internal/archival/archiver/yt_dlp.go

@@ -7,9 +7,11 @@ import (
 	"encoding/json"
 	"fmt"
 	"io"
+	"log/slog"
 	"os"
 	"os/exec"
 	"path/filepath"
+	"regexp"
 	"strings"
 	"time"
 
@@ -23,6 +25,9 @@ type YtDlpConfig struct {
 	WriteThumbnail bool   `json:"write_thumbnail"`
 	WriteSubs      bool   `json:"write_subs"` // Download subtitles as separate files
 	EmbedSubs      bool   `json:"embed_subs"` // Embed subtitles into the video file
+	// IgnoreErrors archives the video even when other assets (subtitles, thumbnail)
+	// fail; those failures are logged as warnings instead of failing the archival.
+	IgnoreErrors bool `json:"ignore_errors"`
 }
 
 // ToMap converts the YtDlpConfig to a map[string]any for use with ApplyConfig
@@ -52,6 +57,7 @@ func NewYtDlpArchiver() *YtDlpArchiver {
 			WriteThumbnail: true,
 			WriteSubs:      true,
 			EmbedSubs:      true,
+			IgnoreErrors:   true,
 		},
 	}
 }
@@ -93,6 +99,7 @@ func (a *YtDlpArchiver) GetDefaultConfig() any {
 		WriteThumbnail: true,
 		WriteSubs:      true,
 		EmbedSubs:      true,
+		IgnoreErrors:   true,
 	}
 }
 
@@ -113,6 +120,9 @@ func (a *YtDlpArchiver) ApplyConfig(config map[string]any) error {
 	if v, ok := config["embed_subs"].(bool); ok {
 		a.config.EmbedSubs = v
 	}
+	if v, ok := config["ignore_errors"].(bool); ok {
+		a.config.IgnoreErrors = v
+	}
 	return nil
 }
 
@@ -164,9 +174,11 @@ func (a *YtDlpArchiver) Archive(ctx context.Context, link *model.Link, stor stor
 	outputTemplate := filepath.Join(tempDir, "%(title)s.%(ext)s")
 	args := []string{
 		"-o", outputTemplate,
-		"--no-warnings",
 		"--no-playlist",
 	}
+	if a.config.IgnoreErrors {
+		args = append(args, "--ignore-errors")
+	}
 	if a.config.WriteThumbnail {
 		args = append(args, "--write-thumbnail")
 	}
@@ -186,12 +198,17 @@ func (a *YtDlpArchiver) Archive(ctx context.Context, link *model.Link, stor stor
 	cmd.Stdout = &combinedOutput
 	cmd.Stderr = &combinedOutput
 
-	if err := cmd.Run(); err != nil {
-		errMsg := err.Error()
-		if exitErr, ok := err.(*exec.ExitError); ok {
+	runErr := cmd.Run()
+	output := strings.TrimSpace(combinedOutput.String())
+	if runErr != nil && (!a.config.IgnoreErrors || !hasMediaFile(tempDir)) {
+		errMsg := runErr.Error()
+		if exitErr, ok := runErr.(*exec.ExitError); ok {
 			errMsg = fmt.Sprintf("exit code %d", exitErr.ExitCode())
 		}
-		return nil, fmt.Errorf("yt-dlp failed (%s): %s", errMsg, strings.TrimSpace(combinedOutput.String()))
+		return nil, fmt.Errorf("yt-dlp failed (%s): %s", errMsg, output)
+	}
+	if problems := ytDlpProblems(output); len(problems) > 0 {
+		slog.Warn("yt-dlp archived the video with missing assets", "url", link.URL, "problems", problems)
 	}
 
 	var files []FileInfo
@@ -239,7 +256,7 @@ func (a *YtDlpArchiver) Archive(ctx context.Context, link *model.Link, stor stor
 	}
 
 	if len(files) == 0 {
-		return nil, fmt.Errorf("yt-dlp produced no files. output: %s", strings.TrimSpace(combinedOutput.String()))
+		return nil, fmt.Errorf("yt-dlp produced no files. output: %s", output)
 	}
 
