hyperguild/ingestion/internal/pipeline/wing_layout_test.go

package pipeline

import (
	"context"
	"os"
	"path/filepath"
	"testing"

	"github.com/stretchr/testify/assert"
	"github.com/stretchr/testify/require"
)

func wingCfg(wing string, pages ...RawPage) Config {
	return Config{
		Wing:   wing,
		Schema: "# Schema",
		Complete: func(context.Context, string, string) (string, error) {
			return mustJSON(pages), nil
		},
	}
}

func sourceEntityConcept() []RawPage {
	return []RawPage{
		{Title: "Dark Factory Talk", Type: "source", Subtype: "video", Domain: "ai",
			Content: "## Summary\n\nA talk.\n\n## Entities Mentioned\n\n[[Jev]]\n\n## Concepts Introduced or Reinforced\n\n[[Dark Factory]]\n"},
		{Title: "Jev", Type: "entity", Subtype: "tool", Domain: "ai",
			Content: "## Definition\n\nA decision model.\n\n## Sources\n"},
		{Title: "Dark Factory", Type: "concept", Domain: "ai",
			Content: "## Definition\n\nA fully automated build line.\n\n## Sources\n"},
	}
}

func exists(dir, rel string) bool {
	_, err := os.Stat(filepath.Join(dir, rel))
	return err == nil
}

func read(t *testing.T, dir, rel string) string {
	t.Helper()
	b, err := os.ReadFile(filepath.Join(dir, rel))
	require.NoError(t, err)
	return string(b)
}

// With a wing, the SOURCE belongs to the wing; concepts and entities are the
// shared cross-wing layer (wiki/concepts, wiki/entities) because their slug is
// the graph's global identity and they are cited from many wings.
func TestRun_WithAWingTheSourceIsPerWingAndConceptsEntitiesAreShared(t *testing.T) {
	dir := t.TempDir()
	_, err := Run(context.Background(), wingCfg("web-clip", sourceEntityConcept()...), dir, "body of the talk", "dark-factory-talk", false)
	require.NoError(t, err)

	src := read(t, dir, "wiki/web-clip/sources/dark-factory-talk.md")
	assert.Contains(t, src, "wing: web-clip")
	assert.Contains(t, src, "hall: sources")
	assert.True(t, exists(dir, "wiki/entities/jev.md"))
	assert.True(t, exists(dir, "wiki/concepts/dark-factory.md"))
	assert.NotContains(t, read(t, dir, "wiki/entities/jev.md"), "wing:", "a shared page belongs to no single wing")
	assert.False(t, exists(dir, "wiki/web-clip/facts"), "no per-wing copy of a shared page")
	assert.False(t, exists(dir, "wiki/sources"), "the retired source layout is not used")
}

func TestRun_WithAWingTheSharedPageLinksBackToTheSource(t *testing.T) {
	dir := t.TempDir()
	_, err := Run(context.Background(), wingCfg("web-clip", sourceEntityConcept()...), dir, "body", "dark-factory-talk", false)
	require.NoError(t, err)
	assert.Contains(t, read(t, dir, "wiki/entities/jev.md"), "[[dark-factory-talk|Dark Factory Talk]]")
}

func TestRun_TheSameEntityCitedFromTwoWingsIsOnePageWithBothSources(t *testing.T) {
	dir := t.TempDir()
	_, err := Run(context.Background(), wingCfg("web-clip", sourceEntityConcept()...), dir, "first body", "talk-one", false)
	require.NoError(t, err)
	// A wing must already exist to be named; create it the way a note would.
	require.NoError(t, os.MkdirAll(filepath.Join(dir, "wiki/homelab/facts"), 0o755))

	again := []RawPage{
		{Title: "Second Talk", Type: "source", Subtype: "video", Content: "## Summary\n\nMore.\n\n## Entities Mentioned\n\n[[The Jev]]\n"},
		{Title: "The Jev", Type: "entity", Subtype: "tool", Content: "## Definition\n\nStill a decision model.\n\n## Sources\n"},
	}
	_, err = Run(context.Background(), wingCfg("homelab", again...), dir, "second body", "talk-two", false)
	require.NoError(t, err)

	files, err := filepath.Glob(filepath.Join(dir, "wiki/entities/*.md"))
	require.NoError(t, err)
	assert.Len(t, files, 1, "one entity page, matched by title with the leading article ignored")
	jev := read(t, dir, "wiki/entities/jev.md")
	assert.Contains(t, jev, "talk-one")
	assert.Contains(t, jev, "talk-two")
	assert.False(t, exists(dir, "wiki/homelab/facts/jev.md"), "no second copy in the second wing")
}

// A source slug is the graph's identity too: a second wing must not take over
// the slug of a source that lives elsewhere.
func TestRun_ASourceSlugAlreadyUsedInAnotherWingIsRefused(t *testing.T) {
	dir := t.TempDir()
	_, err := Run(context.Background(), wingCfg("web-clip", sourceEntityConcept()...), dir, "body", "talk", false)
	require.NoError(t, err)
	require.NoError(t, os.MkdirAll(filepath.Join(dir, "wiki/homelab/facts"), 0o755))

	_, err = Run(context.Background(), wingCfg("homelab", sourceEntityConcept()...), dir, "different body", "talk", false)
	require.Error(t, err)
	assert.Contains(t, err.Error(), "already used")
	assert.False(t, exists(dir, "wiki/homelab/sources/talk.md"))
}

func TestRun_ASourceSlugAlreadyInTheRetiredLayoutIsRefused(t *testing.T) {
	dir := t.TempDir()
	require.NoError(t, os.MkdirAll(filepath.Join(dir, "wiki/sources"), 0o755))
	require.NoError(t, os.WriteFile(filepath.Join(dir, "wiki/sources/talk.md"), []byte("---\ntitle: Talk\n---\nold"), 0o644))

	_, err := Run(context.Background(), wingCfg("web-clip", sourceEntityConcept()...), dir, "body", "talk", false)
	require.Error(t, err)
	assert.Contains(t, err.Error(), "already used")
}

func TestRun_ReIngestingTheSameSourceIntoItsOwnWingIsNotACollision(t *testing.T) {
	dir := t.TempDir()
	_, err := Run(context.Background(), wingCfg("web-clip", sourceEntityConcept()...), dir, "body one", "talk", false)
	require.NoError(t, err)
	_, err = Run(context.Background(), wingCfg("web-clip", sourceEntityConcept()...), dir, "body two, edited", "talk", false)
	require.NoError(t, err)
}

func TestRun_WithAWingRejectsANameThatIsNotAPlainWing(t *testing.T) {
	for _, bad := range []string{"../escape", "a/b", "Radar", "has space", "."} {
		t.Run(bad, func(t *testing.T) {
			dir := t.TempDir()
			_, err := Run(context.Background(), wingCfg(bad, sourceEntityConcept()...), dir, "body", "talk", false)
			require.Error(t, err)
			assert.Len(t, listAll(t, dir), 1, "nothing is written for a rejected wing (only the temp dir itself exists)")
		})
	}
}

func listAll(t *testing.T, dir string) []string {
	t.Helper()
	var out []string
	_ = filepath.WalkDir(dir, func(p string, _ os.DirEntry, _ error) error { out = append(out, p); return nil })
	return out
}