From 4a632c1c94ba771ac5a17df28d39bbfd20184ba6 Mon Sep 17 00:00:00 2001
From: Bob Mottram <bob@libreserver.org>
Date: Fri, 20 May 2022 16:18:02 +0100
Subject: [PATCH] Convert ID3 metadata field on audio attachments into hashtags

---
 media.py | 32 ++++++++++++++++++++++++++++++++
 posts.py | 20 ++++++++++++++++++++
 2 files changed, 52 insertions(+)

diff --git a/media.py b/media.py
index 1f84f7a10..c5c61b195 100644
--- a/media.py
+++ b/media.py
@@ -162,6 +162,38 @@ def _spoof_meta_data(base_dir: str, nickname: str, domain: str,
         return
 
 
+def get_music_metadata(filename: str) -> {}:
+    """Returns metadata for a music file
+    """
+    result = None
+    try:
+        result = subprocess.run(['exiftool', '-v3', filename],
+                                stdout=subprocess.PIPE)
+    except BaseException as ex:
+        print('EX: get_music_metadata failed ' + str(ex))
+    if not result:
+        return {}
+    if not result.stdout:
+        return {}
+    try:
+        id3_lines = result.stdout.decode('utf-8').split('\n')
+    except BaseException:
+        print('EX: get_music_metadata unable to decode output')
+        return {}
+    fieldnames = (
+        'Title', 'Artist', 'Genre', 'Track', 'Album', 'Length', 'Band'
+    )
+    music_metadata = {}
+    for line in id3_lines:
+        for field in fieldnames:
+            if field + ' = ' in line:
+                field_value = line.split(field + ' = ')[1]
+                if '>' in field_value:
+                    field_value = field_value.split('>')[0].strip()
+                music_metadata[field.lower()] = field_value
+    return music_metadata
+
+
 def convert_image_to_low_bandwidth(image_filename: str) -> None:
     """Converts an image to a low bandwidth version
     """
diff --git a/posts.py b/posts.py
index 1780999d7..da64d85fd 100644
--- a/posts.py
+++ b/posts.py
@@ -32,6 +32,8 @@ from webfinger import webfinger_handle
 from httpsig import create_signed_header
 from siteactive import site_is_active
 from languages import understood_post_language
+from utils import valid_hash_tag
+from utils import get_audio_extensions
 from utils import get_summary_from_post
 from utils import get_user_paths
 from utils import invalid_ciphertext
@@ -70,6 +72,7 @@ from utils import remove_html
 from utils import dangerous_markup
 from utils import acct_dir
 from utils import local_actor_url
+from media import get_music_metadata
 from media import attach_media
 from media import replace_you_tube
 from media import replace_twitter
@@ -1423,6 +1426,23 @@ def _create_post_base(base_dir: str,
     else:
         mentioned_recipients = ''
 
+    # add hashtags from audio file ID3 tags, such as Artist, Album, etc
+    if attach_image_filename and media_type:
+        audio_types = get_audio_extensions()
+        music_metadata = None
+        for ext in audio_types:
+            if ext in media_type:
+                music_metadata = get_music_metadata(attach_image_filename)
+                break
+        if music_metadata:
+            for audio_tag, audio_value in music_metadata.items():
+                if audio_tag == 'title' or audio_tag == 'track':
+                    continue
+                audio_value = audio_value.title().replace(' ', '')
+                if valid_hash_tag(audio_value) and \
+                   '#' + audio_value not in content:
+                    content += ' #' + audio_value
+
     tags = []
     hashtags_dict = {}