ui: cover art in Beatport, Spotify e YouTube search

- core/beatport.py: BeatportTrack.image_url estratto da image.dynamic_uri (95x95)
- core/spotify_client.py: _track_to_dict include image_url (Spotify torna 3 taglie,
  prendo la ~64px). search_artist_discography inietta album.images anche negli
  album-tracks per la cover
- core/youtube_search.py: image_url = i.ytimg.com/vi/<id>/mqdefault.jpg
- webui/index.html: nuova <th class="col-cover"></th> nelle 3 tabelle
- webui/js/app.js: 3 renderTable inseriscono <td class="col-cover"><img>
  con loading=lazy, referrerpolicy=no-referrer e onerror fallback
- webui/css/style.css: .beatport-table .col-cover img { 40x40, border-radius 4px,
  object-fit: cover, bg fallback var(--bg-input) }

core/beatport.py

@@ -23,6 +23,7 @@ class BeatportTrack:
     artists: str      # es. "A, B & C" già formattato
     duration_sec: int
     beatport_id: int
+    image_url: str = ""   # URL cover art (default vuoto per retrocompatibilità test)
 
     @property
     def display(self) -> str:
@@ -150,6 +151,17 @@ def _format_artists(artists_field: object) -> str:
     return ", ".join(names[:-1]) + " & " + names[-1]
 
 
+def _extract_image_url(image_field: object, size: int = 95) -> str:
+    """Estrae URL cover art dall'oggetto `image` di Beatport.
+    Preferisce `dynamic_uri` sostituendo {w}x{h}, fallback a `uri` fisso."""
+    if not isinstance(image_field, dict):
+        return ""
+    dyn = image_field.get("dynamic_uri") or ""
+    if isinstance(dyn, str) and "{w}" in dyn and "{h}" in dyn:
+        return dyn.replace("{w}", str(size)).replace("{h}", str(size))
+    return image_field.get("uri") or ""
+
+
 def _parse_tracks(data: dict) -> list:
     """Trasforma i track dict di Beatport in BeatportTrack ordinati per posizione."""
     raw = _find_tracks_results(data)
@@ -164,6 +176,7 @@ def _parse_tracks(data: dict) -> list:
                 artists=_format_artists(item.get("artists")),
                 duration_sec=length_ms // 1000,
                 beatport_id=int(item.get("id") or 0),
+                image_url=_extract_image_url(item.get("image")),
             )
         except (TypeError, ValueError) as e:
             raise BeatportParseError(f"track[{i}] shape inattesa: {e}") from e

core/spotify_client.py

@@ -308,6 +308,12 @@ def search_tracks(token: str, query: str, limit: int = 50) -> list:
 
 def _track_to_dict(t: dict) -> dict:
     """Mappa il track object Spotify sul nostro schema uniforme."""
+    images = t.get("album", {}).get("images", []) or []
+    # Spotify torna 3 taglie ordinate large->small. Prendo la più piccola (~64px)
+    # se disponibile, altrimenti la prima che c'è.
+    image_url = ""
+    if images:
+        image_url = images[-1].get("url", "") or images[0].get("url", "")
     return {
         "id": t.get("id", ""),
         "url": t.get("external_urls", {}).get("spotify", ""),
@@ -315,6 +321,7 @@ def _track_to_dict(t: dict) -> dict:
         "artists": ", ".join(a.get("name", "") for a in t.get("artists", [])),
         "album": t.get("album", {}).get("name", ""),
         "duration_sec": int(t.get("duration_ms", 0)) // 1000,
+        "image_url": image_url,
     }
 
 
@@ -372,10 +379,12 @@ def search_artist_discography(token: str, artist_name: str) -> list:
     r_alb.raise_for_status()
     albums = r_alb.json().get("items", [])
 
-    # 4. Per ogni album, tracce (album/track object non ha "album" sub-field, iniettiamola)
+    # 4. Per ogni album, tracce (album/track object non ha "album" sub-field, iniettiamola
+    # includendo anche le images dell'album così _track_to_dict riesce a estrarre la cover)
     for alb in albums:
         alb_id = alb.get("id")
         alb_name = alb.get("name", "")
+        alb_images = alb.get("images", []) or []
         if not alb_id:
             continue
         time.sleep(0.1)  # rate limit interno
@@ -388,8 +397,7 @@ def search_artist_discography(token: str, artist_name: str) -> list:
         r_at.raise_for_status()
         for t in r_at.json().get("items", []):
             t = dict(t)
-            # Album tracks non hanno "album" nested; iniettiamo il nome
-            t.setdefault("album", {"name": alb_name})
+            t.setdefault("album", {"name": alb_name, "images": alb_images})
             collected.append(_track_to_dict(t))
 
     # 5. Dedupe

core/youtube_search.py

@@ -62,11 +62,13 @@ def search_youtube(query: str, limit: int = 50) -> list:
             continue
         video_id = e.get("id") or ""
         url = e.get("url") or (f"https://www.youtube.com/watch?v={video_id}" if video_id else "")
+        image_url = f"https://i.ytimg.com/vi/{video_id}/mqdefault.jpg" if video_id else ""
         result.append({
             "id": video_id,
             "url": url,
             "title": e.get("title") or "",
             "channel": e.get("uploader") or e.get("channel") or "",
             "duration_sec": int(e.get("duration") or 0),
+            "image_url": image_url,
         })
     return result