Merge branch 'import-zip'

Closes #156

Author: Thomas Scholtes

beets/importer.py

@@ -22,6 +22,8 @@
 import pickle
 import itertools
 from collections import defaultdict
+from tempfile import mkdtemp
+import shutil
 
 from beets import autotag
 from beets import library
@@ -525,6 +527,11 @@ def imported_items(self):
         else:
             return [self.item]
 
+    def cleanup(self):
+        """Perform clean up during `finalize` stage.
+        """
+        pass
+
     # Utilities.
 
     def prune(self, filename):
@@ -540,6 +547,78 @@ def prune(self, filename):
                             clutter=config['clutter'].as_str_seq())
 
 
+class ArchiveImportTask(ImportTask):
+    """Additional methods for handling archives.
+
+    Use when `toppath` points to a `zip`, `tar`, or `rar` archive.
+    """
+
+    def __init__(self, toppath):
+        super(ArchiveImportTask, self).__init__(toppath)
+        self.sentinel = True
+        self.extracted = False
+
+    @classmethod
+    def is_archive(cls, path):
+        """Returns true if the given path points to an archive that can
+        be handled.
+        """
+        if not os.path.isfile(path):
+            return False
+
+        for path_test, _ in cls.handlers():
+            if path_test(path):
+                return True
+        return False
+
+    @classmethod
+    def handlers(cls):
+        """Returns a list of archive handlers.
+
+        Each handler is a `(path_test, ArchiveClass)` tuple. `path_test`
+        is a function that returns `True` if the given path can be
+        handled by `ArchiveClass`. `ArchiveClass` is a class that
+        implements the same interface as `tarfile.TarFile`.
+        """
+        if not hasattr(cls, '_handlers'):
+            cls._handlers = []
+            from zipfile import is_zipfile, ZipFile
+            cls._handlers.append((is_zipfile, ZipFile))
+            from tarfile import is_tarfile, TarFile
+            cls._handlers.append((is_tarfile, TarFile))
+            try:
+                from rarfile import is_rarfile, RarFile
+            except ImportError:
+                pass
+            else:
+                cls._handlers.append((is_rarfile, RarFile))
+
+        return cls._handlers
+
+    def cleanup(self):
+        """Removes the temporary directory the archive was extracted to.
+        """
+        if self.extracted:
+            shutil.rmtree(self.toppath)
+
+    def extract(self):
+        """Extracts the archive to a temporary directory and sets
+        `toppath` to that directory.
+        """
+        for path_test, handler_class in self.handlers():
+            if path_test(self.toppath):
+                break
+
+        try:
+            extract_to = mkdtemp()
+            archive = handler_class(self.toppath, mode='r')
+            archive.extractall(extract_to)
+        finally:
+            archive.close()
+        self.extracted = True
+        self.toppath = extract_to
+
+
 # Full-album pipeline stages.
 
 def read_tasks(session):
@@ -573,6 +652,24 @@ def read_tasks(session):
         history_dirs = history_get()
 
     for toppath in session.paths:
+        # Extract archives
+        archive_task = None
+        if ArchiveImportTask.is_archive(syspath(toppath)):
+            if not (config['import']['move'] or config['import']['copy']):
+                log.warn("Cannot import archive. Please set "
+                         "the 'move' or 'copy' option.")
+                continue
+
+            log.debug('extracting archive {0}'
+                      .format(displayable_path(toppath)))
+            archive_task = ArchiveImportTask(toppath)
+            try:
+                archive_task.extract()
+            except Exception as exc:
+                log.error('extraction failed: {0}'.format(exc))
+                continue
+            toppath = archive_task.toppath
+
         # Check whether the path is to a file.
         if not os.path.isdir(syspath(toppath)):
             try:
@@ -627,7 +724,11 @@ def read_tasks(session):
                 yield ImportTask(toppath, paths, items)
 
         # Indicate the directory is finished.
-        yield ImportTask.done_sentinel(toppath)
+        # FIXME hack to delete extraced archives
+        if archive_task is None:
+            yield ImportTask.done_sentinel(toppath)
+        else:
+            yield archive_task
 
     # Show skipped directories.
     if config['import']['incremental'] and incremental_skipped:
@@ -937,6 +1038,7 @@ def finalize(session):
                 task.save_progress()
             if config['import']['incremental']:
                 task.save_history()
+            task.cleanup()
             continue
 
         items = task.imported_items()
@@ -971,6 +1073,7 @@ def finalize(session):
             task.save_progress()
         if config['import']['incremental']:
             task.save_history()
+        task.cleanup()
 
 
 # Singleton pipeline stages.

docs/changelog.rst

@@ -25,6 +25,7 @@ New stuff:
 * There is also a new :doc:`/plugins/keyfinder` that runs a command line
   program to get the key from audio data and store it in the
   `initial_key` field.
+* Beets can now import `zip`, `tar` and `rar` archives.
 
