feat(metadata): audio tag rewrite via ffmpeg + CBZ ComicInfo injection

Phases 5b and 5c of the metadata writer. 5b — single-file audio (MP3/FLAC/M4A/M4B): the container's tags are rewritten with ffmpeg -c copy (no re-encode; -map 0 keeps cover art, -map_chapters 0 keeps chapters). The key set is exactly what the scanner's applyAudioTagFallbacks reads back — title+album, subtitle, artist/album_artist, composer (narrators, Audible's convention), grouping + iTunes series-part, date, genre — so a write + rescan round-trips. Output is ffprobe-VALIDATED (duration within 1s/0.5% of the source) before the atomic rename; a lossy container rewrite never replaces a good book. Fields the format can't carry per our own read chain are deliberately not written, and fields_written reports only what really round-trips (subtitle rides MP3/FLAC — this image's mp4 muxer drops subtitle keys, discovered empirically and pinned by test). 5c — CBZ: ComicInfo.xml injected (or replaced) with the same atomic zip rewrite as the EPUB path, extended with an append-if-missing mode since a comic may not carry one yet. Field map mirrors the scanner's comic block: Title/Series/Number/Count/Volume, Writer (authors), Summary, Publisher, Year/Month/Day, LanguageISO, Web, Genre+Tags; fractional series numbers are omitted rather than truncated to a lie. CBR/CB7 stay on the sidecar path (RAR/7z can't be rewritten in place). Integration tests generate media at test runtime with ffmpeg (m4b) and in-memory zips (cbz) — no binary fixtures — pinning the tag round-trip through ffprobe, post-write rescan titles, the ComicInfo injection and the validation-gate behavior.
2026-10-27 16:04:34 -04:00
1 parent eaf1b2c8dd
commit 59eddc1386
@@ -9,7 +9,9 @@ import (
 	"io"
 	"net/http"
 	"os"
+	"os/exec"
 	"path/filepath"
+	"strconv"
 	"strings"
 	"testing"

@@ -219,6 +221,157 @@ func TestMetadataWriteToFile(t *testing.T) {
 		_ = itemB
 	})

