feat: persist passive browser activity batches

This commit is contained in:
2026-07-12 17:29:07 +08:00
parent 39052ffe39
commit bcf49ed32c
10 changed files with 787 additions and 43 deletions
+184 -8
View File
@@ -12,18 +12,20 @@ import (
"log"
"net"
"net/http"
"net/url"
"strings"
"sync"
"time"
"github.com/verstak/verstak-desktop/internal/core/events"
"github.com/verstak/verstak-desktop/internal/core/hostname"
)
const (
capturePath = "/api/browser-inbox/v1/captures"
activityBatchPath = "/api/browser-activity/v1/batches"
DefaultAddr = "127.0.0.1:47731"
DefaultCaptureURL = "http://" + DefaultAddr + capturePath
DefaultActivityURL = "http://" + DefaultAddr + activityBatchPath
receiverTokenHeader = "X-Verstak-Receiver-Token"
)
@@ -43,6 +45,9 @@ const (
maxFileTextBytes = 2 * 1024 * 1024
maxFileBytes = 8 * 1024 * 1024
maxFileDataBase64Bytes = 4 * ((maxFileBytes + 2) / 3)
maxActivityBodyBytes = 256 * 1024
maxActivityEntries = 100
maxActivityDuration = 10 * time.Minute
)
type Receiver struct {
@@ -55,10 +60,12 @@ type Receiver struct {
type WorkspaceProvider func() string
type Options struct {
RequireToken bool
ReceiverToken string
Available func() bool
Persist func(events.Event) error
RequireToken bool
ReceiverToken string
Available func() bool
Persist func(events.Event) error
ActivityAvailable func() bool
PersistActivity func(events.Event) error
}
type Server struct {
@@ -107,6 +114,23 @@ type CaptureBrowser struct {
Name string `json:"name"`
}
// ActivityBatchPayload contains only domain-level time accounting. It never
// accepts or emits page URLs, titles, content, or navigation history.
type ActivityBatchPayload struct {
SchemaVersion int `json:"schemaVersion"`
BatchID string `json:"batchId"`
CreatedAt string `json:"createdAt"`
Source string `json:"source"`
Entries []ActivityEntry `json:"entries"`
}
type ActivityEntry struct {
Hostname string `json:"hostname"`
StartedAt string `json:"startedAt"`
EndedAt string `json:"endedAt"`
DurationSeconds int64 `json:"durationSeconds"`
}
func New(bus *events.Bus, providers ...WorkspaceProvider) *Receiver {
return NewWithOptions(bus, Options{}, providers...)
}
@@ -141,6 +165,18 @@ func (r *Receiver) SetPersistence(available func() bool, persist func(events.Eve
r.options.Persist = persist
}
// SetActivityPersistence configures durable passive browser activity storage.
// An acknowledgement is sent only after persist returns successfully.
func (r *Receiver) SetActivityPersistence(available func() bool, persist func(events.Event) error) {
if r == nil {
return
}
r.optionsMu.Lock()
defer r.optionsMu.Unlock()
r.options.ActivityAvailable = available
r.options.PersistActivity = persist
}
func Start(addr string, receiver *Receiver) (*Server, error) {
if receiver == nil {
return nil, fmt.Errorf("receiver is required")
@@ -181,6 +217,10 @@ func (s *Server) Close() error {
func (r *Receiver) ServeHTTP(w http.ResponseWriter, req *http.Request) {
w.Header().Set("Content-Type", "application/json")
if req.URL.Path == activityBatchPath {
r.serveActivityBatch(w, req)
return
}
if req.URL.Path != capturePath {
http.NotFound(w, req)
return
@@ -256,6 +296,59 @@ func (r *Receiver) ServeHTTP(w http.ResponseWriter, req *http.Request) {
})
}
func (r *Receiver) serveActivityBatch(w http.ResponseWriter, req *http.Request) {
if req.Method != http.MethodPost {
w.WriteHeader(http.StatusMethodNotAllowed)
_ = json.NewEncoder(w).Encode(map[string]string{"error": "method not allowed"})
return
}
if err := r.validateReceiverToken(req); err != nil {
writeError(w, http.StatusUnauthorized, err.Error())
return
}
options := r.currentOptions()
if options.ActivityAvailable == nil || !options.ActivityAvailable() || options.PersistActivity == nil {
writeError(w, http.StatusServiceUnavailable, "activity storage unavailable")
return
}
defer req.Body.Close()
decoder := json.NewDecoder(http.MaxBytesReader(w, req.Body, maxActivityBodyBytes))
var batch ActivityBatchPayload
if err := decoder.Decode(&batch); err != nil {
writeActivityDecodeError(w, err)
return
}
if err := decoder.Decode(&struct{}{}); err != io.EOF {
writeActivityDecodeError(w, err)
return
}
if err := batch.NormalizeAndValidate(); err != nil {
writeError(w, http.StatusBadRequest, err.Error())
return
}
event := events.Event{
Name: "browser.activity.batch",
Timestamp: time.Now().UTC().Format(time.RFC3339Nano),
Payload: batch.EventPayload(),
}
if err := options.PersistActivity(event); err != nil {
log.Printf("[browserreceiver] persist activity %s: %v", batch.BatchID, err)
writeError(w, http.StatusServiceUnavailable, "activity storage unavailable")
return
}
if r.bus != nil {
r.bus.Publish(event)
}
w.WriteHeader(http.StatusAccepted)
_ = json.NewEncoder(w).Encode(map[string]string{
"status": "accepted",
"batchId": batch.BatchID,
})
}
func (r *Receiver) currentOptions() Options {
if r == nil {
return Options{}
@@ -370,6 +463,80 @@ func (p CapturePayload) Validate() error {
return nil
}
// NormalizeAndValidate validates a batch before it is made durable and replaces
// each submitted hostname with the shared canonical A-label form.
