textmachine/platform/internal/httpapi/intake_test.go

package httpapi

import (
	"bytes"
	"context"
	"errors"
	"fmt"
	"io"
	"mime/multipart"
	"net/http"
	"net/http/httptest"
	"os"
	"strings"
	"testing"
	"time"

	"textmachine/platform/internal/books"
	"textmachine/platform/internal/pgstore"
)

// The upload is the only streaming route on this surface, and the only one whose body limit is not
// the default. Both facts are asserted here on the REAL handler chain — through New, with the guard,
// the CSRF layer and the per-route cap in place — because a limit tested on a bare handler is a
// limit tested somewhere other than where it applies (the shape of PD-46).

type fakeIntake struct {
	book pgstore.Book
	err  error
	got  books.Intake
	// read is what the service actually pulled out of the part, which is how a test tells "the file
	// was streamed" from "the handler read it into memory and handed over a copy".
	read int64
}

func (f *fakeIntake) Accept(_ context.Context, in books.Intake) (pgstore.Book, error) {
	f.got = in
	if in.File != nil {
		n, err := io.Copy(io.Discard, in.File)
		f.read = n
		if err != nil {
			return pgstore.Book{}, err
		}
	}
	if f.err != nil {
		return pgstore.Book{}, f.err
	}
	return f.book, nil
}

// form builds a multipart body. The order of the parts is the caller's on purpose: the file MUST
// come last, and the test that says so builds it the other way round.
func form(t *testing.T, file string, size int, fieldsFirst bool) (string, io.Reader) {
	t.Helper()
	var buf bytes.Buffer
	w := multipart.NewWriter(&buf)
	writeFields := func() {
		for k, v := range map[string]string{"source_lang": "zh", "target_lang": "ru", "genre": "xianxia"} {
			if err := w.WriteField(k, v); err != nil {
				t.Fatal(err)
			}
		}
	}
	if fieldsFirst {
		writeFields()
	}
	part, err := w.CreateFormFile("file", file)
	if err != nil {
		t.Fatal(err)
	}
	if _, err := part.Write(bytes.Repeat([]byte("a"), size)); err != nil {
		t.Fatal(err)
	}
	if !fieldsFirst {
		writeFields()
	}
	if err := w.Close(); err != nil {
		t.Fatal(err)
	}
	return w.FormDataContentType(), &buf
}

func upload(t *testing.T, h http.Handler, contentType string, body io.Reader) *httptest.ResponseRecorder {
	t.Helper()
	r := httptest.NewRequest("POST", "/v0/books", body)
	r.Header.Set("Content-Type", contentType)
	r.Header.Set("Authorization", "Bearer token")
	w := httptest.NewRecorder()
	h.ServeHTTP(w, r)
	return w
}

func TestAnAcceptedUploadAnswersWithTheBookTheContractDescribes(t *testing.T) {
	in := &fakeIntake{book: pgstore.Book{
		ID: "bk_7c1", Title: "蛊真人", SourceLang: "zh", TargetLang: "ru", Genre: "xianxia",
		Status: "parsing", AddedAt: time.Unix(0, 0).UTC(),
	}}
	h := v0ServerWith(t, Deps{Library: &fakeLibrary{}, Intake: in})
	ct, body := form(t, "蛊真人.txt", 1024, true)
	w := upload(t, h, ct, body)
	if w.Code != http.StatusCreated {
		t.Fatalf("status %d: %s", w.Code, w.Body)
	}
	got := decode(t, w)
	// Every field the contract marks required on Book, present on the intake's own response.
	for _, k := range []string{"id", "title", "source_lang", "target_lang", "status", "chapter_count",
		"added_at", "progress", "note_count"} {
		if _, ok := got[k]; !ok {
			t.Errorf("the created book is missing the required field %q", k)
		}
	}
	if got["status"] != "parsing" {
		t.Errorf("status %v, want the state the book is actually in", got["status"])
	}
	if in.got.SourceLang != "zh" || in.got.TargetLang != "ru" || in.got.Genre != "xianxia" {
		t.Errorf("the form did not reach the service: %+v", in.got)
	}
	if in.got.Filename != "蛊真人.txt" {
		t.Errorf("filename %q did not reach the service", in.got.Filename)
	}
	if in.read != 1024 {
		t.Errorf("the service was handed %d bytes of the file, want 1024", in.read)
	}
}

