feat: fix content sharing — nginx proxy, file path resolution, catalog filtering

- Add /content and /dwn proxy locations to nginx config (both HTTP and HTTPS)
  so peer requests reach the backend instead of the SPA catch-all
- Update content_file_path() to check FileBrowser data dir as fallback when
  files aren't in the dedicated content/files/ directory
- Populate size_bytes from actual file metadata in content.add
- Filter out availability:nobody items from the public catalog endpoint

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

core/archipelago/src/content_server.rs

@@ -89,8 +89,26 @@ pub async fn save_catalog(data_dir: &Path, catalog: &ContentCatalog) -> Result<(
 }
 
 /// Get the full filesystem path for a content item.
+/// Checks the dedicated content/files/ directory first, then falls back to the
+/// FileBrowser data directory (where users manage files via the web UI).
 pub fn content_file_path(data_dir: &Path, item: &ContentItem) -> PathBuf {
-    data_dir.join(CONTENT_DIR).join(&item.filename)
+    // Strip leading slash from filename for path joining
+    let clean_name = item.filename.trim_start_matches('/');
+
+    // Primary: dedicated content directory
+    let primary = data_dir.join(CONTENT_DIR).join(clean_name);
+    if primary.exists() {
+        return primary;
+    }
+
+    // Fallback: FileBrowser data directory (users share files managed via FileBrowser)
+    let fb_path = data_dir.join("filebrowser").join(clean_name);
+    if fb_path.exists() {
+        return fb_path;
+    }
+
+    // Return primary path even if it doesn't exist (caller checks existence)
+    primary
 }
 
 /// Add a content item to the catalog.