From fb031fcaf6331fbe6cad52b19d8acdebc257f191 Mon Sep 17 00:00:00 2001
From: Jakob Moser <moser@cl.uni-heidelberg.de>
Date: Sun, 1 Jun 2025 12:11:35 +0200
Subject: [PATCH] Implement move instruction generation

---
 karaokatalog/organize/move.py | 45 ++++++++++++++++++++++++++++++++++-
 1 file changed, 44 insertions(+), 1 deletion(-)

diff --git a/karaokatalog/organize/move.py b/karaokatalog/organize/move.py
index 7f9f1a6..f83c774 100644
--- a/karaokatalog/organize/move.py
+++ b/karaokatalog/organize/move.py
@@ -1,13 +1,56 @@
+import re
 from collections.abc import Sequence
 from pathlib import Path
 
 from karaokatalog.instructions.MoveInstruction import MoveInstruction
 from karaokatalog.Song import Song
 
+FORBIDDEN_CHARACTERS = re.compile(r'[/<>:"\\|?*]')
+
+
+def _get_canonical_song_dir(song: Song, variant: int = 0) -> Path:
+    """
+    Get the (relative) canonical directory in which this song should be located, that is
+    the directory f"{song.artist}/{song.title}".
+
+    Illegal characters are replaced. If variant > 0, f" ({variant})" is appended to the result.
+    """
+    variant_suffix = f" ({variant})" if variant > 0 else ""
+    return Path(FORBIDDEN_CHARACTERS.sub("", song.artist)) / Path(
+        FORBIDDEN_CHARACTERS.sub("", song.title) + variant_suffix
+    )
+
 
 def move(songs: Sequence[Song], base_dir: Path) -> Sequence[MoveInstruction]:
     """
     Create move instructions to move every song into the proper song directory
     within the given base_dir.
     """
-    raise NotImplementedError()  # TODO
+    song_dirs = set(song.dir.relative_to(base_dir) for song in songs)
+    move_instructions = []
+
+    for song in songs:
+        absolute_song_dir = song.dir
+        song_dir = absolute_song_dir.relative_to(base_dir)
+        canonical_song_dir = _get_canonical_song_dir(song)
+
+        if song_dir not in song_dirs:
+            # A move instruction has already been generated for the dir this song is in
+            # (which is possible, because some dirs contain multiple (variants of) songs)
+            continue
+
+        if song_dir != canonical_song_dir:
+            # song_dir is not a good name, we want to replace it, so this path will soon be free
+            song_dirs.remove(song_dir)
+
+            # Find a canonical song dir variant that does not yet exist
+            variant = 1
+            while canonical_song_dir in song_dirs:
+                canonical_song_dir = _get_canonical_song_dir(song, variant)
+
+            song_dirs.add(canonical_song_dir)
+            move_instructions.append(
+                MoveInstruction(absolute_song_dir, base_dir / canonical_song_dir)
+            )
+
+    return tuple(move_instructions)