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 sourceAndEntity() []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"},
		{Title: "Jev", Type: "entity", Subtype: "tool", Domain: "ai",
			Content: "## Definition\n\nA decision model.\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)
}

func TestRun_WithAWingWritesIntoThatWingsHalls(t *testing.T) {
	dir := t.TempDir()
	_, err := Run(context.Background(), wingCfg("radar", sourceAndEntity()...), dir, "body of the talk", "dark-factory-talk", false)
	require.NoError(t, err)

	src := read(t, dir, "wiki/radar/sources/dark-factory-talk.md")
	assert.Contains(t, src, "wing: radar")
	assert.Contains(t, src, "hall: sources")

	ent := read(t, dir, "wiki/radar/facts/jev.md")
	assert.Contains(t, ent, "wing: radar")
	assert.Contains(t, ent, "hall: facts")
	assert.Contains(t, ent, "kind: entity")
}

func TestRun_WithAWingDoesNotTouchTheRetiredLayout(t *testing.T) {
	dir := t.TempDir()
	_, err := Run(context.Background(), wingCfg("radar", sourceAndEntity()...), dir, "body of the talk", "dark-factory-talk", false)
	require.NoError(t, err)
	for _, retired := range []string{"wiki/sources", "wiki/concepts", "wiki/entities"} {
		assert.False(t, exists(dir, retired), "%s is the retired layout and must not be written", retired)
	}
}

func TestRun_WithAWingTheSourceLinksBackFromTheEntity(t *testing.T) {
	dir := t.TempDir()
	_, err := Run(context.Background(), wingCfg("radar", sourceAndEntity()...), dir, "body of the talk", "dark-factory-talk", false)
	require.NoError(t, err)
	assert.Contains(t, read(t, dir, "wiki/radar/facts/jev.md"), "[[dark-factory-talk|Dark Factory Talk]]")
}

func TestRun_WithAWingASecondSourceReusesTheSameEntityPage(t *testing.T) {
	dir := t.TempDir()
	_, err := Run(context.Background(), wingCfg("radar", sourceAndEntity()...), dir, "first body", "talk-one", false)
	require.NoError(t, err)

	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("radar", again...), dir, "second body", "talk-two", false)
	require.NoError(t, err)

	files, err := filepath.Glob(filepath.Join(dir, "wiki/radar/facts/*.md"))
	require.NoError(t, err)
	assert.Len(t, files, 1, "an entity matched by title (leading article ignored) is merged, not duplicated")
	assert.Contains(t, read(t, dir, "wiki/radar/facts/jev.md"), "talk-two")
}

func TestRun_WithAWingNeverMergesIntoAnotherWingsPage(t *testing.T) {
	dir := t.TempDir()
	other := "---\ntitle: Jev\nwing: client-x\nhall: facts\nkind: entity\n---\n\n## Definition\n\nPrivate to another wing.\n"
	require.NoError(t, os.MkdirAll(filepath.Join(dir, "wiki/client-x/facts"), 0o755))
	require.NoError(t, os.WriteFile(filepath.Join(dir, "wiki/client-x/facts/jev.md"), []byte(other), 0o644))

	_, err := Run(context.Background(), wingCfg("radar", sourceAndEntity()...), dir, "body", "dark-factory-talk", false)
	require.NoError(t, err)

	assert.Equal(t, other, read(t, dir, "wiki/client-x/facts/jev.md"), "another wing's page is never read into or rewritten")
	assert.True(t, exists(dir, "wiki/radar/facts/jev.md"))
}

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, sourceAndEntity()...), 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
}

// The dangerous path is the back-reference: a source that links [[Jev]] without
// proposing a Jev page makes the pipeline go and edit an existing Jev page. That
// lookup must be confined to the target wing.
func TestRun_WithAWingASourceLinkingToAnotherWingsEntityDoesNotEditIt(t *testing.T) {
	dir := t.TempDir()
	other := "---\ntitle: Jev\nwing: client-x\nhall: facts\nkind: entity\n---\n\n## Definition\n\nPrivate to another wing.\n\n## Sources\n"
	require.NoError(t, os.MkdirAll(filepath.Join(dir, "wiki/client-x/facts"), 0o755))
	require.NoError(t, os.WriteFile(filepath.Join(dir, "wiki/client-x/facts/jev.md"), []byte(other), 0o644))

	onlySource := sourceAndEntity()[:1]
	_, err := Run(context.Background(), wingCfg("radar", onlySource...), dir, "body", "dark-factory-talk", false)
	require.NoError(t, err)

	assert.Equal(t, other, read(t, dir, "wiki/client-x/facts/jev.md"), "no back-reference may be written into another wing")
}