Fixed a path-duplication bug and added regression coverage

internal/artifacts/paths.go

@@ -4,6 +4,7 @@ import "path/filepath"
 
 // SessionPaths contains canonical local paths for one session work directory.
 type SessionPaths struct {
+	WorkspaceRoot            string
 	Root                     string
 	InputsDir                string
 	AudioDir                 string
@@ -26,6 +27,7 @@ func buildSessionPaths(workspaceRoot, sessionID string) SessionPaths {
 	root := SessionWorkDir(workspaceRoot, sessionID)
 	transcripts := filepath.Join(root, "transcripts")
 	return SessionPaths{
+		WorkspaceRoot:            workspaceRoot,
 		Root:                     root,
 		InputsDir:                filepath.Join(root, "inputs"),
 		AudioDir:                 filepath.Join(root, "audio"),

internal/artifacts/resolve.go

@@ -0,0 +1,58 @@
+package artifacts
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+)
+
+// ResolveSessionLocalPathForRead resolves manifest/local artifact paths for read use-cases.
+// Invariants for stage consumers:
+//   - local paths may be absolute, workspace-root qualified, or session-workdir relative.
+//   - callers should resolve through this helper instead of manually joining session/workdir roots.
+func ResolveSessionLocalPathForRead(paths SessionPaths, localPath string) string {
+	p := filepath.Clean(strings.TrimSpace(localPath))
+	if p == "" {
+		return ""
+	}
+	if filepath.IsAbs(p) {
+		return p
+	}
+
+	// Already qualified against known roots.
+	if underRoot(p, paths.WorkspaceRoot) || underRoot(p, paths.Root) {
+		return p
+	}
+
+	candidates := []string{
+		p,
+		filepath.Clean(filepath.Join(paths.WorkspaceRoot, p)),
+		filepath.Clean(filepath.Join(paths.Root, p)),
+	}
+	for _, c := range candidates {
+		if c == "" {
+			continue
+		}
+		if _, err := os.Stat(c); err == nil {
+			return c
+		}
+	}
+
+	// Deterministic fallback for ambiguous relative values.
+	if strings.TrimSpace(paths.WorkspaceRoot) != "" {
+		return candidates[1]
+	}
+	return candidates[2]
+}
+
+func underRoot(path, root string) bool {
+	p := filepath.Clean(strings.TrimSpace(path))
+	r := filepath.Clean(strings.TrimSpace(root))
+	if p == "" || r == "" {
+		return false
+	}
+	if p == r {
+		return true
+	}
+	return strings.HasPrefix(p, r+string(filepath.Separator))
+}

internal/artifacts/resolve_test.go

@@ -0,0 +1,58 @@
+package artifacts
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+)
+
+func TestResolveSessionLocalPathForRead(t *testing.T) {
+	workspace := t.TempDir()
+	paths := buildSessionPaths(workspace, "s-1")
+	if err := os.MkdirAll(paths.TranscriptsRawDir, 0o755); err != nil {
+		t.Fatalf("MkdirAll() error = %v", err)
+	}
+	target := filepath.Join(paths.TranscriptsRawDir, "alice.json")
+	if err := os.WriteFile(target, []byte(`{"segments":[]}`), 0o644); err != nil {
+		t.Fatalf("WriteFile() error = %v", err)
+	}
+
+	abs := target
+	got := ResolveSessionLocalPathForRead(paths, abs)
+	if got != abs {
+		t.Fatalf("absolute path resolution = %q, want %q", got, abs)
+	}
+
+	sessionRelative := filepath.Join("transcripts", "raw", "alice.json")
+	got = ResolveSessionLocalPathForRead(paths, sessionRelative)
+	if got != target {
+		t.Fatalf("session-relative resolution = %q, want %q", got, target)
+	}
+}
+
+func TestResolveSessionLocalPathForReadRelativeWorkspaceRootQualifiedPath(t *testing.T) {
+	cwd, err := os.Getwd()
+	if err != nil {
+		t.Fatalf("Getwd() error = %v", err)
+	}
+	workspaceAbs := t.TempDir()
+	workspaceRel, err := filepath.Rel(cwd, workspaceAbs)
+	if err != nil {
+		t.Fatalf("Rel() error = %v", err)
+	}
+
+	paths := buildSessionPaths(workspaceRel, "s-1")
+	target := filepath.Join(paths.TranscriptsRawDir, "alice.json")
+	if err := os.MkdirAll(filepath.Dir(target), 0o755); err != nil {
+		t.Fatalf("MkdirAll() error = %v", err)
+	}
+	if err := os.WriteFile(target, []byte(`{"segments":[]}`), 0o644); err != nil {
+		t.Fatalf("WriteFile() error = %v", err)
+	}
+
+	manifestPath := filepath.Join(workspaceRel, "work", "s-1", "transcripts", "raw", "alice.json")
+	got := ResolveSessionLocalPathForRead(paths, manifestPath)
+	if got != filepath.Clean(manifestPath) {
+		t.Fatalf("resolution = %q, want %q", got, filepath.Clean(manifestPath))
+	}
+}

internal/stage/merge.go

@@ -190,9 +190,7 @@ func discoverRawTranscripts(m *manifest.Manifest, paths artifacts.SessionPaths)
 				if p == "" {
 					continue
 				}
-				if !filepath.IsAbs(p) {
-					p = filepath.Join(paths.Root, p)
-				}
+				p = artifacts.ResolveSessionLocalPathForRead(paths, p)
 				fromManifest = append(fromManifest, filepath.Clean(p))
 			}
 		}

internal/stage/merge_test.go

@@ -2,6 +2,7 @@ package stage
 
 import (
 	"context"
+	"os"
 	"path/filepath"
 	"strings"
 	"testing"
@@ -153,9 +154,90 @@ func TestMergeStageFallsBackToRawDirectoryWhenTranscribeOutputsMissing(t *testin
 	}
 }
 