 	if title == "" {
@@ -269,3 +286,36 @@ func mimeTypeFromExt(ext string) string {
 func isImageMimeType(mimeType string) bool {
 	return strings.HasPrefix(mimeType, "image/")
 }
+
+// ytDlpFormatPart matches the per-format streams (title.f400.mp4, title.f251-drc.webm)
+// that yt-dlp keeps only when merging them into the final file failed.
+// ponytail: numeric format ids only (YouTube), extend for extractors with named ids like fhls-720p.
+var ytDlpFormatPart = regexp.MustCompile(`\.f\d+[^.]*\.[^.]+$`)
+
+func hasMediaFile(dir string) bool {
+	entries, err := os.ReadDir(dir)
+	if err != nil {
+		return false
+	}
+	for _, entry := range entries {
+		if ytDlpFormatPart.MatchString(entry.Name()) {
+			continue
+		}
+		mimeType := mimeTypeFromExt(filepath.Ext(entry.Name()))
+		if strings.HasPrefix(mimeType, "video/") || strings.HasPrefix(mimeType, "audio/") {
+			return true
+		}
+	}
+	return false
+}
+
+// ytDlpProblems returns the WARNING and ERROR lines of yt-dlp output.
+func ytDlpProblems(output string) []string {
+	var problems []string
+	for _, line := range strings.Split(output, "\n") {
+		if strings.HasPrefix(line, "WARNING:") || strings.HasPrefix(line, "ERROR:") {
+			problems = append(problems, line)
+		}
+	}
+	return problems
+}

internal/archival/archiver/yt_dlp_test.go

@@ -1,6 +1,8 @@
 package archiver
 
 import (
+	"os"
+	"path/filepath"
 	"testing"
 )
 
@@ -127,3 +129,48 @@ func TestYtDlpConfig_ToMap(t *testing.T) {
 		t.Errorf("Expected embed_subs true, got %v", m["embed_subs"])
 	}
 }
+
+func TestYtDlpArchiver_IgnoreErrorsDefaultAndConfig(t *testing.T) {
+	a := NewYtDlpArchiver()
+	if !a.config.IgnoreErrors || !a.GetDefaultConfig().(YtDlpConfig).IgnoreErrors {
+		t.Fatal("Expected IgnoreErrors to be true by default")
+	}
+	if err := a.ApplyConfig(map[string]any{"ignore_errors": false}); err != nil {
+		t.Fatalf("ApplyConfig failed: %v", err)
+	}
+	if a.config.IgnoreErrors {
+		t.Error("Expected IgnoreErrors to be false after config")
+	}
+}
+
+func TestHasMediaFile(t *testing.T) {
+	dir := t.TempDir()
+	if err := os.WriteFile(filepath.Join(dir, "video.en.vtt"), nil, 0o600); err != nil {
+		t.Fatal(err)
+	}
+	if hasMediaFile(dir) {
+		t.Error("Expected no media file with only subtitles")
+	}
+	for _, name := range []string{"video.f400.mp4", "video.f251-drc.webm", "video.mp4.part"} {
+		if err := os.WriteFile(filepath.Join(dir, name), nil, 0o600); err != nil {
+			t.Fatal(err)
+		}
+	}
+	if hasMediaFile(dir) {
+		t.Error("Expected no media file with only unmerged streams and partial downloads")
+	}
+	if err := os.WriteFile(filepath.Join(dir, "video.webm"), nil, 0o600); err != nil {
+		t.Fatal(err)
+	}
+	if !hasMediaFile(dir) {
+		t.Error("Expected a media file after adding a video")
+	}
+}
+
+func TestYtDlpProblems(t *testing.T) {
+	output := "[youtube] abc: Downloading webpage\nWARNING: Unable to download video subtitles for 'en': HTTP Error 429\nERROR: Postprocessing: something\n[info] done"
+	problems := ytDlpProblems(output)
+	if len(problems) != 2 {
+		t.Fatalf("Expected 2 problems, got %d: %v", len(problems), problems)
+	}
+}