[pp:classify] rewrite & simplify (#5213)

Do not manually build paths, which get later overwritten by
pathfmt.build_path() anyway. Just update the target directory and let
the rest of the "path logic" handle it.

Fixes skipping previously downloaded and categorized files,
which was broken since 8124c16a50

gallery_dl/postprocessor/classify.py

@@ -1,6 +1,6 @@
 # -*- coding: utf-8 -*-
 
-# Copyright 2018-2021 Mike Fährmann
+# Copyright 2018-2024 Mike Fährmann
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License version 2 as
@@ -15,44 +15,43 @@ import os
 class ClassifyPP(PostProcessor):
 
     DEFAULT_MAPPING = {
-        "Music" : ("mp3", "aac", "flac", "ogg", "wma", "m4a", "wav"),
-        "Video" : ("flv", "ogv", "avi", "mp4", "mpg", "mpeg", "3gp", "mkv",
-                   "webm", "vob", "wmv"),
-        "Pictures" : ("jpg", "jpeg", "png", "gif", "bmp", "svg", "webp"),
+        "Pictures" : ("jpg", "jpeg", "png", "gif", "bmp", "svg", "webp",
+                      "avif", "heic", "heif", "ico", "psd"),
+        "Video"    : ("flv", "ogv", "avi", "mp4", "mpg", "mpeg", "3gp", "mkv",
+                      "webm", "vob", "wmv", "m4v", "mov"),
+        "Music"    : ("mp3", "aac", "flac", "ogg", "wma", "m4a", "wav"),
         "Archives" : ("zip", "rar", "7z", "tar", "gz", "bz2"),
+        "Documents": ("txt", "pdf"),
     }
 
     def __init__(self, job, options):
         PostProcessor.__init__(self, job)
-        mapping = options.get("mapping", self.DEFAULT_MAPPING)
+        self.directory = self.realdirectory = ""
 
+        mapping = options.get("mapping", self.DEFAULT_MAPPING)
         self.mapping = {
             ext: directory
             for directory, exts in mapping.items()
             for ext in exts
         }
-        job.register_hooks(
-            {"prepare": self.prepare, "file": self.move}, options)
+
+        job.register_hooks({
+            "post"   : self.initialize,
+            "prepare": self.prepare,
+        }, options)
+
+    def initialize(self, pathfmt):
+        # store base directory paths
+        self.directory = pathfmt.directory
+        self.realdirectory = pathfmt.realdirectory
 
     def prepare(self, pathfmt):
+        # extend directory paths depending on file extension
         ext = pathfmt.extension
         if ext in self.mapping:
-            # set initial paths to enable download skips
-            self._build_paths(pathfmt, self.mapping[ext])
-
-    def move(self, pathfmt):
-        ext = pathfmt.extension
-        if ext in self.mapping:
-            # rebuild paths in case the filename extension changed
-            path = self._build_paths(pathfmt, self.mapping[ext])
-            os.makedirs(path, exist_ok=True)
-
-    @staticmethod
-    def _build_paths(pathfmt, extra):
-        path = pathfmt.realdirectory + extra
-        pathfmt.realpath = path + os.sep + pathfmt.filename
-        pathfmt.path = pathfmt.directory + extra + os.sep + pathfmt.filename
-        return path
+            extra = self.mapping[ext] + os.sep
+            pathfmt.directory = self.directory + extra
+            pathfmt.realdirectory = self.realdirectory + extra
 
 
 __postprocessor__ = ClassifyPP

test/test_postprocessor.py

@@ -120,30 +120,37 @@ class ClassifyTest(BasePostprocessorTest):
             for directory, exts in pp.DEFAULT_MAPPING.items()
             for ext in exts
         })
-        self.pathfmt.set_extension("jpg")
-        self.pathfmt.build_path()
 
-        pp.prepare(self.pathfmt)
+        self.assertEqual(pp.directory, "")
+        self._trigger(("post",))
+        self.assertEqual(pp.directory, self.pathfmt.directory)
+
+        self.pathfmt.set_extension("jpg")
+        self._trigger(("prepare",))
+        self.pathfmt.build_path()
         path = os.path.join(self.dir.name, "test", "Pictures")
         self.assertEqual(self.pathfmt.path, path + "/file.jpg")
         self.assertEqual(self.pathfmt.realpath, path + "/file.jpg")
 
-        with patch("os.makedirs") as mkdirs:
-            self._trigger()
-            mkdirs.assert_called_once_with(path, exist_ok=True)
+        self.pathfmt.set_extension("mp4")
+        self._trigger(("prepare",))
+        self.pathfmt.build_path()
+        path = os.path.join(self.dir.name, "test", "Video")
+        self.assertEqual(self.pathfmt.path, path + "/file.mp4")
+        self.assertEqual(self.pathfmt.realpath, path + "/file.mp4")
 
     def test_classify_noop(self):
         pp = self._create()
         rp = self.pathfmt.realpath
 
-        pp.prepare(self.pathfmt)
+        self.assertEqual(pp.directory, "")
+        self._trigger(("post",))
+        self._trigger(("prepare",))
+
+        self.assertEqual(pp.directory, self.pathfmt.directory)
         self.assertEqual(self.pathfmt.path, rp)
         self.assertEqual(self.pathfmt.realpath, rp)
 
-        with patch("os.makedirs") as mkdirs:
-            self._trigger()
-            self.assertEqual(mkdirs.call_count, 0)
-
     def test_classify_custom(self):
         pp = self._create({"mapping": {
             "foo/bar": ["foo", "bar"],
@@ -153,18 +160,18 @@ class ClassifyTest(BasePostprocessorTest):
             "foo": "foo/bar",
             "bar": "foo/bar",
         })
-        self.pathfmt.set_extension("foo")
-        self.pathfmt.build_path()
 
-        pp.prepare(self.pathfmt)
+        self.assertEqual(pp.directory, "")
+        self._trigger(("post",))
+        self.assertEqual(pp.directory, self.pathfmt.directory)
+
+        self.pathfmt.set_extension("foo")
+        self._trigger(("prepare",))
+        self.pathfmt.build_path()
         path = os.path.join(self.dir.name, "test", "foo", "bar")
         self.assertEqual(self.pathfmt.path, path + "/file.foo")
         self.assertEqual(self.pathfmt.realpath, path + "/file.foo")
 
-        with patch("os.makedirs") as mkdirs:
-            self._trigger()
-            mkdirs.assert_called_once_with(path, exist_ok=True)
-
 
 class ExecTest(BasePostprocessorTest):