refactor(modal): reserved inline area for add-tag autocomplete

Replace the floating autocomplete popup with a reserved ~5-row inline
area below the add-tag input. It never covers provenance or suggestion
chips — they stay visible side-by-side with the autocomplete list while
the user types. The area scrolls when more matches are loaded, and is
prefilled with top tags on every image load so it's never empty.

The base .tag-autocomplete class (still used by bulk-select and the
fandom picker) keeps its floating behavior; the modal opts in to the
new inline layout via an added .tag-autocomplete-inline modifier.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

app/static/style.css

@@ -1076,6 +1076,24 @@ header {
 .tag-autocomplete.active {
   display: block;
 }
+/* Inline variant: the modal uses a reserved area that sits below the add-tag
+   input instead of floating over provenance/suggestions. Height fits ~5 rows;
+   overflow scrolls when more matches are loaded. */
+.tag-autocomplete-inline {
+  display: block;
+  position: static;
+  margin-top: 0.5rem;
+  background: var(--panel);
+  border: 1px solid rgba(255, 255, 255, 0.15);
+  border-radius: 6px;
+  height: 11.5rem;
+  max-height: none;
+  min-width: 0;
+  max-width: none;
+  overflow-y: auto;
+  box-shadow: none;
+  z-index: auto;
+}
 .tag-autocomplete-item {
   padding: 0.5rem 0.75rem;
   cursor: pointer;
@@ -1083,6 +1101,7 @@ header {
   align-items: center;
   gap: 0.5rem;
   font-size: 0.9rem;
+  line-height: 1.2;
   color: var(--text);
   border-bottom: 1px solid rgba(255, 255, 255, 0.05);
 }