textmachine/backend/internal/pipeline/statusartifacts_test.go

package pipeline

import (
	"context"
	"encoding/json"
	"os"
	"path/filepath"
	"strings"
	"testing"

	"textmachine/backend/internal/obs"
)

// statusartifacts_test.go: backlog row 213 and 17-seam-inbound-law п.3 — the paths the engine REPORTS
// instead of letting another zone derive, and the versioned envelope every JSON output carries.
//
// The defect row 213 records is not hypothetical: the platform parses book.yaml for two keys and, when
// `project_db` is absent (which is what the operator template ships), computes `<workdir>/<book_id>.db`
// itself — a byte-for-byte copy of config.LoadBook's convention — then appends its own copy of the bank
// read-out's `.bank.json` suffix. Changing either default here would have quietly pointed another zone's
// read at a file that does not exist.

func TestStatusReportsTheEngineOwnedPaths(t *testing.T) {
	rec := &reqRec{}
	srv := newJSONProvider(rec, draftEdit)
	defer srv.Close()
	bookPath := setupProject(t, srv.URL)
	dir := filepath.Dir(bookPath)

	r := newRunner(t, bookPath)
	defer r.Close()
	rep, err := r.Status(context.Background())
	if err != nil {
		t.Fatal(err)
	}
	if rep.Version != statusVersion {
		t.Errorf("status_version = %q, want %q — the bank read-out and the manifest both carry one and this was the gap", rep.Version, statusVersion)
	}
	want := map[string]string{
		"project_db":    filepath.Join(dir, "test-book.db"),
		"bank_export":   filepath.Join(dir, "test-book.db.bank.json"),
		"mined_delta":   filepath.Join(dir, "test-book.mined-delta.yaml"),
		"mined_rejects": filepath.Join(dir, "test-book.mined-rejects.yaml"),
	}
	got := map[string]string{
		"project_db":    rep.Artifacts.ProjectDB,
		"bank_export":   rep.Artifacts.BankExport,
		"mined_delta":   rep.Artifacts.MinedDelta,
		"mined_rejects": rep.Artifacts.MinedRejects,
	}
	for k, w := range want {
		if got[k] != w {
			t.Errorf("artifacts.%s = %q, want %q", k, got[k], w)
		}
		if !filepath.IsAbs(got[k]) {
			t.Errorf("artifacts.%s must be absolute — the consumer reads it from another working directory: %q", k, got[k])
		}
	}
	// The suffix stays what it is. Renaming the bank read-out to a fixed name beside events.jsonl is a
	// BREAKING change and its window is the chapter-structure re-cut (backlog row 161), not this one.
	if rep.Artifacts.BankExport != r.bankExportPath() {
		t.Errorf("the reported path must be the one the engine actually writes: %q vs %q", rep.Artifacts.BankExport, r.bankExportPath())
	}
}

// TestStatusJSONCarriesTheEnvelopeAdditively: the platform allowlists the fields it materializes and
// ignores the rest, so a new field is safe — but only if it is genuinely ADDED and nothing existing moved.
func TestStatusJSONCarriesTheEnvelopeAdditively(t *testing.T) {
	rec := &reqRec{}
	srv := newJSONProvider(rec, draftEdit)
	defer srv.Close()
	bookPath := setupProject(t, srv.URL)
	r := newRunner(t, bookPath)
	defer r.Close()
	rep, err := r.Status(context.Background())
	if err != nil {
		t.Fatal(err)
	}
	body, err := json.Marshal(rep)
	if err != nil {
		t.Fatal(err)
	}
	var doc map[string]json.RawMessage
	if err := json.Unmarshal(body, &doc); err != nil {
		t.Fatal(err)
	}
	for _, k := range []string{"status_version", "artifacts", "book_id", "total_units", "committed_usd", "reserved_usd", "progress", "chapters"} {
		if _, ok := doc[k]; !ok {
			t.Errorf("status --json lost the field %q", k)
		}
	}
}