// PD-72, and the reason the register kept the row open until this route landed: the per-route cap is
// only OBSERVABLE once a route asks for more than the default, and the test has to arrive with it.
func TestAnUploadLargerThanTheRouteAllowsIsRefusedAsTooLarge(t *testing.T) {
	in := &fakeIntake{}
	h := v0ServerWith(t, Deps{
		Library: &fakeLibrary{}, Intake: in,
		Upload: UploadLimits{MaxBytes: 4096, Deadline: time.Minute},
	})
	ct, body := form(t, "book.txt", 16*1024, true)
	w := upload(t, h, ct, body)
	if w.Code != http.StatusRequestEntityTooLarge {
		t.Fatalf("status %d, want 413: %s", w.Code, w.Body)
	}
	if ct := w.Header().Get("Content-Type"); ct != "application/problem+json" {
		t.Errorf("content-type %q, want application/problem+json", ct)
	}
	problem := decode(t, w)
	if problem["status"] != float64(http.StatusRequestEntityTooLarge) || problem["title"] == "" {
		t.Errorf("problem body %v", problem)
	}
	// And a body UNDER the limit goes through the same route untouched: a cap that refuses
	// everything is not a cap.
	ct2, body2 := form(t, "book.txt", 1024, true)
	if w := upload(t, h, ct2, body2); w.Code != http.StatusCreated {
		t.Fatalf("a 1 KiB upload under a 4 KiB cap: %d %s", w.Code, w.Body)
	}
}

// The other half of the same property (PD-35): raising the limit for ONE route must not raise it for
// the rest of the surface.
func TestTheUploadLimitBelongsToTheUploadRouteAlone(t *testing.T) {
	h := v0ServerWith(t, Deps{
		Library: &fakeLibrary{}, Runs: &fakeRuns{}, Intake: &fakeIntake{},
		Upload: UploadLimits{MaxBytes: 8 << 20, Deadline: time.Minute},
	})
	big := strings.Repeat("a", DefaultMaxBody+1)
	w := call(t, h, "POST", "/v0/books/bk_1/runs", `{"verify_bank":false,"ceiling_chapters":1,"pad":"`+big+`"}`)
	if w.Code == http.StatusAccepted {
		t.Fatal("a run request larger than the default body cap was accepted: the upload's limit leaked onto every route")
	}
}

// A deployment with nowhere to put a file mounts no upload route, exactly like every other contract
// route it cannot serve — and an anonymous caller still meets 401 before 404.
func TestWithoutAnIntakeTheUploadRouteIsAGuarded404(t *testing.T) {
	h := v0ServerWith(t, Deps{Library: &fakeLibrary{}})
	ct, body := form(t, "book.txt", 16, true)
	if w := upload(t, h, ct, body); w.Code != http.StatusNotFound {
		t.Fatalf("status %d, want 404 where no intake is configured", w.Code)
	}
	r := httptest.NewRequest("POST", "/v0/books", nil)
	w := httptest.NewRecorder()
	h.ServeHTTP(w, r)
	if w.Code != http.StatusUnauthorized {
		t.Fatalf("an anonymous caller got %d, want 401 before anything else", w.Code)
	}
}

