet()
+    tract_file = test_data.files["tract_file"]
+    output = tmp_path / "count.csv"
+    output.write_text("old\n")
+
+    # Without --force and without an interactive stdin, do not prompt: fail
+    ret = script_runner.run(
+        ["tract_math", tract_file, "count", str(output)],
+        stdin=io.StringIO("")
+    )
+    assert not ret.success
+    assert "--force" in ret.stderr
+    assert output.read_text() == "old\n"
+
+    ret = script_runner.run(
+        ["tract_math", tract_file, "count", str(output), "--force"],
+        stdin=io.StringIO("")
+    )
+    assert ret.success
+    with open(output, newline='') as f:
+        rows = list(csv.reader(f))
+    assert len(rows) == 2
+    assert "6783" in rows[1]

tract_querier/tests/test_tract_querier.py

@@ -12,6 +12,27 @@ def test_help_option(script_runner):
     assert ret.success
 
 
+def test_tract_querier_unsupported_output_format(script_runner, tmp_path):
+    # ".01" is taken as the output format: it must be rejected before the
+    # files are loaded and the queries computed.
+    test_data = datasets.TestDataSet()
+    ret = script_runner.run([
+        "tract_querier",
+        "-a", test_data.files["atlas_file"],
+        "-t", test_data.files["tract_file"],
+        "-q", test_data.files["query_cst_file"],
+        "-o", str(tmp_path / "subj.01"),
+    ])
+    assert not ret.success
+    assert "'.01' not supported" in ret.stderr
+    assert "Loading files" not in ret.stdout
+
+
+def test_syntax_error_message_is_not_quoted():
+    from tract_querier import TractQuerierSyntaxError
+    assert str(TractQuerierSyntaxError("line 1\nline 2")) == "line 1\nline 2"
+
+
 def test_tract_querier_query(script_runner):
     test_data = datasets.TestDataSet()
     atlas_file = test_data.files["atlas_file"]

tract_querier/tract_math/decorator.py

@@ -4,6 +4,7 @@
 import io
 import csv
 from os import path
+import sys
 
 import nibabel
 
@@ -98,6 +99,7 @@ def wrapper(*input_args):
             # find optional flags and put them as a dictionary
             # at the beginning of args
             args, options_dict = find_optional_args(input_args)
+            force = options_dict.pop('--force', None) is not None
 
             total_args = len(args)
             argspec = inspect.getfullargspec(func)
@@ -136,7 +138,7 @@ def wrapper(*input_args):
                 option_and_args = (options_dict,) + args
                 out = func(*option_and_args)
 
-                process_output(out, file_output=file_output)
+                process_output(out, file_output=file_output, force=force)
             elif (
                 total_args >= func_total_args or
                 len(args) == func_args
@@ -160,11 +162,16 @@ def wrapper(*input_args):
     return internal_decorator
 
 
-def process_output(output, file_output=None):
+def process_output(output, file_output=None, force=False):
     if output is None:
         return
 
-    if file_output is not None and path.exists(file_output):
+    if file_output is not None and path.exists(file_output) and not force:
+        if not sys.stdin.isatty():
+            raise TractMathWrongArgumentsError(
+                'Output file %s exists, use --force to overwrite it'
+                % file_output
+            )
         in_key = input("Overwrite file %s (y/N)? " % file_output)
         if in_key.lower().strip() != 'y':
             return
@@ -187,33 +194,34 @@ def process_output(output, file_output=None):
                 dialect = 'excel-tab'
             else:
                 dialect = 'excel'
-            f = open(file_output, 'w')
-        writer = csv.DictWriter(f, output.keys(), dialect=dialect)
+            f = open(file_output, 'w', newline='')
+        with f:
+            writer = csv.DictWriter(f, output.keys(), dialect=dialect)
 
-        if hasattr(writer, 'writeheader'):
-            writer.writeheader()
-        else:
-            header = dict(zip(writer.fieldnames, writer.fieldnames))
-            writer.writerow(header)
-
-        first_value = next(iter(output.values()))
-        if (
-            not isinstance(first_value, str) and
-            isinstance(first_value, Iterable)
-        ):
-            rows = (
-                dict(zip(*row))
-                for row in zip(
-                    repeat(output.keys(), len(first_value)),
-                    zip(*output.values())
+            if hasattr(writer, 'writeheader'):
+                writer.writeheader()
+            else:
+                header = dict(zip(writer.fieldnames, writer.fieldnames))
+                writer.writerow(header)
+
+            first_value = next(iter(output.values()))
+            if (
+                not isinstance(first_value, str) and
+                isinstance(first_value, Iterable)
+            ):
+                rows = (
+                    dict(zip(*row))
+                    for row in zip(
+                        repeat(output.keys(), len(first_value)),
+                        zip(*output.values())
+                    )
                 )
-            )
-            writer.writerows(rows)
-        else:
-            writer.writerow(output)
+                writer.writerows(rows)
+            else:
+                writer.writerow(output)
 
-        if file_output is None:
-            print(f.getvalue())
+            if file_output is None:
+                print(f.getvalue())
 
 
 class TractMathWrongArgumentsError(TypeError):

tract_querier/tract_math/operations.py

@@ -220,7 +220,7 @@ def tract_dump_endpoints(optional_flags, tractography):
     return res
 
 
-@tract_math_operation(': Minimum and maximum distance between two consecutive points')
+@tract_math_operation(': minimum and maximum tract length')
 def tract_point_distance_min_max(optional_flags, tractography):
     dist_min = numpy.empty(len(tractography.tracts()))
     dist_max = numpy.empty(len(tractography.tracts()))
@@ -241,7 +241,7 @@ def tract_subsample(optional_flags, tractography, points_per_tract, file_output)
     )
 
 
-@tract_math_operation('<mm per tract> <tractography_file_output>: subsamples tracts to a maximum number of points')
+@tract_math_operation('<min length in mm> <tractography_file_output>: keeps only the tracts longer than the given length')
 def tract_remove_short_tracts(optional_flags, tractography, min_tract_length, file_output):
 
     min_tract_length = float(min_tract_length)
@@ -351,9 +351,10 @@ def tract_deform(optional_flags, tractography, image, file_output=None):
 
 
 @tract_math_operation(
-    '<transform> [invert] <tractography_file_output>: apply a '
+    '<transform> <ref_image> [invert] <tractography_file_output>: apply a '
     'affine transform to a tractography. '
-    'transform is assumed to be in RAS format like Nifti.'
+    'transform is assumed to be in RAS format like Nifti. '
+    'ref_image gives the affine and dimensions of the output.'
 )
 def tract_affine_transform(optional_flags,
                            tractography, transform_file, ref_image,