From d678c8fd0ff14bee4d3de4addd3456282d704c78 Mon Sep 17 00:00:00 2001
From: Tanner Collin <git@tannercollin.com>
Date: Mon, 3 Apr 2023 21:04:37 -0600
Subject: [PATCH 1/4] Add audio hashing

---
 main.py | 22 ++++++++++++++++++++--
 1 file changed, 20 insertions(+), 2 deletions(-)

diff --git a/main.py b/main.py
index f949256..6055557 100644
--- a/main.py
+++ b/main.py
@@ -3,6 +3,7 @@ import os
 import hashlib
 from PIL import Image, UnidentifiedImageError
 
+import acoustid
 from imagehash import average_hash
 from videohash import VideoHash
 from videohash.exceptions import FFmpegFailedToExtractFrames
@@ -49,6 +50,12 @@ def get_image_hash(filename):
     except UnidentifiedImageError:
         return None
 
+def get_audio_hash(filename):
+    try:
+        return str(acoustid.fingerprint_file(filename)[1].decode()) + '_audio'
+    except acoustid.FingerprintGenerationError:
+        return None
+
 def get_video_hash(filename):
     try:
         return str(VideoHash(path=filename).hash_hex) + '_video'
@@ -65,7 +72,10 @@ for filename in filenames:
     if filename in delete: continue
 
     try:
-        digest = get_image_hash(filename) or get_video_hash(filename)
+        digest = get_image_hash(filename) or get_audio_hash(filename) or get_video_hash(filename)
+    except KeyboardInterrupt:
+        print('Skipping media hashing.')
+        break
     except BaseException as e:
         print()
         print('Exception', e.__class__.__name__, str(e), 'while hashing:')
@@ -79,8 +89,16 @@ for filename in filenames:
     else:
         hashes.add(digest)
 
+for dupe in delete:
+    print(dupe)
 print()
-print('Found', len(delete), 'total duplicate files.')
+print('Found', len(delete), 'total duplicate files. Delete them?')
+print('ENTER to continue, ctrl+c to cancel.')
+try:
+    input()
+except KeyboardInterrupt:
+    print('\nCancelled.')
+    os._exit(0)
 
 print('Deleting...')
 for dupe in delete:

From dddfbb07247801b4bbcead7446362ef37131cf12 Mon Sep 17 00:00:00 2001
From: Tanner Collin <git@tannercollin.com>
Date: Mon, 3 Apr 2023 22:11:34 -0600
Subject: [PATCH 2/4] Compare bits on audio hashes

---
 main.py | 23 +++++++++++++++++++----
 1 file changed, 19 insertions(+), 4 deletions(-)

diff --git a/main.py b/main.py
index 6055557..0ec2daa 100644
--- a/main.py
+++ b/main.py
@@ -4,11 +4,13 @@ import hashlib
 from PIL import Image, UnidentifiedImageError
 
 import acoustid
+import chromaprint
 from imagehash import average_hash
 from videohash import VideoHash
 from videohash.exceptions import FFmpegFailedToExtractFrames
 
 hashes = set()
+audio_hashes = []
 delete = set()
 
 filenames = [x for x in glob.glob('**', recursive=True) if os.path.isfile(x)]
@@ -52,7 +54,9 @@ def get_image_hash(filename):
 
 def get_audio_hash(filename):
     try:
-        return str(acoustid.fingerprint_file(filename)[1].decode()) + '_audio'
+        fp = acoustid.fingerprint_file(filename)[1]
+        values, _ = chromaprint.decode_fingerprint(fp)
+        return chromaprint.hash_fingerprint(values)
     except acoustid.FingerprintGenerationError:
         return None
 
@@ -84,13 +88,24 @@ for filename in filenames:
 
     if not digest: continue
 
-    if digest in hashes:
-        delete.add(filename)
+    if type(digest) == int:
+        for h in audio_hashes:
+            if bin(digest ^ h).count('1') <= 5:   # TODO adjust?
+                delete.add(filename)
+                break
+        else:  # for
+            audio_hashes.append(digest)
     else:
-        hashes.add(digest)
+        if digest in hashes:
+            delete.add(filename)
+        else:
+            hashes.add(digest)
 
+print()
+print()
 for dupe in delete:
     print(dupe)
+
 print()
 print('Found', len(delete), 'total duplicate files. Delete them?')
 print('ENTER to continue, ctrl+c to cancel.')

From 230460c1539a5964047f97403b3ff915b9974fe2 Mon Sep 17 00:00:00 2001
From: Tanner Collin <git@tannercollin.com>
Date: Mon, 3 Apr 2023 22:11:53 -0600
Subject: [PATCH 3/4] Freeze requirements

---
 .gitignore       |  2 +-
 requirements.txt | 19 +++++++++++++++++++
 2 files changed, 20 insertions(+), 1 deletion(-)
 create mode 100644 requirements.txt

diff --git a/.gitignore b/.gitignore
index 28d346b..afef025 100644
--- a/.gitignore
+++ b/.gitignore
@@ -102,4 +102,4 @@ ENV/
 *.swp
 *.swo
 
-mcdata/
+test/
diff --git a/requirements.txt b/requirements.txt
new file mode 100644
index 0000000..5192096
--- /dev/null
+++ b/requirements.txt
@@ -0,0 +1,19 @@
+audioread==3.0.0
+Brotli==1.0.9
+certifi==2022.12.7
+charset-normalizer==3.1.0
+idna==3.4
+imagedominantcolor==1.0.1
+ImageHash==4.3.1
+mutagen==1.46.0
+numpy==1.24.2
+Pillow==9.5.0
+pyacoustid==1.2.2
+pycryptodomex==3.17
+PyWavelets==1.4.1
+requests==2.28.2
+scipy==1.10.1
+urllib3==1.26.15
+videohash==3.0.1
+websockets==11.0
+yt-dlp==2023.3.4

From c226c3807ad069a750375ffb34e2d2465e459f5b Mon Sep 17 00:00:00 2001
From: Tanner Collin <git@tannercollin.com>
Date: Tue, 17 Oct 2023 13:49:24 -0600
Subject: [PATCH 4/4] IDK

---
 main.py | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/main.py b/main.py
index 0ec2daa..5e75a8d 100644
--- a/main.py
+++ b/main.py
@@ -12,6 +12,7 @@ from videohash.exceptions import FFmpegFailedToExtractFrames
 hashes = set()
 audio_hashes = []
 delete = set()
+hash_lookup = {}
 
 filenames = [x for x in glob.glob('**', recursive=True) if os.path.isfile(x)]
 
@@ -88,10 +89,14 @@ for filename in filenames:
 
     if not digest: continue
 
+    hash_lookup[digest] = filename
+
     if type(digest) == int:
         for h in audio_hashes:
             if bin(digest ^ h).count('1') <= 5:   # TODO adjust?
                 delete.add(filename)
+                print()
+                print(digest, filename, 'close to', h, hash_lookup[h])
                 break
         else:  # for
             audio_hashes.append(digest)