// The wire rule a streaming reader forces: the languages have to be known before the file part is
// handed over, so a form that sends them afterwards is refused rather than half-processed.
func TestAFormWhoseFileComesBeforeItsFieldsIsRefused(t *testing.T) {
	in := &fakeIntake{}
	h := v0ServerWith(t, Deps{Library: &fakeLibrary{}, Intake: in})
	ct, body := form(t, "book.txt", 64, false)
	w := upload(t, h, ct, body)
	if w.Code != http.StatusBadRequest {
		t.Fatalf("status %d, want 400: %s", w.Code, w.Body)
	}
	if in.read != 0 {
		t.Errorf("the file was streamed before the form was known to be usable (%d bytes)", in.read)
	}
}

func TestAnUploadWithNoFileAtAllIsRefused(t *testing.T) {
	h := v0ServerWith(t, Deps{Library: &fakeLibrary{}, Intake: &fakeIntake{}})
	var buf bytes.Buffer
	w := multipart.NewWriter(&buf)
	if err := w.WriteField("source_lang", "zh"); err != nil {
		t.Fatal(err)
	}
	if err := w.Close(); err != nil {
		t.Fatal(err)
	}
	if got := upload(t, h, w.FormDataContentType(), &buf); got.Code != http.StatusBadRequest {
		t.Fatalf("status %d, want 400", got.Code)
	}
	// And a body that is not multipart at all.
	r := httptest.NewRequest("POST", "/v0/books", strings.NewReader(`{"not":"multipart"}`))
	r.Header.Set("Content-Type", "application/json")
	r.Header.Set("Authorization", "Bearer token")
	rec := httptest.NewRecorder()
	h.ServeHTTP(rec, r)
	if rec.Code != http.StatusBadRequest {
		t.Fatalf("a JSON body on the intake route: %d, want 400", rec.Code)
	}
}

// A field the contract does not name is IGNORED, like an unknown JSON property: inside 0.x a minor
// bump is where optional fields appear, and refusing the upload would break a client generated
// against a later version for something it was free to ignore.
func TestAnUnknownFormFieldDoesNotRefuseTheUpload(t *testing.T) {
	in := &fakeIntake{book: pgstore.Book{ID: "bk_1", Status: "parsing"}}
	h := v0ServerWith(t, Deps{Library: &fakeLibrary{}, Intake: in})
	var buf bytes.Buffer
	w := multipart.NewWriter(&buf)
	for _, kv := range [][2]string{{"source_lang", "zh"}, {"target_lang", "ru"}, {"title", "as yet unratified"}} {
		if err := w.WriteField(kv[0], kv[1]); err != nil {
			t.Fatal(err)
		}
	}
	part, err := w.CreateFormFile("file", "b.txt")
	if err != nil {
		t.Fatal(err)
	}
	if _, err := part.Write([]byte("text")); err != nil {
		t.Fatal(err)
	}
	if err := w.Close(); err != nil {
		t.Fatal(err)
	}
	if got := upload(t, h, w.FormDataContentType(), &buf); got.Code != http.StatusCreated {
		t.Fatalf("status %d, want 201: an unknown field must not refuse the upload", got.Code)
	}
}

// A form field long enough to be a payload of its own is refused rather than truncated.
func TestAFormFieldThatIsAPayloadIsRefused(t *testing.T) {
	h := v0ServerWith(t, Deps{Library: &fakeLibrary{}, Intake: &fakeIntake{},
		Upload: UploadLimits{MaxBytes: 8 << 20, Deadline: time.Minute}})
	var buf bytes.Buffer
	w := multipart.NewWriter(&buf)
	if err := w.WriteField("genre", strings.Repeat("g", maxIntakeField+1)); err != nil {
		t.Fatal(err)
	}
	if err := w.Close(); err != nil {
		t.Fatal(err)
	}
	if got := upload(t, h, w.FormDataContentType(), &buf); got.Code != http.StatusBadRequest {
		t.Fatalf("status %d, want 400", got.Code)
	}
}

