Files
narratio/internal/adapters/whisperx/multipart_body_test.go

328 lines
9.2 KiB
Go

package whisperx
import (
"bytes"
"context"
"errors"
"io"
"mime"
"mime/multipart"
"strings"
"sync"
"sync/atomic"
"testing"
"time"
)
func TestMultipartBodyReplayUsesIdenticalFraming(t *testing.T) {
factory := newTestMultipartBodyFactory(t, context.Background(), "audio-data")
defer factory.Close()
original, err := factory.NewBody()
if err != nil {
t.Fatalf("NewBody() original error = %v", err)
}
replay, err := factory.NewBody()
if err != nil {
t.Fatalf("NewBody() replay error = %v", err)
}
originalBytes, err := io.ReadAll(original)
if err != nil {
t.Fatalf("ReadAll(original) error = %v", err)
}
replayBytes, err := io.ReadAll(replay)
if err != nil {
t.Fatalf("ReadAll(replay) error = %v", err)
}
if !bytes.Equal(originalBytes, replayBytes) {
t.Fatal("original and replay multipart bytes differ")
}
parts := parseMultipartParts(t, factory.contentType, replayBytes)
if len(parts) != 2 {
t.Fatalf("part count = %d, want 2", len(parts))
}
if parts[0].name != "file" || parts[0].filename != "audio.flac" || string(parts[0].data) != "audio-data" {
t.Fatalf("file part = %#v, want file/audio.flac with exact audio", parts[0])
}
if parts[1].name != "language" || string(parts[1].data) != "en" {
t.Fatalf("language part = %#v, want language=en", parts[1])
}
}
func TestMultipartBodyReplayRestartsAfterPartialOriginal(t *testing.T) {
factory := newTestMultipartBodyFactory(t, context.Background(), "audio-data")
defer factory.Close()
original, err := factory.NewBody()
if err != nil {
t.Fatalf("NewBody() original error = %v", err)
}
partial := make([]byte, len(factory.prefix)+3)
if _, err := io.ReadFull(original, partial); err != nil {
t.Fatalf("ReadFull(original) error = %v", err)
}
if err := original.Close(); err != nil {
t.Fatalf("Close(original) error = %v", err)
}
replay, err := factory.NewBody()
if err != nil {
t.Fatalf("NewBody() replay error = %v", err)
}
replayBytes, err := io.ReadAll(replay)
if err != nil {
t.Fatalf("ReadAll(replay) error = %v", err)
}
parts := parseMultipartParts(t, factory.contentType, replayBytes)
if got := string(parts[0].data); got != "audio-data" {
t.Fatalf("replayed audio = %q, want complete payload", got)
}
if err := factory.SourceError(false); err != nil {
t.Fatalf("SourceError(false) = %v, want abandoned original ignored", err)
}
}
func TestMultipartBodySimultaneousBodiesUseIndependentSources(t *testing.T) {
var opens atomic.Int32
var closes atomic.Int32
factory, err := newMultipartBodyFactory(context.Background(), "audio.flac", "en", func(string) (io.ReadCloser, error) {
opens.Add(1)
return &countedReadCloser{Reader: strings.NewReader("audio-data"), closes: &closes}, nil
})
if err != nil {
t.Fatalf("newMultipartBodyFactory() error = %v", err)
}
first, err := factory.NewBody()
if err != nil {
t.Fatalf("NewBody() first error = %v", err)
}
second, err := factory.NewBody()
if err != nil {
t.Fatalf("NewBody() second error = %v", err)
}
var firstBytes, secondBytes []byte
var firstErr, secondErr error
var group sync.WaitGroup
group.Add(2)
go func() {
defer group.Done()
firstBytes, firstErr = io.ReadAll(first)
}()
go func() {
defer group.Done()
secondBytes, secondErr = io.ReadAll(second)
}()
group.Wait()
if firstErr != nil || secondErr != nil {
t.Fatalf("concurrent reads returned (%v, %v)", firstErr, secondErr)
}
if !bytes.Equal(firstBytes, secondBytes) {
t.Fatal("independent body bytes differ")
}
if got := opens.Load(); got != 2 {
t.Fatalf("audio opens = %d, want 2", got)
}
if err := factory.Close(); err != nil {
t.Fatalf("Close() error = %v", err)
}
if err := factory.Close(); err != nil {
t.Fatalf("second Close() error = %v", err)
}
if got := closes.Load(); got != 2 {
t.Fatalf("audio closes = %d, want 2", got)
}
}
func TestMultipartBodyReportsSourceFailures(t *testing.T) {
t.Run("read", func(t *testing.T) {
sourceErr := errors.New("read source")
factory, err := newMultipartBodyFactory(context.Background(), "audio.flac", "en", func(string) (io.ReadCloser, error) {
return &failingReadCloser{first: []byte("partial"), err: sourceErr}, nil
})
if err != nil {
t.Fatalf("newMultipartBodyFactory() error = %v", err)
}
body, err := factory.NewBody()
if err != nil {
t.Fatalf("NewBody() error = %v", err)
}
if _, err := io.ReadAll(body); !errors.Is(err, sourceErr) {
t.Fatalf("ReadAll() error = %v, want source failure", err)
}
_ = body.Close()
if err := factory.SourceError(false); !errors.Is(err, sourceErr) {
