fix: split HTTP connect/read timeouts (Jellyfin build + qBit stats)

Every HTTP client passed an integer timeout to requests, applying the same
value to BOTH connect and read phases. A slow Jellyfin /Items page or qBit
/sync/maindata blew through the 10s read budget → ReadTimeoutError. Split
into a (connect=5s, read=60s default) tuple via shared http_timeout() helper.
The media index build worker uses a 180s read floor. Existing services with
low timeout_seconds benefit from bumping to 60+.

backend/src/media_library_viewer_api/clients/http_timeout.py

@@ -0,0 +1,36 @@
+"""Shared HTTP timeout helpers.
+
+``requests`` accepts a single integer timeout and applies it to BOTH the
+connect and read phases. For slow upstream services (large Jellyfin
+libraries, qBittorrent with many torrents), the read phase needs a much
+larger budget than connect. These helpers produce ``(connect, read)`` tuples
+so the two phases are decoupled.
+"""
+
+from __future__ import annotations
+
+#: Short connect timeout — fail fast on unreachable/dead hosts.
+DEFAULT_CONNECT_TIMEOUT = 5.0
+
+#: Generous read timeout — let slow responses complete.
+DEFAULT_READ_TIMEOUT = 60.0
+
+
+def http_timeout(
+    read_timeout: float | int | None = None,
+    connect_timeout: float = DEFAULT_CONNECT_TIMEOUT,
+) -> tuple[float, float]:
+    """Build a ``(connect, read)`` timeout tuple for ``requests``.
+
+    ``read_timeout`` is the per-response read budget (seconds). When omitted
+    or non-positive, :data:`DEFAULT_READ_TIMEOUT` applies.
+    """
+    effective_read = DEFAULT_READ_TIMEOUT
+    if read_timeout is not None:
+        try:
+            parsed = float(read_timeout)
+            if parsed > 0:
+                effective_read = parsed
+        except (TypeError, ValueError):
+            pass  # fall back to default on non-numeric input
+    return (connect_timeout, effective_read)