// The service's own refusals reach the wire as the contract's codes and never as engine or database
// text (§Problem).
func TestWhatTheIntakeRefusesBecomesAProductPhrase(t *testing.T) {
	cases := []struct {
		err  error
		want int
	}{
		{fmt.Errorf("%w: the languages must be codes", books.ErrBadIntake), http.StatusBadRequest},
		{pgstore.ErrNoAccount, http.StatusNotFound},
		{errors.New("pgstore: connection refused: dial tcp 127.0.0.1:5432"), http.StatusInternalServerError},
	}
	for _, tc := range cases {
		h := v0ServerWith(t, Deps{Library: &fakeLibrary{}, Intake: &fakeIntake{err: tc.err}})
		ct, body := form(t, "b.txt", 16, true)
		w := upload(t, h, ct, body)
		if w.Code != tc.want {
			t.Errorf("%v -> %d, want %d", tc.err, w.Code, tc.want)
		}
		if strings.Contains(w.Body.String(), "pgstore") || strings.Contains(w.Body.String(), "5432") {
			t.Errorf("the response carries internals: %s", w.Body)
		}
	}
}

// FP5-8(в) (acceptance): the cap allowed one part past its own number. The assertion is not the
// status — a form that simply runs out of parts is a 400 too, so a status-only test says nothing —
// but that the part BEHIND the cap is never handed to the intake. The file sits at exactly
// maxIntakeParts+1 here: with the cap off by one it is reached, accepted and streamed, which is the
// work the cap exists to refuse.
func TestAFormWithTooManyPartsIsRefusedAtTheCap(t *testing.T) {
	in := &fakeIntake{book: pgstore.Book{ID: "bk_cap"}}
	h := v0ServerWith(t, Deps{Library: &fakeLibrary{}, Intake: in,
		Upload: UploadLimits{MaxBytes: 8 << 20, Deadline: time.Minute}})
	var buf bytes.Buffer
	w := multipart.NewWriter(&buf)
	for k, v := range map[string]string{"source_lang": "zh", "target_lang": "ru"} {
		if err := w.WriteField(k, v); err != nil {
			t.Fatal(err)
		}
	}
	for i := range maxIntakeParts - 2 {
		if err := w.WriteField(fmt.Sprintf("filler-%d", i), "x"); err != nil {
			t.Fatal(err)
		}
	}
	fw, err := w.CreateFormFile("file", "book.txt")
	if err != nil {
		t.Fatal(err)
	}
	if _, err := fw.Write(bytes.Repeat([]byte("x"), 1024)); err != nil {
		t.Fatal(err)
	}
	if err := w.Close(); err != nil {
		t.Fatal(err)
	}
	got := upload(t, h, w.FormDataContentType(), &buf)
	if got.Code != http.StatusBadRequest {
		t.Fatalf("status %d with %d parts, want 400 at the cap of %d", got.Code, maxIntakeParts+1, maxIntakeParts)
	}
	if in.read != 0 {
		t.Fatalf("the file behind the cap was streamed anyway: %d bytes", in.read)
	}
}

// FP5-6 (acceptance): a body that outlives the route's own deadline is the client being slow, not
// this service being broken — and a 500 tells the client the opposite of the truth about retrying.
func TestAnUploadThatOutlivesItsDeadlineIsNotAnInternalError(t *testing.T) {
	h := v0ServerWith(t, Deps{Library: &fakeLibrary{}, Intake: &fakeIntake{err: os.ErrDeadlineExceeded}})
	ct, body := form(t, "book.txt", 64, true)
	w := upload(t, h, ct, body)
	if w.Code == http.StatusInternalServerError {
		t.Fatalf("an expired upload deadline answered 500: %s", w.Body)
	}
	if w.Code != http.StatusRequestTimeout {
		t.Fatalf("status %d, want 408", w.Code)
	}
	if w.Header().Get("Content-Type") != "application/problem+json" {
		t.Errorf("content-type %q", w.Header().Get("Content-Type"))
	}
}