StandardBear: Add fixing capabilities

bears/js/JSStandardBear.py

@@ -1,11 +1,17 @@
+from collections import defaultdict
+import re
+from subprocess import Popen, PIPE
+from typing import List, Iterable, Tuple
+
 from coalib.bearlib.abstractions.Linter import linter
 from dependency_management.requirements.NpmRequirement import NpmRequirement
+from coalib.results.Result import Result
+from coalib.results.Diff import Diff
 
 
 @linter(executable='standard',
-        output_format='regex',
-        output_regex=r'\s*[^:]+:(?P<line>\d+):(?P<column>\d+):'
-                     r'\s*(?P<message>.+)')
+        use_stdin=True,
+        use_stderr=True)
 class JSStandardBear:
     """
     One JavaScript Style to Rule Them All.
@@ -24,5 +30,65 @@ class JSStandardBear:
     CAN_FIX = {'Formatting'}
     SEE_MORE = 'https://standardjs.com/rules.html'
 
+    issue_regex = re.compile(
+        r'\s*[^:]+:(?P<line>\d+):(?P<column>\d+):'
+        r'\s*(?P<message>.+)')
+
     def create_arguments(self, filename, file, config_file):
         return (filename, '--verbose')
+
+    @staticmethod
+    def _get_corrected_code(old_code: List[str]) -> List[str]:
+        """
+        Pipes the code to JSStandard and returns the corrected code.
+        """
+        p = Popen(
+            ('standard', '--stdin', '--fix'),
+            stdin=PIPE, stdout=PIPE, stderr=PIPE)
+        p.stdin.write(bytes(''.join(old_code), 'UTF-8'))
+        out, err = p.communicate()
+        return out.decode('UTF-8').splitlines(True)
+
+    def _get_issues(self, stdout: str) -> Iterable[Tuple[int, str]]:
+        """
+        Gets the issues from the output of JSStandard.
+
+        The issues get parsed with `self.issue_regex` and merged if they
+        concern the same line.
+
+        :param stdout: Output from which the issues get parsed.
+        :return: List of tuples containing the line number and the message.
+        """
+        match_objects = (
+            self.issue_regex.match(line) for line in stdout.splitlines())
+        issues = (
+            match_object.groupdict()
+            for match_object in match_objects if match_object is not None)
+        line_number_to_messages = defaultdict(list)
+        for issue in issues:
+            line_number = int(issue['line'])
+            line_number_to_messages[line_number].append(issue['message'])
+        return (
+            (line_number, '\n'.join(messages))
+            for line_number, messages
+            in sorted(line_number_to_messages.items()))
+
+    def process_output(self, output, filename, file):
+        stdout, stderr = output
+        corrected_code = None
+        if '--fix' in stderr:
+            corrected_code = self._get_corrected_code(file)
+            if len(file) != len(corrected_code):
+                corrected_code = None
+
+        for line_number, message in self._get_issues(stdout):
+            diff = None
+            if corrected_code:
+                diff = Diff(file)
+                diff.modify_line(line_number, corrected_code[line_number-1])
+            yield Result.from_values(
+                origin=self,
+                message=message,
+                file=filename,
+                line=line_number,
+                diffs={filename: diff})

tests/js/JSStandardBearTest.py

@@ -1,8 +1,12 @@
-from coalib.testing.LocalBearTestHelper import verify_local_bear
+from queue import Queue
 
+from coalib.results.Diff import Diff
+from coalib.settings.Section import Section
+from coalib.testing.BearTestHelper import generate_skip_decorator
+from coalib.testing.LocalBearTestHelper import verify_local_bear
+from coalib.testing.LocalBearTestHelper import LocalBearTestHelper
 from bears.js.JSStandardBear import JSStandardBear
 
-
 good_file = """
 var foo = {
   bar: 1,
@@ -77,3 +81,95 @@
                                                       bad_file_undef,
                                                       bad_file_ifelse,
                                                       bad_file_func_name,))
+
+
+@generate_skip_decorator(JSStandardBear)
+class JSStandardBearTestClass(LocalBearTestHelper):
+    bad_code = """
+(function () {
+    console.log('wrong indentation')
+    console.log('wrong indentation and semicolon');
+}())
+""".splitlines(True)
+
+    def setUp(self):
+        section = Section('name')
+        self.js_standard_bear = JSStandardBear(section, Queue())
+
+    def _check_diff(self, result, line_number, corrected_code):
+        self.assertEqual(1, len(result.affected_code))
+        affected_code = result.affected_code[0]
+        self.assertEqual(line_number, affected_code.start.line)
+        self.assertEqual(line_number, affected_code.end.line)
+
+        self.assertEqual(1, len(result.diffs))
+        diff = list(result.diffs.values())[0]
+        expected_diff = Diff(self.bad_code)
+        expected_diff.modify_line(line_number, corrected_code)
+        self.assertEqual(expected_diff, diff)
+
+    def test_diffs(self):
+        results = self.check_invalidity(
+            self.js_standard_bear,
+            self.bad_code)
+        self.assertEqual(2, len(results))
+
+        corrected_code = "  console.log('wrong indentation')"
+        self._check_diff(results[0], 3, corrected_code)
+
+        corrected_code = "  console.log('wrong indentation and semicolon')"
+        self._check_diff(results[1], 4, corrected_code)
+
+    def test_messages(self):
+        results = self.check_invalidity(
+            self.js_standard_bear,
+            self.bad_code)
+        self.assertEqual(2, len(results))
+        self.assertIn('(indent)', results[0].message)
+
+        # Check that messages for lines with multiple issues get merged.
+        self.assertIn('(indent)', results[1].message)
+        self.assertIn('(semi)', results[1].message)
+
+    def test_corrected_code(self):
+        """
+        Check that the corrected code is valid.
+        """
+        corrected_code = self.js_standard_bear._get_corrected_code(
+            self.bad_code)
+        self.check_validity(
+            self.js_standard_bear,
+            corrected_code)
+
+    def test_get_issues_from_output(self):
+        output = """
+coala-bears/test.js:2:5: Expected indentation of 2 spaces but found 4. (indent)
+coala-bears/test.js:3:5: Expected indentation of 2 spaces but found 4. (indent)
+coala-bears/test.js:3:51: Extra semicolon. (semi)
+"""
+        issues = list(self.js_standard_bear._get_issues(output))
+        line_numbers = [line_number for line_number, _ in issues]
+        self.assertEqual(line_numbers, [2, 3])
+        messages = [message for _, message in issues]
+        expected_messages = [
+            'Expected indentation of 2 spaces but found 4. (indent)',
+            ('Expected indentation of 2 spaces but found 4. (indent)\n'
+             'Extra semicolon. (semi)')]
+        self.assertEqual(messages, expected_messages)
+
+    def test_different_number_of_lines(self):
+        """
+        Check a case where the number of lines of the invalid and the
+        corrected code differs.
+        """
+        bad_code_with_empty_lines = ['\n', '\n'] + self.bad_code
+        results = self.check_invalidity(
+            self.js_standard_bear,
+            bad_code_with_empty_lines)
+        self.assertEqual(3, len(results))
+        self.assertIn('(no-multiple-empty-lines)', results[0].message)
+
+        # When there the number of lines differ, the diff is `None`.
+        self.assertEqual(1, len(results[0].diffs))
+        diff = list(results[0].diffs.values())[0]
+        self.assertEqual(None, diff)