// TestExportJSONCarriesItsVersion: the law's п.3 is general ("every JSON output of the engine carries a
// versioned envelope") and names `status --json` because that is where the gap was noticed. The export
// was the second surface without one.
func TestExportJSONCarriesItsVersion(t *testing.T) {
	rec := &reqRec{}
	srv := newJSONProvider(rec, draftEdit)
	defer srv.Close()
	bookPath := setupProject(t, srv.URL)
	r := newRunner(t, bookPath)
	defer r.Close()
	exp, err := r.Export(false)
	if err != nil {
		t.Fatal(err)
	}
	if exp.Version != exportVersion {
		t.Errorf("export_version = %q, want %q", exp.Version, exportVersion)
	}
}

// --- the dofix pack ---------------------------------------------------------------------------------

func TestTheReportedPathsAreAbsoluteEvenWhenTheConfigIsNOT(t *testing.T) {
	// The existing pin checks filepath.IsAbs, but its fixture hands the engine an ABSOLUTE --config, so
	// every path is already absolute before absPath sees it and removing absPath changed nothing. Measured:
	// with a relative --config the call is load-bearing, and without it a consumer running in its own
	// working directory — which is what the platform is — resolves the path against a different directory
	// and reads a different file, or none.
	rec := &reqRec{}
	srv := newJSONProvider(rec, draftEdit)
	defer srv.Close()
	bookPath := setupProject(t, srv.URL)
	dir := filepath.Dir(bookPath)

	t.Chdir(dir)
	r := newRunner(t, filepath.Base(bookPath)) // RELATIVE, the way an operator types it
	defer r.Close()
	rep, err := r.Status(context.Background())
	if err != nil {
		t.Fatal(err)
	}
	for name, got := range map[string]string{
		"project_db": rep.Artifacts.ProjectDB, "bank_export": rep.Artifacts.BankExport,
		"mined_delta": rep.Artifacts.MinedDelta, "mined_rejects": rep.Artifacts.MinedRejects,
	} {
		if !filepath.IsAbs(got) {
			t.Errorf("artifacts.%s = %q — a consumer in another working directory reads the wrong file", name, got)
		}
	}
}

func TestTheManifestCarriesTheArtifactEnvelope(t *testing.T) {
	// Where the envelope has to LAND. `status --json` publishes it, and the platform's read-model does not
	// call status at all — its engine interface is {Manifest, Export}, and the one place it needs an
	// engine-owned path it re-derives `<workdir>/<book_id>.db` itself, which is the duplicate unified
	// backlog row 213 records. An envelope on a surface nobody reads removes nothing.
	rec := &reqRec{}
	srv := newJSONProvider(rec, draftEdit)
	defer srv.Close()
	bookPath := setupProject(t, srv.URL)
	dir := filepath.Dir(bookPath)

