textmachine/backend/internal/pipeline/systemmessages_wave_test.go

package pipeline

import (
	"context"
	"strings"
	"testing"
)

// systemmessages_wave_test.go proves the SystemMessages fix on the two paths that actually carry
// the memory bank to a model — the DRAFT wave (the translator's src→dst glossary) and the EDIT
// wave (the editor's CONFIRMED-dst constraint block). The edit wave matters on its own account:
// it is the wave that delivers the OWNER'S SIGNED decisions, so a hop that ate its system message
// would silently eat the owner's edits, not merely a machine's proposals.
//
// ⚠ The assertion is the POST-FIX invariant, deliberately: "the injection is in the request" is
// true of unfixed code too (the assembler always built it; the endpoint dropped it), so it would
// be green on a broken engine. What separates fixed from unfixed is ONE system message with the
// injection INSIDE it.

// systemMessagesOf returns the wire body's system messages, in order.
func systemMessagesOf(t *testing.T, body string) []wireMsg {
	t.Helper()
	var out []wireMsg
	for _, m := range bodyMessages(t, body) {
		if m.Role == "system" {
			out = append(out, m)
		}
	}
	return out
}

// TestSingleSystemEndpointCarriesTheBankOnBothWaves runs a real two-wave book against a provider
// declared single-system and asserts, for the draft AND the edit request, that the endpoint got
// exactly one system message and that the bank block is inside it.
func TestSingleSystemEndpointCarriesTheBankOnBothWaves(t *testing.T) {
	rec := &reqRec{}
	srv := newJSONProvider(rec, func(body string) (string, string) {
		if isEditBody(body) {
			return "ОТРЕДАКТИРОВАННЫЙ ПЕРЕВОД про Судзуки.", "stop"
		}
		return "Судзуки пошёл в тихую библиотеку.", "stop"
	})
	defer srv.Close()
	bookPath := setupProjectOpts(t, srv.URL, projectOpts{
		regenerate: 1, source: suzukiSource, glossarySeed: suzukiSeed, systemMessagesSingle: true,
	})

	r := newRunner(t, bookPath)
	res, err := r.TranslateBook(context.Background())
	if err != nil {
		t.Fatal(err)
	}
	if res.Flagged != 0 {
		t.Fatalf("the join must not disturb the verdict: flagged=%d", res.Flagged)
	}

	cases := []struct {
		wave      string
		isEdit    bool
		wantBlock string // the bank text this wave's role receives
	}{
		// waverun.go runDraftChunk: the translator's src→dst glossary.
		{wave: "draft", isEdit: false, wantBlock: "ГЛОССАРИЙ"},
		// waverun.go runEditUnit: the editor's CONFIRMED-dst constraint block — the carrier of the
		// owner's signed decisions.
		{wave: "edit", isEdit: true, wantBlock: "КАНОНИЧЕСКИЕ ПЕРЕВОДЫ"},
	}
	for _, c := range cases {
		t.Run(c.wave, func(t *testing.T) {
			seen := 0
			for _, body := range rec.all() {
				if isEditBody(body) != c.isEdit {
					continue
				}
				seen++
				sys := systemMessagesOf(t, body)
				if len(sys) != 1 {
					t.Fatalf("%s wave: a single-system endpoint received %d system messages — the bank is being dropped by the provider",
						c.wave, len(sys))
				}
				if !strings.Contains(sys[0].Content, c.wantBlock) {
					t.Fatalf("%s wave: the bank block %q is not inside the single system message: %q",
						c.wave, c.wantBlock, sys[0].Content)
				}
				if !strings.Contains(sys[0].Content, "Судзуки") {
					t.Fatalf("%s wave: the approved rendering did not reach the model: %q", c.wave, sys[0].Content)
				}
				// The stable prompt prefix must stay FIRST inside the join, or the provider's prefix
				// cache stops hitting on every call.
				iPrefix := strings.Index(sys[0].Content, "Редактируй перевод.")
				if !c.isEdit {
					iPrefix = strings.Index(sys[0].Content, "Переводи с")
				}
				if iPrefix != 0 {
					t.Fatalf("%s wave: the stable prefix is not first in the join (index %d): %q", c.wave, iPrefix, sys[0].Content)
				}
				// The user turn survives the join untouched, and the bank never leaks into it.
				msgs := bodyMessages(t, body)
				if len(msgs) != 2 || msgs[1].Role != "user" {
					t.Fatalf("%s wave: wire shape = %+v, want [system user]", c.wave, msgs)
				}
				if strings.Contains(msgs[1].Content, c.wantBlock) {
					t.Fatalf("%s wave: the bank leaked into the user message: %q", c.wave, msgs[1].Content)
				}
			}
			if seen == 0 {
				t.Fatalf("%s wave recorded no request", c.wave)
			}
		})
	}
}

// TestMultiSystemEndpointIsUnchangedOnBothWaves is the control on the SAME book: a provider that
// declares nothing still receives the two-message wire, so the fix cannot have changed what every
// other provider in the stack sees.
func TestMultiSystemEndpointIsUnchangedOnBothWaves(t *testing.T) {
	rec := &reqRec{}
	srv := newJSONProvider(rec, func(body string) (string, string) {
		if isEditBody(body) {
			return "ОТРЕДАКТИРОВАННЫЙ ПЕРЕВОД про Судзуки.", "stop"
		}
		return "Судзуки пошёл в тихую библиотеку.", "stop"
	})
	defer srv.Close()
	bookPath := setupProjectOpts(t, srv.URL, projectOpts{
		regenerate: 1, source: suzukiSource, glossarySeed: suzukiSeed,
	})

	r := newRunner(t, bookPath)
	if _, err := r.TranslateBook(context.Background()); err != nil {
		t.Fatal(err)
	}
	for _, body := range rec.all() {
		if sys := systemMessagesOf(t, body); len(sys) != 2 {
			t.Fatalf("an undeclared endpoint must keep the two-message wire, got %d: %+v", len(sys), sys)
		}
	}
}