 Fixes:
 

docs/reference/cli.rst

@@ -43,21 +43,25 @@ import
 ``````
 ::
 
-    beet import [-CWAPRqst] [-l LOGPATH] DIR...
+    beet import [-CWAPRqst] [-l LOGPATH] PATH...
     beet import [options] -L QUERY
 
 Add music to your library, attempting to get correct tags for it from
 MusicBrainz.
 
-Point the command at a directory full of music. The directory can be a single
-album or a directory whose leaf subdirectories are albums (the latter case is
-true of typical Artist/Album organizations and many people's "downloads"
-folders). The music will be copied to a configurable directory structure (see
-below) and added to a library database (see below). The command is interactive
-and will try to get you to verify MusicBrainz tags that it thinks are suspect.
-(This means that importing a large amount of music is therefore very tedious
-right now; this is something we need to work on. Read the
-:doc:`autotagging guide </guides/tagger>` if you need help.)
+Point the command at a directory full of music. The directory can be a
+single album or a directory whose leaf subdirectories are albums (the
+latter case is true of typical Artist/Album organizations and many
+people's "downloads" folders). The path can also be a single file or an
+archive. Beets supports `zip` and `tar` archives out of the box. To
+extract `rar` files you need to install the `rarfile`_ package and the
+`unrar` command. The music will be copied to a configurable directory
+structure (see below) and added to a library database (see below). The
+command is interactive and will try to get you to verify MusicBrainz
+tags that it thinks are suspect.  (This means that importing a large
+amount of music is therefore very tedious right now; this is something
+we need to work on. Read the :doc:`autotagging guide </guides/tagger>`
+if you need help.)
 
 * By default, the command copies files your the library directory and
   updates the ID3 tags on your music. If you'd like to leave your music
@@ -116,6 +120,8 @@ right now; this is something we need to work on. Read the
   ``--group-albums`` option to split the files based on their metadata before
   matching them as separate albums.
 
+.. _rarfile: https://pypi.python.org/pypi/rarfile/2.2
+
 .. only:: html
 
     Reimporting

setup.py

@@ -99,6 +99,7 @@ def _read(fn):
         'echonest_tempo': ['pyechonest'],
         'lastgenre': ['pylast'],
         'web': ['flask'],
+        'import': ['rarfile'],
     },
     # Non-Python/non-PyPI plugin dependencies:
     # replaygain: mp3gain || aacgain

test/rsrc/archive.rar

@@ -17,10 +17,13 @@
 import os
 import shutil
 import StringIO
+from tempfile import mkstemp
+from zipfile import ZipFile
+from tarfile import TarFile
 
 import _common
 from _common import unittest
-from helper import TestImportSession, TestHelper
+from helper import TestImportSession, TestHelper, has_program
 from beets import library
 from beets import importer
 from beets.mediafile import MediaFile
@@ -299,6 +302,60 @@ def test_import_with_delete_prunes_directory_empty(self):
         self.assertNotExists(os.path.join(self.import_dir, 'the_album'))
 
 
+class ImportZipTest(unittest.TestCase, ImportHelper):
+
+    def setUp(self):
+        self.setup_beets()
+
+    def tearDown(self):
+        self.teardown_beets()
+
+    def test_import_zip(self):
+        zip_path = self.create_archive()
+        self.assertEqual(len(self.lib.items()), 0)
+        self.assertEqual(len(self.lib.albums()), 0)
+
+        self._setup_import_session(autotag=False, import_dir=zip_path)
+        self.importer.run()
+        self.assertEqual(len(self.lib.items()), 1)
+        self.assertEqual(len(self.lib.albums()), 1)
+
+    def create_archive(self):
+        (handle, path) = mkstemp(dir=self.temp_dir)
+        os.close(handle)
+        archive = ZipFile(path, mode='w')
+        archive.write(os.path.join(_common.RSRC, 'full.mp3'),
+                      'full.mp3')
+        archive.close()
+        return path
+
+
+class ImportTarTest(ImportZipTest):
+
+    def create_archive(self):
+        (handle, path) = mkstemp(dir=self.temp_dir)
+        os.close(handle)
+        archive = TarFile(path, mode='w')
+        archive.add(os.path.join(_common.RSRC, 'full.mp3'),
+                    'full.mp3')
+        archive.close()
+        return path
+
+
+@unittest.skipIf(not has_program('unrar'), 'unrar program not found')
+class ImportRarTest(ImportZipTest):
+
+    def create_archive(self):
+        return os.path.join(_common.RSRC, 'archive.rar')
+
+
+@unittest.skip('Implment me!')
+class ImportPasswordRarTest(ImportZipTest):
+
+    def create_archive(self):
+        return os.path.join(_common.RSRC, 'password.rar')
+
+
 class ImportSingletonTest(_common.TestCase, ImportHelper):
     """Test ``APPLY`` and ``ASIS`` choices for an import session with singletons
     config set to True.

test/test_importer.py

@@ -14,6 +14,7 @@ deps =
     flask
     responses
     pyechonest
+    rarfile
 commands =
     nosetests {posargs}