	r := newRunner(t, bookPath)
	defer r.Close()
	m, err := r.BuildAndPersistManifest()
	if err != nil {
		t.Fatal(err)
	}
	want := map[string]string{
		"project_db":    filepath.Join(dir, "test-book.db"),
		"bank_export":   filepath.Join(dir, "test-book.db.bank.json"),
		"mined_delta":   filepath.Join(dir, "test-book.mined-delta.yaml"),
		"mined_rejects": filepath.Join(dir, "test-book.mined-rejects.yaml"),
	}
	got := map[string]string{
		"project_db": m.Artifacts.ProjectDB, "bank_export": m.Artifacts.BankExport,
		"mined_delta": m.Artifacts.MinedDelta, "mined_rejects": m.Artifacts.MinedRejects,
	}
	for k, w := range want {
		if got[k] != w {
			t.Errorf("manifest artifacts.%s = %q, want %q", k, got[k], w)
		}
	}
	// It is in the DOCUMENT a consumer parses, not only in the Go struct.
	body, err := json.Marshal(m)
	if err != nil {
		t.Fatal(err)
	}
	var doc map[string]any
	if err := json.Unmarshal(body, &doc); err != nil {
		t.Fatal(err)
	}
	art, ok := doc["artifacts"].(map[string]any)
	if !ok || art["mined_delta"] != want["mined_delta"] {
		t.Fatalf("the emitted manifest must carry the envelope: %v", doc["artifacts"])
	}
	// DERIVED on every read, never trusted from the file: a path stored in a sidecar that travels with
	// the book directory is wrong the first time the directory moves. A stored manifest with a stale
	// envelope must come back with the current one.
	raw, err := os.ReadFile(r.manifestPath())
	if err != nil {
		t.Fatal(err)
	}
	stale := strings.Replace(string(raw), want["mined_delta"], "/gone/elsewhere.yaml", 1)
	if stale == string(raw) {
		t.Fatal("test premise broken: the stored manifest does not carry the path")
	}
	writeFile(t, r.manifestPath(), stale)
	if loaded := r.loadManifest(); loaded == nil {
		t.Fatal("the doctored manifest must still be structurally valid (only the path was changed)")
	} else if loaded.Artifacts.MinedDelta != want["mined_delta"] {
		t.Fatalf("a stored path must be OVERWRITTEN with the current truth, got %q", loaded.Artifacts.MinedDelta)
	}
}

// TestTheBankExportCarriesItsFreshnessAnchor pins fix3 §4.6: the export's write discipline is
// loud-but-not-fatal, so the file on disk may be a previous boundary's projection — and «may be stale»
// used to live only in a log line. The document itself now names the boundary that produced it and the
// run identity a journal reader can compare against; a later boundary moves the anchor.
func TestTheBankExportCarriesItsFreshnessAnchor(t *testing.T) {
	rec := &reqRec{}
	srv := newJSONProvider(rec, draftEdit)
	defer srv.Close()
	r := newRunner(t, setupProject(t, srv.URL))
	defer r.Close()
	ctx := obs.WithReqInfo(context.Background(), obs.ReqInfo{TraceID: "fx3-freshness-trace"})
	if _, err := r.TranslateBook(ctx); err != nil {
		t.Fatal(err)
	}
	var exp BankExport
	mustParseJSON(t, r.bankExportPath(), &exp)
	if exp.AsOf != "run-finished" {
		t.Fatalf("the last boundary of a completed run is run-finished, got %q", exp.AsOf)
	}
	if exp.RunID != "fx3-freshness-trace" {
		t.Fatalf("the anchor must carry the run identity the journal streams, got %q", exp.RunID)
	}
	// A later boundary REPLACES the anchor — that is the whole point: two projections are tellable apart.
	ctx2 := obs.WithReqInfo(context.Background(), obs.ReqInfo{TraceID: "fx3-second-run"})
	r.exportBank(ctx2, "redrive/re-seeded")
	mustParseJSON(t, r.bankExportPath(), &exp)
	if exp.AsOf != "redrive/re-seeded" || exp.RunID != "fx3-freshness-trace" {
		// The emitter is authoritative once the journal is open: the SAME runner re-exporting still
		// belongs to its own stream, whatever a caller stamped on the context.
		t.Fatalf("anchor after a later boundary: as_of=%q run_id=%q", exp.AsOf, exp.RunID)
	}
	// …and a runner with no journal and no traced context says «unknown», never «current».
	r2 := newRunner(t, setupProject(t, srv.URL))
	defer r2.Close()
	r2.exportBank(context.Background(), "run-start/seeded")
	mustParseJSON(t, r2.bankExportPath(), &exp)
	if exp.RunID != "" || exp.AsOf != "run-start/seeded" {
		t.Fatalf("no run identity must export as empty, got as_of=%q run_id=%q", exp.AsOf, exp.RunID)
	}
}