record-asr/internal/server/upload_large_test.go

package server

import (
	"bytes"
	"crypto/sha256"
	"encoding/hex"
	"encoding/json"
	"io"
	"net/http"
	"net/http/httptest"
	"os"
	"path/filepath"
	"strconv"
	"testing"

	"git.misaka.ren/m1saka/record-asr/internal/config"
	"git.misaka.ren/m1saka/record-asr/internal/store"
)

// posReader emits a deterministic, position-dependent byte stream so that no
// two slices of the generated file are identical — a reorder would then
// change the reassembled hash and be caught.
type posReader struct{ pos int64 }

func (r *posReader) Read(p []byte) (int, error) {
	for i := range p {
		x := r.pos + int64(i)
		p[i] = byte((x*127 ^ (x >> 7)) & 0xff)
	}
	r.pos += int64(len(p))
	return len(p), nil
}

// sha256sum returns the hex SHA-256 of a file, streaming so 120MB never sits
// in memory at once.
func sha256sum(t *testing.T, path string) string {
	t.Helper()
	f, err := os.Open(path)
	if err != nil {
		t.Fatalf("open %s: %v", path, err)
	}
	defer f.Close()
	h := sha256.New()
	if _, err := io.Copy(h, f); err != nil {
		t.Fatalf("hash %s: %v", path, err)
	}
	return hex.EncodeToString(h.Sum(nil))
}

// TestChunkedUploadLargeFile uploads a 120MB file — larger than the 100MB CDN
// body cap the feature exists to bypass — as 25MB slices and checks that the
// reassembled file is byte-identical to the original. Guarded so the default
// test suite stays fast; run with:
//
//	ASR_TEST_LARGE=1 go test -run Large ./internal/server/...
func TestChunkedUploadLargeFile(t *testing.T) {
	if os.Getenv("ASR_TEST_LARGE") == "" {
		t.Skip("set ASR_TEST_LARGE=1 to run the large-file chunked upload test (~120MB on disk)")
	}
	const total = 120 << 20 // 120 MiB — over the 100MB cap, no per-request limit here
	const chunk = 25 << 20  // 25 MiB slices → 5 chunks (last is 20 MiB, partial)

	dir := t.TempDir()
	st, err := store.Open(dir)
	if err != nil {
		t.Fatalf("store.Open: %v", err)
	}
	cfg := config.Config{DataDir: dir, ChunkSize: chunk, MaxUploadBytes: 512 << 20, MaxRetries: 3}
	srv := New(cfg, st, nil, nil) // upload path never touches the transcribe/model stack
	ts := httptest.NewServer(srv.Router())
	t.Cleanup(ts.Close)

	// 1. Generate the 120MB original on disk, hashing as we go.
	origPath := filepath.Join(dir, "original.wav")
	of, err := os.Create(origPath)
	if err != nil {
		t.Fatalf("create orig: %v", err)
	}
	if _, err := io.Copy(of, io.LimitReader(&posReader{}, total)); err != nil {
		t.Fatalf("write orig: %v", err)
	}
	of.Close()
	wantHash := sha256sum(t, origPath)

	// 2. Open a chunked-upload session.
	body, _ := json.Marshal(map[string]any{"filename": "large.wav", "size": total})
	resp, err := http.Post(ts.URL+"/api/uploads", "application/json", bytes.NewReader(body))
	if err != nil {
		t.Fatalf("init: %v", err)
	}
	var init struct {
		UploadID    string `json:"upload_id"`
		ChunkSize   int64  `json:"chunk_size"`
		TotalChunks int    `json:"total_chunks"`
	}
	json.NewDecoder(resp.Body).Decode(&init)
	resp.Body.Close()
	if init.UploadID == "" || init.ChunkSize != chunk || init.TotalChunks != 5 {
		t.Fatalf("init = %+v", init)
	}

	// 3. Stream each slice straight from the original file (no full copy in RAM).
	src, err := os.Open(origPath)
	if err != nil {
		t.Fatalf("open src: %v", err)
	}
	defer src.Close()
	for i := 0; i < init.TotalChunks; i++ {
		offset := int64(i) * chunk
		size := int64(chunk)
		if offset+size > total {
			size = total - offset
		}
		req, _ := http.NewRequest("POST",
			ts.URL+"/api/uploads/"+init.UploadID+"?index="+strconv.Itoa(i),
			io.NewSectionReader(src, offset, size))
		req.Header.Set("Content-Type", "application/octet-stream")
		cresp, cerr := http.DefaultClient.Do(req)
		if cerr != nil {
			t.Fatalf("chunk %d: %v", i, cerr)
		}
		if cresp.StatusCode != http.StatusOK {
			t.Fatalf("chunk %d status = %d", i, cresp.StatusCode)
		}
		cresp.Body.Close()
	}

	// 4. Complete → 202 + pending task.
	req, _ := http.NewRequest("POST", ts.URL+"/api/uploads/"+init.UploadID+"/complete", nil)
	resp, err = http.DefaultClient.Do(req)
	if err != nil {
		t.Fatalf("complete: %v", err)
	}
	defer resp.Body.Close()
	if resp.StatusCode != http.StatusAccepted {
		t.Fatalf("complete status = %d", resp.StatusCode)
	}
	var task store.Task
	json.NewDecoder(resp.Body).Decode(&task)
	if task.Status != store.StatusPending {
		t.Fatalf("task status = %q, want %q", task.Status, store.StatusPending)
	}

	// 5. The reassembled file must be byte-identical (same SHA-256) and exact size.
	gotHash := sha256sum(t, st.AudioPath(&task))
	if gotHash != wantHash {
		t.Fatalf("reassembled file hash mismatch:\n want %s\n got  %s", wantHash, gotHash)
	}
	fi, err := os.Stat(st.AudioPath(&task))
	if err != nil || fi.Size() != total {
		t.Fatalf("final size = %d (%v), want %d", fi.Size(), err, total)
	}
}