t.Fatalf("SourceError(false) = %v, want source failure", err)
}
})
t.Run("close after complete read", func(t *testing.T) {
closeErr := errors.New("close source")
factory, err := newMultipartBodyFactory(context.Background(), "audio.flac", "en", func(string) (io.ReadCloser, error) {
return &errorCloseReadCloser{Reader: strings.NewReader("audio-data"), err: closeErr}, nil
})
if err != nil {
t.Fatalf("newMultipartBodyFactory() error = %v", err)
}
body, err := factory.NewBody()
if err != nil {
t.Fatalf("NewBody() error = %v", err)
}
if _, err := io.ReadAll(body); err != nil {
t.Fatalf("ReadAll() error = %v", err)
}
if err := body.Close(); !errors.Is(err, closeErr) {
t.Fatalf("Close() error = %v, want source close failure", err)
}
if err := factory.SourceError(false); !errors.Is(err, closeErr) {
t.Fatalf("SourceError(false) = %v, want source close failure", err)
}
})
}
func TestMultipartBodyCloseAndCancelReleaseBlockedSource(t *testing.T) {
for _, test := range []struct {
name string
release func(context.CancelFunc, io.ReadCloser)
}{
{name: "close", release: func(_ context.CancelFunc, body io.ReadCloser) { _ = body.Close() }},
{name: "cancel", release: func(cancel context.CancelFunc, _ io.ReadCloser) { cancel() }},
} {
t.Run(test.name, func(t *testing.T) {
ctx, cancel := context.WithCancel(context.Background())
defer cancel()
source := newSignalingBlockedReadCloser()
factory, err := newMultipartBodyFactory(ctx, "audio.flac", "en", func(string) (io.ReadCloser, error) {
return source, nil
})
if err != nil {
t.Fatalf("newMultipartBodyFactory() error = %v", err)
}
defer factory.Close()
body, err := factory.NewBody()
if err != nil {
t.Fatalf("NewBody() error = %v", err)
}
readDone := make(chan error, 1)
go func() {
_, err := io.ReadAll(body)
readDone <- err
}()
select {
case <-source.blocked:
case <-time.After(time.Second):
t.Fatal("audio source did not block")
}
test.release(cancel, body)
select {
case <-source.closed:
case <-time.After(time.Second):
t.Fatal("audio source was not closed")
}
select {
case <-readDone:
case <-time.After(time.Second):
t.Fatal("blocked body read did not return")
}
if test.name == "cancel" {
if _, err := factory.NewBody(); !errors.Is(err, context.Canceled) {
t.Fatalf("NewBody() after cancellation error = %v, want context.Canceled", err)
}
}
})
}
}
func newTestMultipartBodyFactory(t *testing.T, ctx context.Context, audio string) *multipartBodyFactory {
t.Helper()
factory, err := newMultipartBodyFactory(ctx, "audio.flac", "en", func(string) (io.ReadCloser, error) {
return io.NopCloser(strings.NewReader(audio)), nil
})
if err != nil {
t.Fatalf("newMultipartBodyFactory() error = %v", err)
}
return factory
}
type multipartTestPart struct {
name string
filename string
data []byte
}
func parseMultipartParts(t *testing.T, contentType string, body []byte) []multipartTestPart {
t.Helper()
mediaType, params, err := mime.ParseMediaType(contentType)
if err != nil {
t.Fatalf("ParseMediaType() error = %v", err)
}
if mediaType != "multipart/form-data" {
t.Fatalf("media type = %q, want multipart/form-data", mediaType)
}
reader := multipart.NewReader(bytes.NewReader(body), params["boundary"])
var parts []multipartTestPart
for {
part, err := reader.NextPart()
if errors.Is(err, io.EOF) {
return parts
}
if err != nil {
t.Fatalf("NextPart() error = %v", err)
}
data, err := io.ReadAll(part)
if err != nil {
t.Fatalf("ReadAll(part %q) error = %v", part.FormName(), err)
}
parts = append(parts, multipartTestPart{name: part.FormName(), filename: part.FileName(), data: data})
}
}
type countedReadCloser struct {
io.Reader
closes *atomic.Int32
once sync.Once
}
func (reader *countedReadCloser) Close() error {
reader.once.Do(func() { reader.closes.Add(1) })
return nil
}
type errorCloseReadCloser struct {
io.Reader
err error
}
func (reader *errorCloseReadCloser) Close() error { return reader.err }
type signalingBlockedReadCloser struct {
blocked chan struct{}
closed chan struct{}
once sync.Once
read bool
}
func newSignalingBlockedReadCloser() *signalingBlockedReadCloser {
return &signalingBlockedReadCloser{blocked: make(chan struct{}), closed: make(chan struct{})}
}
func (reader *signalingBlockedReadCloser) Read(buffer []byte) (int, error) {
if !reader.read {
reader.read = true
return copy(buffer, "audio-data"), nil
}
reader.once.Do(func() { close(reader.blocked) })
<-reader.closed
return 0, errors.New("audio source closed")
}
func (reader *signalingBlockedReadCloser) Close() error {
select {
case <-reader.closed:
default:
close(reader.closed)
}
return nil
}