feat: add acoustic dedupe pipeline and docker runtime

am_deduper/metadata.py

@@ -1,24 +1,33 @@
 from __future__ import annotations
 
+import re
 from pathlib import Path
 
+from .constants import MUSIC_EXTENSIONS, VERSION_MARKERS
+
 try:
     from tinytag import TinyTag
-except ImportError:  # optional until metadata phase is invoked
+except ImportError:
     TinyTag = None
 
 
 def read_metadata(path: Path) -> dict:
+    data = {"format": path.suffix.lower().lstrip(".")}
     if TinyTag is None:
-        return {}
-    tag = TinyTag.get(str(path), image=False)
-    return {
+        return data
+    try:
+        tag = TinyTag.get(str(path), image=False)
+    except Exception:
+        return data
+    data.update({
         "duration": tag.duration,
         "artist": tag.artist,
         "title": tag.title,
         "album": tag.album,
-        "format": path.suffix.lower().lstrip("."),
         "bitrate": int(tag.bitrate) if tag.bitrate else None,
         "sample_rate": int(tag.samplerate) if tag.samplerate else None,
         "bit_depth": int(tag.bitdepth) if tag.bitdepth else None,
-    }
+    })
+    text = " ".join(str(data.get(k) or "") for k in ("title", "album", "file_name"))
+    data["version_flags"] = [m for m in VERSION_MARKERS if re.search(rf"(?<!\w){re.escape(m)}(?!\w)", text, re.I)]
+    return data