+func TestMergeStageResolvesWorkspaceQualifiedManifestOutputsWithoutDuplication(t *testing.T) {
+	env, m := setupMergeEnvWithRelativeWorkspaceRoot(t)
+	paths := env.ArtifactStore.SessionPaths(m.SessionID)
+
+	rawPath := filepath.Join(paths.TranscriptsRawDir, "alice.json")
+	writeFile(t, rawPath, `{"segments":[]}`)
+	writeFile(t, filepath.Join(paths.InputsDir, "speakers.yml"), "match: []\n")
+	writeFile(t, filepath.Join(paths.InputsDir, "autocorrect.yml"), "rules: []\n")
+
+	m.MarkStageSucceeded("transcribe", time.Now().UTC(), []manifest.ArtifactRecord{
+		{Kind: "transcript_raw", LocalPath: filepath.Join(env.Config.Pipeline.Workspace.Root, "work", m.SessionID, "transcripts", "raw", "alice.json")},
+	})
+
+	fake := &seriatim.FakeRunner{}
+	env.Seriatim = fake
+
+	if _, err := (mergeStage{}).Run(context.Background(), env, m); err != nil {
+		t.Fatalf("merge.Run() error = %v", err)
+	}
+	if len(fake.Requests) != 1 {
+		t.Fatalf("fake requests = %d, want 1", len(fake.Requests))
+	}
+	got := fake.Requests[0].InputTranscriptPaths
+	if len(got) != 1 {
+		t.Fatalf("input transcript paths = %#v, want len 1", got)
+	}
+	if got[0] != filepath.Clean(rawPath) {
+		t.Fatalf("resolved transcript path = %q, want %q", got[0], filepath.Clean(rawPath))
+	}
+}
+
+func TestMergeStageResolvesSessionRelativeManifestOutputs(t *testing.T) {
+	env, m := setupMergeEnv(t)
+	paths := env.ArtifactStore.SessionPaths(m.SessionID)
+
+	rawPath := filepath.Join(paths.TranscriptsRawDir, "alice.json")
+	writeFile(t, rawPath, `{"segments":[]}`)
+	writeFile(t, filepath.Join(paths.InputsDir, "speakers.yml"), "match: []\n")
+	writeFile(t, filepath.Join(paths.InputsDir, "autocorrect.yml"), "rules: []\n")
+
+	m.MarkStageSucceeded("transcribe", time.Now().UTC(), []manifest.ArtifactRecord{
+		{Kind: "transcript_raw", LocalPath: filepath.Join("transcripts", "raw", "alice.json")},
+	})
+
+	fake := &seriatim.FakeRunner{}
+	env.Seriatim = fake
+
+	if _, err := (mergeStage{}).Run(context.Background(), env, m); err != nil {
+		t.Fatalf("merge.Run() error = %v", err)
+	}
+	if len(fake.Requests) != 1 {
+		t.Fatalf("fake requests = %d, want 1", len(fake.Requests))
+	}
+	got := fake.Requests[0].InputTranscriptPaths
+	if len(got) != 1 {
+		t.Fatalf("input transcript paths = %#v, want len 1", got)
+	}
+	if got[0] != filepath.Clean(rawPath) {
+		t.Fatalf("resolved transcript path = %q, want %q", got[0], filepath.Clean(rawPath))
+	}
+}
+
 func setupMergeEnv(t *testing.T) (*Env, *manifest.Manifest) {
 	t.Helper()
 	workspace := t.TempDir()
+	return setupMergeEnvWithWorkspace(t, workspace)
+}
+
+func setupMergeEnvWithRelativeWorkspaceRoot(t *testing.T) (*Env, *manifest.Manifest) {
+	t.Helper()
+	cwd, err := os.Getwd()
+	if err != nil {
+		t.Fatalf("Getwd() error = %v", err)
+	}
+	workspaceAbs := t.TempDir()
+	workspaceRel, err := filepath.Rel(cwd, workspaceAbs)
+	if err != nil {
+		t.Fatalf("Rel() error = %v", err)
+	}
+	return setupMergeEnvWithWorkspace(t, workspaceRel)
+}
+
+func setupMergeEnvWithWorkspace(t *testing.T, workspace string) (*Env, *manifest.Manifest) {
+	t.Helper()
 	cfgDir := t.TempDir()
 	sessionPath := filepath.Join(cfgDir, "session.yml")
 	pipelinePath := filepath.Join(cfgDir, "pipeline.yml")