Split the main function into pieces

Have there be a scanning function that contained the
previous file scanning logic, then a validating function
that did the checking logic and then have the main function
call into those functions instead of containing all the
same logic in its own function.

Change-Id: Ie35665a016164122be294fee41e7ce54605d832e

1 files changed, 81 insertions, 71 deletions

diff --git a/doc8/main.py b/doc8/main.py
index 95ebab0..eaabdd1 100644
--- a/doc8/main.py
+++ b/doc8/main.py
@@ -148,86 +148,34 @@ def setup_logging(verbose):
                         format='%(levelname)s: %(message)s', stream=sys.stdout)
 
 
-def main():
-    parser = argparse.ArgumentParser(
-        prog='doc8',
-        description=__doc__,
-        formatter_class=argparse.RawDescriptionHelpFormatter)
-    default_configs = ", ".join(CONFIG_FILENAMES)
-    parser.add_argument("paths", metavar='path', type=str, nargs='*',
-                        help=("Path to scan for doc files"
-                              " (default: current directory)."),
-                        default=[os.getcwd()])
-    parser.add_argument("--config", metavar='path', action="append",
-                        help="User config file location"
-                             " (default: %s)." % default_configs,
-                        default=[])
-    parser.add_argument("--allow-long-titles", action="store_true",
-                        help="Allow long section titles (default: False).",
-                        default=False)
-    parser.add_argument("--ignore", action="append", metavar="code",
-                        help="Ignore the given error code(s).",
-                        type=split_set_type,
-                        default=[])
-    parser.add_argument("--no-sphinx", action="store_false",
-                        help="Do not ignore sphinx specific false positives.",
-                        default=True, dest='sphinx')
-    parser.add_argument("--ignore-path", action="append", default=[],
-                        help="Ignore the given directory or file (globs"
-                             " are supported).", metavar='path')
-    parser.add_argument("--max-line-length", action="store", metavar="int",
-                        type=int,
-                        help="Maximum allowed line"
-                             " length (default: %s)." % MAX_LINE_LENGTH,
-                        default=MAX_LINE_LENGTH)
-    parser.add_argument("-e", "--extension", action="append",
-                        metavar="extension",
-                        help="Check file extensions of the given type"
-                             " (default: %s)." % ", ".join(FILE_PATTERNS),
-                        default=list(FILE_PATTERNS))
-    parser.add_argument("-v", "--verbose", dest="verbose", action='store_true',
-                        help="Run in verbose mode.", default=False)
-    parser.add_argument("--version", dest="version", action='store_true',
-                        help="Show the version and exit.", default=False)
-    args = vars(parser.parse_args())
-    if args.get('version'):
-        print(version.version_string())
-        return 0
-    args['ignore'] = merge_sets(args['ignore'])
-    cfg = extract_config(args)
-    args['ignore'].update(cfg.pop("ignore", set()))
-    if 'sphinx' in cfg:
-        args['sphinx'] = cfg.pop("sphinx")
-    args['extension'].extend(cfg.pop('extension', []))
-    args['ignore_path'].extend(cfg.pop('ignore_path', []))
-    args.update(cfg)
-    setup_logging(args.get('verbose'))
-
+def scan(cfg):
     print("Scanning...")
     files = collections.deque()
-    ignored_paths = args.pop('ignore_path')
+    ignored_paths = cfg.pop('ignore_path')
     files_ignored = 0
-    files_selected = 0
-    file_iter = utils.find_files(args.pop('paths', []),
-                                 args.pop('extension', []), ignored_paths)
+    file_iter = utils.find_files(cfg.pop('paths', []),
+                                 cfg.pop('extension', []), ignored_paths)
     for filename, ignoreable in file_iter:
         if ignoreable:
             files_ignored += 1
-            if args.get('verbose'):
+            if cfg.get('verbose'):
                 print("  Ignoring '%s'" % (filename))
         else:
-            files_selected += 1
             files.append(file_parser.parse(filename))
-            if args.get('verbose'):
+            if cfg.get('verbose'):
                 print("  Selecting '%s'" % (filename))
+    return (files, files_ignored)
+
 
