From 137b18517f63b0bba3d39b4ced48ed4cfacb7984 Mon Sep 17 00:00:00 2001
From: lars <Lars.Noack@outlook.de>
Date: Mon, 24 Oct 2022 13:23:07 +0200
Subject: [PATCH] moved linguistics stuff to another file

---
 requirements.txt                                  |   3 ++-
 src/__pycache__/phonetic_compares.cpython-310.pyc | Bin 0 -> 196 bytes
 src/phonetic_compares.py                          |   6 ++++++
 src/youtube_music.py                              |  10 +++++-----
 4 files changed, 13 insertions(+), 6 deletions(-)
 create mode 100644 src/__pycache__/phonetic_compares.cpython-310.pyc

diff --git a/requirements.txt b/requirements.txt
index 823a91c..d525e14 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -3,4 +3,5 @@ pandas~=1.5.0
 mutagen~=1.46.0
 musicbrainzngs~=0.7.1
 jellyfish~=0.9.0
-pydub~=0.25.1
\ No newline at end of file
+pydub~=0.25.1
+youtube_dl
\ No newline at end of file
diff --git a/src/__pycache__/phonetic_compares.cpython-310.pyc b/src/__pycache__/phonetic_compares.cpython-310.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..8286e70db6bd2c436db3e93ccd84330eb0af37c2
GIT binary patch
literal 196
zcmYk0F%H5o3`J8|0JRGzVC?|8079GqA&@$-ppqDxlGGt?g?b{ymAW!<1twf3zT{`U
zzu1^06Tx}hGBxG>>g7KJmNgH>3@vnI#eumO>EO+c*rZW6F2UONY^ZpOFy!)w&rUkf
zCYLr3Br~*)1Gx=Y!nG!OGMAa}MX|VsjGAk_R~B;(K~hlC7g#}U)KQ`Hc>r4Ywi_)S
O>c+w<;xp7g7mF`qs55;4

literal 0
HcmV?d00001

diff --git a/src/phonetic_compares.py b/src/phonetic_compares.py
index a98f09c..556c563 100644
--- a/src/phonetic_compares.py
+++ b/src/phonetic_compares.py
@@ -1,2 +1,8 @@
 import jellyfish
 
+TITLE_THRESHOLD_LEVENSHTEIN = 1
+
+
+def match_titles(title_1: str, title_2: str) -> (bool, int):
+    distance = jellyfish.levenshtein_distance(title_1, title_2)
+    return distance > 1, distance
diff --git a/src/youtube_music.py b/src/youtube_music.py
index c469378..3f9ac99 100644
--- a/src/youtube_music.py
+++ b/src/youtube_music.py
@@ -1,9 +1,10 @@
 import youtube_dl
 import pandas as pd
-import jellyfish
 import logging
 import time
 
+import phonetic_compares
+
 YDL_OPTIONS = {'format': 'bestaudio', 'noplaylist': 'True'}
 YOUTUBE_URL_KEY = 'webpage_url'
 WAIT_BETWEEN_BLOCK = 10
@@ -31,12 +32,11 @@ def get_youtube_url(row):
     result = get_youtube_from_isrc(row['isrc'])
     video_title = result['title'].lower()
 
-    phonetic_distance = jellyfish.levenshtein_distance(real_title, video_title)
+    match, distance = phonetic_compares.match_titles(video_title, real_title)
 
-    print(real_title, video_title, phonetic_distance)
-    if phonetic_distance > 1:
+    if match:
         logging.warning(
-            f"dont downloading {result['url']} cuz the phonetic distance ({phonetic_distance}) between {real_title} and {video_title} is to high.")
+            f"dont downloading {result['url']} cuz the phonetic distance ({distance}) between {real_title} and {video_title} is to high.")
         return None
     return result['url']