Add support to limit checks to git diff (#1)

Author: anapaulagomes

README.rst

@@ -46,6 +46,27 @@ The following command will lint all files in the current directory:
 The default configuration file name is `.relint.yaml` within your working
 directory, but you can provide any YAML or JSON file.
 
+If you prefer linting changed files (cached on git) you can use the option
+`--diff [-d]`:
+
+.. code-block:: bash
+
+    git diff | relint my_file.py --diff
+
+This option is useful for pre-commit purposes. Here an example of how to use it
+with `pre-commit`_ framework:
+
+.. code-block:: YAML
+
+    - repo: local
+      hooks:
+      - id: relint
+          name: relint
+          entry: bin/relint-pre-commit.sh
+          language: system
+
+You can find an example of `relint-pre-commit.sh`_ in this repository.
+
 Samples
 -------
 
@@ -76,3 +97,6 @@ Samples
       hint: "Please write to self.stdout or self.stderr in favor of using a logger."
       filename:
         - "*/management/commands/*.py"
+
+.. _`pre-commit`: https://pre-commit.com/
+.. _`relint-pre-commit.sh`: relint-pre-commit.sh

relint-pre-commit.sh

@@ -0,0 +1,4 @@
+#!/usr/bin/env sh
+
+set -eo pipefail
+git diff --staged | relint --diff ${@:1}

relint.py

@@ -2,12 +2,24 @@
 import fnmatch
 import glob
 import re
+import sys
 from collections import namedtuple
 from itertools import chain
 
 import yaml
 
 
+GIT_DIFF_LINE_NUMBERS_PATTERN = re.compile(
+    r"@ -\d+(,\d+)? \+(\d+)(,)?(\d+)? @")
+GIT_DIFF_FILENAME_PATTERN = re.compile(
+    r"(?:\n|^)diff --git a\/.* b\/(.*)(?:\n|$)")
+GIT_DIFF_SPLIT_PATTERN = re.compile(
+    r"(?:\n|^)diff --git a\/.* b\/.*(?:\n|$)")
+
+
+Test = namedtuple('Test', ('name', 'pattern', 'hint', 'filename', 'error'))
+
+
 def parse_args():
     parser = argparse.ArgumentParser()
     parser.add_argument(
@@ -18,19 +30,22 @@ def parse_args():
         help='Path to one or multiple files to be checked.'
     )
     parser.add_argument(
-        '--config',
         '-c',
+        '--config',
         metavar='CONFIG_FILE',
         type=str,
         default='.relint.yml',
         help='Path to config file, default: .relint.yml'
     )
+    parser.add_argument(
+        '-d',
+        '--diff',
+        action='store_true',
+        help='Analyze content from git diff.'
+    )
     return parser.parse_args()
 
 
-Test = namedtuple('Test', ('name', 'pattern', 'hint', 'filename', 'error'))
-
-
 def load_config(path):
     with open(path) as fs:
         for test in yaml.load(fs):
@@ -56,31 +71,77 @@ def lint_file(filename, tests):
         for test in tests:
             if any(fnmatch.fnmatch(filename, fp) for fp in test.filename):
                 for match in test.pattern.finditer(content):
-                    yield filename, test, match
+                    line_number = match.string[:match.start()].count('\n') + 1
+                    yield filename, test, match, line_number
 
 
-def main():
-    args = parse_args()
-    paths = {
-        path
-        for file in args.files
-        for path in glob.iglob(file, recursive=True)
-    }
+def parse_line_numbers(output):
+    """
+    Extract line numbers from ``git diff`` output.
 
-    tests = list(load_config(args.config))
+    Git shows which lines were changed indicating a start line
+    and how many lines were changed from that. If only one
+    line was changed, the output will display only the start line,
+    like this:
+    ``@@ -54 +54 @@ import glob``
+    If more lines were changed from that point, it will show
+    how many after a comma:
+    ``@@ -4,2 +4,2 @@ import glob``
+    It means that line number 4 and the following 2 lines were changed
+    (5 and 6).
 