-    ignoreables = frozenset(args.pop('ignore', []))
+def validate(cfg, files):
+    print("Validating...")
     error_counts = {}
+    ignoreables = frozenset(cfg.pop('ignore', []))
     while files:
         f = files.popleft()
-        if args.get('verbose'):
+        if cfg.get('verbose'):
             print("Validating %s" % f)
-        for c in fetch_checks(args):
+        for c in fetch_checks(cfg):
             try:
                 # http://legacy.python.org/dev/peps/pep-3155/
                 check_name = c.__class__.__qualname__
@@ -241,7 +189,7 @@ def main():
                 pass
             else:
                 if not extension_matcher.match(f.extension):
-                    if args.get('verbose'):
+                    if cfg.get('verbose'):
                         print("  Skipping check '%s' since it does not"
                               " understand parsing a file with extension '%s'"
                               % (check_name, f.extension))
@@ -253,17 +201,17 @@ def main():
             else:
                 reports = reports - ignoreables
                 if not reports:
-                    if args.get('verbose'):
+                    if cfg.get('verbose'):
                         print("  Skipping check '%s', determined to only"
                               " check ignoreable codes" % check_name)
                     continue
-            if args.get('verbose'):
+            if cfg.get('verbose'):
                 print("  Running check '%s'" % check_name)
             if isinstance(c, checks.ContentCheck):
                 for line_num, code, message in c.report_iter(f):
                     if code in ignoreables:
                         continue
-                    if args.get('verbose'):
+                    if cfg.get('verbose'):
                         print('    - %s:%s: %s %s'
                               % (f.filename, line_num, code, message))
                     else:
@@ -275,7 +223,7 @@ def main():
                     for code, message in c.report_iter(line):
                         if code in ignoreables:
                             continue
-                        if args.get('verbose'):
+                        if cfg.get('verbose'):
                             print('    - %s:%s: %s %s'
                                   % (f.filename, line_num, code, message))
                         else:
@@ -285,11 +233,73 @@ def main():
             else:
                 raise TypeError("Unknown check type: %s, %s"
                                 % (type(c), c))
+    return error_counts
+
+
+def main():
+    parser = argparse.ArgumentParser(
+        prog='doc8',
+        description=__doc__,
+        formatter_class=argparse.RawDescriptionHelpFormatter)
+    default_configs = ", ".join(CONFIG_FILENAMES)
+    parser.add_argument("paths", metavar='path', type=str, nargs='*',
+                        help=("Path to scan for doc files"
+                              " (default: current directory)."),
+                        default=[os.getcwd()])
+    parser.add_argument("--config", metavar='path', action="append",
+                        help="User config file location"
+                             " (default: %s)." % default_configs,
+                        default=[])
+    parser.add_argument("--allow-long-titles", action="store_true",
+                        help="Allow long section titles (default: False).",
+                        default=False)
+    parser.add_argument("--ignore", action="append", metavar="code",
+                        help="Ignore the given error code(s).",
+                        type=split_set_type,
+                        default=[])
+    parser.add_argument("--no-sphinx", action="store_false",
+                        help="Do not ignore sphinx specific false positives.",
+                        default=True, dest='sphinx')
+    parser.add_argument("--ignore-path", action="append", default=[],
+                        help="Ignore the given directory or file (globs"
+                             " are supported).", metavar='path')
+    parser.add_argument("--max-line-length", action="store", metavar="int",
+                        type=int,
+                        help="Maximum allowed line"
+                             " length (default: %s)." % MAX_LINE_LENGTH,
+                        default=MAX_LINE_LENGTH)
+    parser.add_argument("-e", "--extension", action="append",
+                        metavar="extension",
+                        help="Check file extensions of the given type"
+                             " (default: %s)." % ", ".join(FILE_PATTERNS),
+                        default=list(FILE_PATTERNS))
+    parser.add_argument("-v", "--verbose", dest="verbose", action='store_true',
+                        help="Run in verbose mode.", default=False)
+    parser.add_argument("--version", dest="version", action='store_true',
+                        help="Show the version and exit.", default=False)
+    args = vars(parser.parse_args())
+    if args.get('version'):
+        print(version.version_string())
+        return 0
+    args['ignore'] = merge_sets(args['ignore'])
+    cfg = extract_config(args)
+    args['ignore'].update(cfg.pop("ignore", set()))
+    if 'sphinx' in cfg:
+        args['sphinx'] = cfg.pop("sphinx")
+    args['extension'].extend(cfg.pop('extension', []))
+    args['ignore_path'].extend(cfg.pop('ignore_path', []))
+    args.update(cfg)
+    setup_logging(args.get('verbose'))
+
+    files, files_ignored = scan(args)
+    files_selected = len(files)
+    error_counts = validate(args, files)
     total_errors = sum(six.itervalues(error_counts))
+
     print("=" * 8)
     print("Total files scanned = %s" % (files_selected))
     print("Total files ignored = %s" % (files_ignored))
-    print("Total accumulated errors = %s" % total_errors)
+    print("Total accumulated errors = %s" % (total_errors))
     if error_counts:
         print("Detailed error counts:")
         for check_name in sorted(six.iterkeys(error_counts)):