introduce shellutils.globfind as an alternative to find that accepts glob patterns (easier to have separate function for backward compatibility reasons)

2 files changed, 47 insertions, 7 deletions

diff --git a/shellutils.py b/shellutils.py
index f6072f1..6a7509d 100644
--- a/shellutils.py
+++ b/shellutils.py
@@ -8,12 +8,13 @@ scripts.
 """
 __docformat__ = "restructuredtext en"
 
-import os        
+import os
 import glob
 import shutil
 import sys
 import tempfile
 import time
+import fnmatch
 from os.path import exists, isdir, islink, basename, join, walk
 
 from logilab.common import STD_BLACKLIST
@@ -141,6 +142,35 @@ def find(directory, exts, exclude=False, blacklist=STD_BLACKLIST):
     return files
 
 
+def globfind(directory, pattern, blacklist=STD_BLACKLIST):
+    """Recursively finds files matching glob `pattern` under `directory`.
+
+    This is an alternative to `logilab.common.shellutils.find`.
+    
+    :type directory: str
+    :param directory:
+      directory where the search should start
+
+    :type pattern: basestring
+    :param pattern:
+      the glob pattern (e.g *.py, foo*.py, etc.)
+
+    :type blacklist: list or tuple
+    :param blacklist:
+      optional list of files or directory to ignore, default to the value of
+      `logilab.common.STD_BLACKLIST`
+
+    :rtype: iterator
+    :return:
+      iterator over the list of all matching files
+    """
+    for curdir, dirnames, filenames in os.walk(directory):
+        for fname in fnmatch.filter(filenames, pattern):
+            yield join(curdir, fname)
+        for skipped in blacklist:
+            if skipped in dirnames:
+                dirnames.remove(skipped)
+
 def unzip(archive, destdir):
     import zipfile
     if not exists(destdir):
@@ -154,7 +184,6 @@ def unzip(archive, destdir):
             outfile.write(zfobj.read(name))
             outfile.close()
 
-
 class Execute:
     """This is a deadlock safe version of popen2 (no stdin), that returns
     an object with errorlevel, out and err.
diff --git a/test/unittest_shellutils.py b/test/unittest_shellutils.py
index 5e63170..396d7c6 100644
--- a/test/unittest_shellutils.py
+++ b/test/unittest_shellutils.py
@@ -5,7 +5,7 @@ from os.path import join
 
 from logilab.common.testlib import TestCase, unittest_main
 
-from logilab.common.shellutils import find, ProgressBar
+from logilab.common.shellutils import globfind, find, ProgressBar
 from StringIO import StringIO
 
 DATA_DIR = join('data','find_test')
@@ -36,10 +36,21 @@ class FindTC(TestCase):
                                                        join('sub', 'doc.txt'),
                                                        'write_protected_file.txt',
                                                        ]]))
-        
-#    def test_exclude_base_dir(self):
-#        self.assertEquals(files_by_ext(DATA_DIR, include_exts=('.py',), exclude_dirs=(DATA_DIR,)),
-#                          [])
+
+    def test_globfind(self):
+        files = set(globfind(DATA_DIR, '*.py'))
+        self.assertSetEqual(files,
+                            set([join(DATA_DIR, f) for f in ['__init__.py', 'module.py',
+                                                       'module2.py', 'noendingnewline.py',
+                                                       'nonregr.py', join('sub', 'momo.py')]]))
+        files = set(globfind(DATA_DIR, 'mo*.py'))
+        self.assertSetEqual(files,
+                            set([join(DATA_DIR, f) for f in ['module.py', 'module2.py',
+                                                             join('sub', 'momo.py')]]))
+        files = set(globfind(DATA_DIR, 'mo*.py', blacklist=('sub',)))
+        self.assertSetEqual(files,
+                            set([join(DATA_DIR, f) for f in ['module.py', 'module2.py']]))
+
 
 class ProgressBarTC(TestCase):
     def test_refresh(self):