ENH: Use flake8 to check for PEP8 violations in doctests

scripts/validate_docstrings.py

@@ -24,6 +24,10 @@
 import inspect
 import importlib
 import doctest
+from contextlib import contextmanager
+
+from flake8.api import legacy as flake8
+
 try:
     from io import StringIO
 except ImportError:
@@ -40,7 +44,6 @@
 from numpydoc.docscrape import NumpyDocString
 from pandas.io.formats.printing import pprint_thing
 
-
 PRIVATE_CLASSES = ['NDFrame', 'IndexOpsMixin']
 DIRECTIVES = ['versionadded', 'versionchanged', 'deprecated']
 
@@ -331,6 +334,43 @@ def parameter_mismatches(self):
 
         return errs
 
+    @property
+    def pep8_violations(self):
+        with self._file_representation() as filename:
+            style_guide = flake8.get_style_guide(doctests=True)
+            report = style_guide.input_file(filename=filename)
+            return report.get_statistics('')
+
+    @contextmanager
+    def _file_representation(self):
+        """
+        Creates a tmp file containing the function without the body
+
+        :returns filename of tmp file
+        """
+        create_function = 'def {name}{signature}:  # noqa: E501\n' \
+                          '    """{doc}"""\n' \
+                          '    pass\n'
+
+        tmp_dir = os.path.join(BASE_PATH, 'build', 'validate_docstring')
+        os.makedirs(tmp_dir, exist_ok=True)
+
+        filename = os.path.join(tmp_dir, self.name + '.py')
+        with open(filename, 'w') as f:
+            name = self.name.split('.')[-1]
+            sig = str(inspect.signature(self.obj))
+            lines = self.raw_doc.split("\n")
+            indented_lines = ['    ' + line if line else ''
+                              for line in lines[1:]]
+            doc = '\n'.join([lines[0], *indented_lines])
+
+            f.write(create_function.format(name=name, signature=sig, doc=doc))
+
+        yield filename
+
+        os.remove(filename)
+        os.rmdir(tmp_dir)
+
     @property
     def correct_parameters(self):
         return not bool(self.parameter_mismatches)
@@ -446,7 +486,7 @@ def validate_one(func_name):
         if doc.summary != doc.summary.lstrip():
             errs.append('Summary contains heading whitespaces.')
         elif (doc.is_function_or_method
-                and doc.summary.split(' ')[0][-1] == 's'):
+              and doc.summary.split(' ')[0][-1] == 's'):
             errs.append('Summary must start with infinitive verb, '
                         'not third person (e.g. use "Generate" instead of '
                         '"Generates")')
@@ -490,6 +530,12 @@ def validate_one(func_name):
         for param_err in param_errs:
             errs.append('\t{}'.format(param_err))
 
+    pep8_errs = doc.pep8_violations
+    if pep8_errs:
+        errs.append('Errors in doctest sections')
+        for pep8_err in pep8_errs:
+            errs.append('\t{}'.format(pep8_err))
+
     if doc.is_function_or_method:
         if not doc.returns and "return" in doc.method_source:
             errs.append('No Returns section found')

setup.cfg

@@ -28,6 +28,31 @@ exclude =
     doc/temp/*.py,
     .eggs/*.py,
     versioneer.py
+    .tox
+    .git
+
+doctests = True
+#TODO fix doctests
+exclude_from_doctest =
+    ./pandas/_libs
+    ./pandas/api
+    ./pandas/compat
+    ./pandas/computation
+    ./pandas/core
+    ./pandas/errors
+    ./pandas/io
+    ./pandas/plotting
+    ./pandas/tests
+    ./pandas/tools
+    ./pandas/tseries
+    ./pandas/types
+    ./pandas/util
+
+[flake8-rst]
+ignore =
+    F821,  # undefined name
+    W391,  # blank line at end of file [Seems to be a bug (v0.4.1)]
+
 
 [yapf]
 based_on_style = pep8