-    matches = chain.from_iterable(
-        lint_file(path, tests)
-        for path in paths
-    )
+    Args:
+        output (int): ``git diff`` output.
+
+    Returns:
+        list: All changed line numbers.
+    """
+    line_numbers = []
+    matches = GIT_DIFF_LINE_NUMBERS_PATTERN.finditer(output)
+
+    for match in matches:
+        start = int(match.group(2))
+        if match.group(4) is not None:
+            end = start + int(match.group(4))
+            line_numbers.extend(range(start, end))
+        else:
+            line_numbers.append(start)
+
+    return line_numbers
+
+
+def parse_filenames(output):
+    return re.findall(GIT_DIFF_FILENAME_PATTERN, output)
 
-    _filename = ''
-    lines = []
 
+def split_diff_content_by_filename(output):
+    """
+    Split the output by filename.
+
+    Args:
+        output (int): ``git diff`` output.
+
+    Returns:
+        dict: Filename and its content.
+    """
+    content_by_filename = {}
+    filenames = parse_filenames(output)
+    splited_content = re.split(GIT_DIFF_SPLIT_PATTERN, output)
+    splited_content = filter(lambda x: x != '', splited_content)
+
+    for filename, content in zip(filenames, splited_content):
+        content_by_filename[filename] = content
+    return content_by_filename
+
+
+def print_culprits(matches):
     exit_code = 0
+    _filename = ''
+    lines = []
 
-    for filename, test, match in matches:
+    for filename, test, match, _ in matches:
         exit_code = test.error if exit_code == 0 else exit_code
+
         if filename != _filename:
             _filename = filename
             lines = match.string.splitlines()
@@ -102,6 +163,46 @@ def main():
         )
         print(*match_lines, sep="\n")
 
+    return exit_code
+
+
+def match_with_diff_changes(content, matches):
+    """Check matches found on diff output."""
+    for filename, test, match, line_number in matches:
+        if content.get(filename) and line_number in content.get(filename):
+            yield filename, test, match, line_number
+
+
+def parse_diff(output):
+    """Parse changed content by file."""
+    changed_content = {}
+    for filename, content in split_diff_content_by_filename(output).items():
+        changed_line_numbers = parse_line_numbers(content)
+        changed_content[filename] = changed_line_numbers
+    return changed_content
+
+
+def main():
+    args = parse_args()
+    paths = {
+        path
+        for file in args.files
+        for path in glob.iglob(file, recursive=True)
+    }
+
+    tests = list(load_config(args.config))
+
+    matches = chain.from_iterable(
+        lint_file(path, tests)
+        for path in paths
+    )
+
+    if args.diff:
+        output = sys.stdin.read()
+        changed_content = parse_diff(output)
+        matches = match_with_diff_changes(changed_content, matches)
+
+    exit_code = print_culprits(matches)
     exit(exit_code)
 
 

test.diff

@@ -0,0 +1,59 @@
+diff --git a/README.rst b/README.rst
+index 43032c5..e7203b3 100644
+--- a/README.rst
++++ b/README.rst
+@@ -51,7 +51,7 @@ If you prefer linting changed files (cached on git) you can use the option
+ 
+ .. code-block:: bash
+ 
+-    relint my_file.py --diff
++    git diff | relint my_file.py --diff
+ 
+ This option is useful for pre-commit purposes.
+ 
+diff --git a/relint.py b/relint.py
+index 31061ec..697a3f0 100644
+--- a/relint.py
++++ b/relint.py
+@@ -113,7 +113,7 @@ def print_culprits(matches):
+     for filename, test, match, _ in matches:
+         exit_code = test.error if exit_code == 0 else exit_code
+ 
+-        if filename != _filename:
++        if filename != _filename:  # TODO check this out
+             _filename = filename
+             lines = match.string.splitlines()
+ 
+@@ -167,7 +167,7 @@ def main():
+         for path in paths
+     )
+ 
+-    if args.diff:
++    if args.diff:  # TODO wow
+         output = sys.stdin.read()
+         changed_content = parse_diff(output)
+         matches = filter_paths_from_diff(changed_content, matches)
+diff --git a/test_relint.py b/test_relint.py
+index 7165fd3..249b783 100644
+--- a/test_relint.py
++++ b/test_relint.py
+@@ -54,8 +54,9 @@ class TestParseGitDiff:
+     def test_split_diff_content(self):
+         output = open('test.diff').read()
+         splited = split_diff_content(output)
++
+         assert isinstance(splited, dict)
+-        assert len(splited) == 2
++        assert len(splited) == 3
+ 
+     def test_return_empty_list_if_can_not_split_diff_content(self):
+         splited = split_diff_content('')
+@@ -120,7 +121,7 @@ class TestParseGitDiff:
+                 "@@ -1,0 +2 @@\n" \
+                 "+# TODO: I'll do it later, promise\n"
+ 
+-        parsed_content = parse_diff(output)
++        parsed_content = parse_diff(output)  # TODO brand new
+         expected = {'test_relint.py': [2]}
+ 
+         assert parsed_content == expected