feat: enrich RSS items without IDs via TMDB title search

Items from RSS feeds that have no media ID (tmdb/imdb/tvdb) are now
enriched automatically using TMDB's search API. The release name is
parsed to extract a clean title and year, then searched against TMDB
to retrieve the TMDB ID before validation.

- Add media/tmdb package with SearchMovies, SearchShows, and
  ExtractTitleAndYear (parses torrent release names)
- Add EnrichFeedItemWithTmdbId to media.Client, called in rss/process.go
  before the ID validation switch
- Add --run-now flag to nabarr run to trigger all feeds immediately
- Wire media.Client through rss.Client and rssJob
- Merge feature/add-tag-option (tag support for Sonarr/Radarr)

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

media/tmdb/parse.go

@@ -0,0 +1,25 @@
+package tmdb
+
+import (
+	"regexp"
+	"strconv"
+	"strings"
+)
+
+var yearRegex = regexp.MustCompile(`\b(19|20)\d{2}\b`)
+
+// ExtractTitleAndYear parses a torrent release name and returns the clean title and year.
+// Example: "The.Housemaid.2025.FRENCH.1080p" → ("The Housemaid", 2025)
+func ExtractTitleAndYear(releaseName string) (string, int) {
+	s := strings.ReplaceAll(releaseName, ".", " ")
+	s = strings.ReplaceAll(s, "_", " ")
+
+	loc := yearRegex.FindStringIndex(s)
+	if loc == nil {
+		return strings.TrimSpace(s), 0
+	}
+
+	year, _ := strconv.Atoi(s[loc[0]:loc[1]])
+	title := strings.TrimSpace(s[:loc[0]])
+	return title, year
+}