From 32e79c46d8f082e28668d2257237a09dab53214c Mon Sep 17 00:00:00 2001
From: joren <jorensc@riseup.net>
Date: Sun, 23 Feb 2025 02:09:23 +0100
Subject: [PATCH] Add script

---
 deduper.sh | 106 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 106 insertions(+)
 create mode 100755 deduper.sh

diff --git a/deduper.sh b/deduper.sh
new file mode 100755
index 0000000..ddbb23b
--- /dev/null
+++ b/deduper.sh
@@ -0,0 +1,106 @@
+#!/bin/bash
+
+LOGFILE="dupe_cleanup.log"
+ARMED=false  
+
+
+if [[ "$1" == "--armed" ]]; then
+    ARMED=true
+fi
+
+
+log() {
+    echo "$1" | tee -a "$LOGFILE"
+}
+
+
+get_quality() {
+    file="$1"
+    bitrate=$(mediainfo --Output="Audio;%BitRate%" "$file" | awk '{print $1}')
+    bitdepth=$(mediainfo --Output="Audio;%BitDepth%" "$file" | awk '{print $1}')
+    
+    
+    bitrate=${bitrate:-0}
+    bitdepth=${bitdepth:-0}
+    
+    echo "$bitrate $bitdepth"
+}
+
+
+echo "===============================" | tee -a "$LOGFILE"
+log "Starting duplicate cleanup at $(date)"
+if ! $ARMED; then
+    log "DRY-RUN MODE: No files will be renamed or deleted."
+else
+    log "ARMED MODE: Files **will** be renamed and deleted!"
+fi
+echo "===============================" | tee -a "$LOGFILE"
+
+
+find . -type f -regextype posix-extended -regex ".*\([0-9]\)\.flac" | while read -r dup_file; do
+    
+    orig_file=$(echo "$dup_file" | sed -E 's/ \([0-9]+\)\.flac/.flac/')
+
+    
+    candidates=()
+    if [[ -f "$orig_file" ]]; then
+        candidates+=("$orig_file")
+    fi
+
+    
+    for n in {1..9}; do
+        candidate=$(echo "$orig_file" | sed "s/.flac$/ ($n).flac/")
+        if [[ -f "$candidate" ]]; then
+            candidates+=("$candidate")
+        fi
+    done
+
+    
+    if [[ ${#candidates[@]} -le 1 ]]; then
+        continue
+    fi
+
+    best_file=""
+    best_bitrate=0
+    best_bitdepth=0
+
+    log "Checking duplicates for: $orig_file"
+
+    
+    for file in "${candidates[@]}"; do
+        read bitrate bitdepth < <(get_quality "$file")
+        log "  - Found: $file (Bitrate: $bitrate, Bitdepth: $bitdepth)"
+
+        if (( bitrate > best_bitrate )) || { (( bitrate == best_bitrate )) && (( bitdepth > best_bitdepth )); }; then
+            best_bitrate=$bitrate
+            best_bitdepth=$bitdepth
+            best_file="$file"
+        fi
+    done
+
+    log "  -> Keeping: $best_file (Bitrate: $best_bitrate, Bitdepth: $best_bitdepth)"
+
+    
+    if [[ "$best_file" != "$orig_file" ]]; then
+        log "  * Would rename: $best_file -> $orig_file"
+        if $ARMED; then
+            mv "$best_file" "$orig_file"
+        fi
+    fi
+
+    
+    for file in "${candidates[@]}"; do
+        if [[ "$file" != "$best_file" ]]; then
+            log "  * Would delete: $file"
+            if $ARMED; then
+                rm "$file"
+            fi
+        fi
+    done
+
+    echo "" | tee -a "$LOGFILE"
+
+done
+
+log "Cleanup completed at $(date)"
+