fix db insert duplicate unique fields error

+ save album date as timestamp in db
+ Rewrite artist image downloader with multiprocessing.Pool

app/lib/artistlib.py

@@ -1,19 +1,19 @@
-from concurrent.futures import ThreadPoolExecutor
-from pathlib import Path
-from io import BytesIO
-
-from PIL import Image, UnidentifiedImageError
-import requests
 import urllib
+from concurrent.futures import ProcessPoolExecutor as Pool
+from io import BytesIO
+from multiprocessing import Pool, cpu_count
+from pathlib import Path
 
+import requests
+from PIL import Image, UnidentifiedImageError
+from requests.exceptions import ConnectionError as ReqConnError
+from requests.exceptions import ReadTimeout
 from tqdm import tqdm
-from requests.exceptions import ConnectionError as ReqConnError, ReadTimeout
 
 from app import settings
-from app.models import Artist, Track, Album
-from app.utils.hashing import create_hash
-
+from app.models import Album, Artist, Track
 from app.store import artists as artist_store
+from app.utils.hashing import create_hash
 
 
 def get_artist_image_link(artist: str):
@@ -74,15 +74,17 @@ class DownloadImage:
 
 class CheckArtistImages:
     def __init__(self):
-        with ThreadPoolExecutor() as pool:
-            list(
+        with Pool(cpu_count()) as pool:
+            res = list(
                 tqdm(
-                    pool.map(self.download_image, artist_store.ArtistStore.artists),
+                    pool.imap_unordered(self.download_image, artist_store.ArtistStore.artists),
                     total=len(artist_store.ArtistStore.artists),
                     desc="Downloading missing artist images",
                 )
             )
 
+            list(res)
+
     @staticmethod
     def download_image(artist: Artist):
         """
@@ -90,7 +92,9 @@ class CheckArtistImages:
 
         :param artist: The artist name
         """
-        img_path = Path(settings.Paths.get_artist_img_sm_path()) / f"{artist.artisthash}.webp"
+        img_path = (
+            Path(settings.Paths.get_artist_img_sm_path()) / f"{artist.artisthash}.webp"
+        )
 
         if img_path.exists():
             return
@@ -155,9 +159,7 @@ def get_albumartists(albums: list[Album]) -> set[str]:
     return artists
 
 
-def get_all_artists(
-        tracks: list[Track], albums: list[Album]
-) -> list[Artist]:
+def get_all_artists(tracks: list[Track], albums: list[Album]) -> list[Artist]:
     artists_from_tracks = get_artists_from_tracks(tracks=tracks)
     artist_from_albums = get_albumartists(albums=albums)
 

app/lib/taglib.py

@@ -1,6 +1,7 @@
 import datetime
 import os
 from io import BytesIO
+import pendulum
 
 from PIL import Image, UnidentifiedImageError
 from tinytag import TinyTag
@@ -64,11 +65,12 @@ def extract_thumb(filepath: str, webp_path: str) -> bool:
     return False
 
 
-def extract_date(date_str: str | None, timestamp: float) -> int:
+def extract_date(date_str: str | None) -> int | None:
     try:
-        return int(date_str.split("-")[0])
+        date = pendulum.parse(date_str, strict=False)
+        return int(date.timestamp())
     except Exception as e:
-        return datetime.datetime.fromtimestamp(timestamp).year
+        return None
 
 
 def get_tags(filepath: str):
@@ -141,7 +143,7 @@ def get_tags(filepath: str):
     tags.image = f"{tags.albumhash}.webp"
     tags.folder = win_replace_slash(os.path.dirname(filepath))
 
-    tags.date = extract_date(tags.year, last_mod)
+    tags.date = extract_date(tags.year) or int(last_mod)
     tags.filepath = win_replace_slash(filepath)
     tags.filetype = filetype
     tags.last_mod = last_mod

app/lib/watchdogg.py

@@ -131,7 +131,6 @@ def add_track(filepath: str) -> None:
     """
     # remove the track if it already exists
     TrackStore.remove_track_by_filepath(filepath)
-    db.remove_tracks_by_filepaths(filepath)
 
     # add the track to the database and store.
     tags = get_tags(filepath)