+	t.Run("m4b tags are rewritten via ffmpeg and survive rescan", func(t *testing.T) {
+		rel := fmt.Sprintf("test-metadata-write-%d/audio/book.m4b", os.Getpid())
+		abs := filepath.Join(uploadsDir, rel)
+		require.NoError(t, os.MkdirAll(filepath.Dir(abs), 0o755))
+		// Generate the fixture at test runtime (AGENTS: no binary media
+		// in git): 2s of silence through the native AAC encoder, with
+		// an original title tag to prove the rewrite replaces it.
+		gen := exec.Command("ffmpeg", "-hide_banner", "-loglevel", "error", "-y",
+			"-f", "lavfi", "-i", "anullsrc=r=44100:cl=mono",
+			"-t", "2", "-c:a", "aac",
+			"-metadata", "title=Generated Fixture",
+			"-metadata", "album=Generated Fixture",
+			abs)
+		out, err := gen.CombinedOutput()
+		require.NoError(t, err, "ffmpeg fixture generation failed: %s", string(out))
+
+		var itemID string
+		err = setup.DBPool.QueryRow(ctx,
+			`INSERT INTO media_items (library_id, title, file_path, format_group, mime_type, file_size, created_at, imported_at, updated_at)
+			 VALUES ($1, 'Generated Fixture', $2, 'audiobook', 'audio/mp4', 100, NOW(), NOW(), NOW()) RETURNING id`,
+			libID, rel,
+		).Scan(&itemID)
+		require.NoError(t, err)
+
+		// Follow the real editor flow: PUT first (records
+		// metadata_overrides, so the inline rescan keeps the edited
+		// values instead of folder-structure derivations).
+		editItem(itemID, map[string]interface{}{
+			"title":          "Audio New Title",
+			"subtitle":       "Audio Sub",
+			"authors":        []string{"Audio Author"},
+			"narrators":      []string{"Narrator Person"},
+			"series":         "Audio Series",
+			"series_number":  1.5,
+			"date_published": "1953-04-16",
+			"tags":           []string{"Science"},
+		})
+
+		code, body := callWrite(setup.Token, itemID)
+		require.Equal(t, http.StatusOK, code)
+		assert.Equal(t, "audio-tags", body["metadata_target"])
+
+		// The container now carries the round-trip keys (probe with the
+		// same tool the scanner uses).
+		probe := func() map[string]string {
+			out, err := exec.Command("ffprobe", "-v", "error", "-show_entries",
+				"format_tags", "-of", "json", abs).Output()
+			require.NoError(t, err)
+			var parsed struct {
+				Format struct {
+					Tags map[string]string `json:"tags"`
+				} `json:"format"`
+			}
+			require.NoError(t, json.Unmarshal(out, &parsed))
+			return parsed.Format.Tags
+		}
+		tags := probe()
+		assert.Equal(t, "Audio New Title", tags["album"], "the scanner's album rung")
+		assert.Equal(t, "Audio New Title", tags["title"])
+		// This ffmpeg's mp4 muxer drops subtitle keys — nothing that
+		// can't be read back may be written (fields_written omits it).
+		assert.Empty(t, tags["subtitle"], "subtitle must not be written where it can't round-trip")
+		assert.Equal(t, "Audio Author", tags["artist"])
+		assert.Equal(t, "Narrator Person", tags["composer"])
+		assert.Equal(t, "Audio Series", tags["grouping"])
+		assert.Equal(t, "Science", tags["genre"])
+		assert.NotContains(t, tags["album"], "Generated Fixture", "old tag replaced")
+
+		writeMap := body["write"].(map[string]interface{})
+		assert.NotContains(t, writeMap["fields_written"], "subtitle", "honest fields_written")
+
+		// Still a valid, same-length audio file after the rewrite.
+		dur := func() float64 {
+			out, err := exec.Command("ffprobe", "-v", "error", "-show_entries",
+				"format=duration", "-of", "default=noprint_wrappers=1:nokey=1", abs).Output()
+			require.NoError(t, err)
+			v, err := strconv.ParseFloat(strings.TrimSpace(string(out)), 64)
+			require.NoError(t, err)
+			return v
+		}
+		assert.Greater(t, dur(), 1.5, "audio survived the rewrite")
+
+		// The inline rescan re-read the tags: DB title stays the written one.
+		var title string
+		require.NoError(t, setup.DBPool.QueryRow(ctx,
+			`SELECT title FROM media_items WHERE id = $1`, itemID).Scan(&title))
+		assert.Equal(t, "Audio New Title", title)
+	})
+
+	t.Run("CBZ gets ComicInfo.xml injected and rescan keeps it", func(t *testing.T) {
+		rel := fmt.Sprintf("test-metadata-write-%d/comic/book.cbz", os.Getpid())
+		abs := filepath.Join(uploadsDir, rel)
+		require.NoError(t, os.MkdirAll(filepath.Dir(abs), 0o755))
+		// Minimal archive: one "page" file (a text stub — cover
+		// extraction failing is tolerated by the scanner) and no
+		// ComicInfo.xml yet.
+		var buf bytes.Buffer
+		w := zip.NewWriter(&buf)
+		page, _ := w.Create("page_0001.jpg")
+		_, _ = page.Write([]byte("not-a-real-jpeg"))
+		require.NoError(t, w.Close())
+		require.NoError(t, os.WriteFile(abs, buf.Bytes(), 0o644))
+
+		var itemID string
+		err := setup.DBPool.QueryRow(ctx,
+			`INSERT INTO media_items (library_id, title, file_path, format_group, mime_type, file_size, created_at, imported_at, updated_at)
+			 VALUES ($1, 'Generated Comic', $2, 'comic_archive', 'application/zip', 100, NOW(), NOW(), NOW()) RETURNING id`,
+			libID, rel,
+		).Scan(&itemID)
+		require.NoError(t, err)
+
+		editItem(itemID, map[string]interface{}{
+			"title":         "Comic New Title",
+			"series":        "Comic Series",
+			"series_number": 4,
+			"authors":       []string{"Comic Writer"},
+			"tags":          []string{"Superhero"},
+		})
+
+		code, body := callWrite(setup.Token, itemID)
+		require.Equal(t, http.StatusOK, code)
+		assert.Equal(t, "comicinfo", body["metadata_target"])
+
+		reader, err := zip.OpenReader(abs)
+		require.NoError(t, err)
+		var infoXML string
+		for _, f := range reader.File {
+			if f.Name == "ComicInfo.xml" {
+				rc, err := f.Open()
+				require.NoError(t, err)
+				data, err := io.ReadAll(rc)
+				require.NoError(t, rc.Close())
+				require.NoError(t, err)
+				infoXML = string(data)
+			}
+		}
+		reader.Close()
+		assert.Contains(t, infoXML, "<Title>Comic New Title</Title>")
+		assert.Contains(t, infoXML, "<Series>Comic Series</Series>")
+		assert.Contains(t, infoXML, "<Number>4</Number>")
+		assert.Contains(t, infoXML, "<Writer>Comic Writer</Writer>")
+
+		// The inline rescan reads the ComicInfo back: the DB title and
+		// series stay the written values.
+		var title, series string
+		require.NoError(t, setup.DBPool.QueryRow(ctx,
+			`SELECT title, series FROM media_items WHERE id = $1`, itemID).Scan(&title, &series))
+		assert.Equal(t, "Comic New Title", title)
+		assert.Equal(t, "Comic Series", series)
+	})
+
 	t.Cleanup(func() {
 		_ = os.RemoveAll(filepath.Join(uploadsDir, fmt.Sprintf("test-metadata-write-%d", os.Getpid())))
 	})