func (p *ActivityBatchPayload) NormalizeAndValidate() error {
if p == nil || p.SchemaVersion != 1 {
return fmt.Errorf("unsupported schemaVersion")
}
if strings.TrimSpace(p.BatchID) == "" {
return fmt.Errorf("batchId is required")
}
if err := validateCaptureText(p.BatchID, "batchId", maxCaptureIDBytes); err != nil {
return err
}
if strings.TrimSpace(p.CreatedAt) == "" {
return fmt.Errorf("createdAt is required")
}
if _, err := time.Parse(time.RFC3339, p.CreatedAt); err != nil {
return fmt.Errorf("createdAt is invalid")
}
if err := validateCaptureText(p.CreatedAt, "createdAt", maxCapturedAtBytes); err != nil {
return err
}
if strings.TrimSpace(p.Source) == "" {
return fmt.Errorf("source is required")
}
if err := validateCaptureText(p.Source, "source", maxCaptureSourceBytes); err != nil {
return err
}
if len(p.Entries) == 0 || len(p.Entries) > maxActivityEntries {
return fmt.Errorf("entries must contain between 1 and %d items", maxActivityEntries)
}
for index := range p.Entries {
entry := &p.Entries[index]
canonical := hostname.NormalizeHostnameV1(entry.Hostname)
if canonical == "" {
return fmt.Errorf("entries[%d].hostname is invalid", index)
}
startedAt, err := time.Parse(time.RFC3339, entry.StartedAt)
if err != nil {
return fmt.Errorf("entries[%d].startedAt is invalid", index)
}
endedAt, err := time.Parse(time.RFC3339, entry.EndedAt)
if err != nil {
return fmt.Errorf("entries[%d].endedAt is invalid", index)
}
interval := endedAt.Sub(startedAt)
if interval <= 0 || interval > maxActivityDuration {
return fmt.Errorf("entries[%d] interval must be between 1 second and %s", index, maxActivityDuration)
}
if entry.DurationSeconds <= 0 || entry.DurationSeconds > int64(maxActivityDuration/time.Second) || time.Duration(entry.DurationSeconds)*time.Second > interval {
return fmt.Errorf("entries[%d].durationSeconds is invalid", index)
}
entry.Hostname = canonical
}
return nil
}
func (p ActivityBatchPayload) EventPayload() map[string]interface{} {
entries := make([]map[string]interface{}, 0, len(p.Entries))
for _, entry := range p.Entries {
entries = append(entries, map[string]interface{}{
"hostname": entry.Hostname,
"startedAt": entry.StartedAt,
"endedAt": entry.EndedAt,
"durationSeconds": entry.DurationSeconds,
})
}
return map[string]interface{}{
"batchId": strings.TrimSpace(p.BatchID),
"createdAt": strings.TrimSpace(p.CreatedAt),
"source": strings.TrimSpace(p.Source),
"entries": entries,
}
}
func (p CapturePayload) validateFile() error {
if p.File == nil || strings.TrimSpace(p.File.Name) == "" {
return fmt.Errorf("file.name is required")
@@ -450,10 +617,19 @@ func (p CapturePayload) EventPayload() map[string]interface{} {
}
func captureDomain(rawURL, fallback string) string {
if u, err := url.Parse(strings.TrimSpace(rawURL)); err == nil && u.Hostname() != "" {
return u.Hostname()
if normalized := hostname.NormalizeURLHostnameV1(rawURL); normalized != "" {
return normalized
}
return strings.TrimSpace(fallback)
return hostname.NormalizeHostnameV1(fallback)
}
func writeActivityDecodeError(w http.ResponseWriter, err error) {
var maxBytesErr *http.MaxBytesError
if errors.As(err, &maxBytesErr) {
writeError(w, http.StatusRequestEntityTooLarge, "activity payload exceeds limit")
return
}
writeError(w, http.StatusBadRequest, "invalid JSON")
}
func writeError(w http.ResponseWriter, status int, message string) {
@@ -238,6 +238,101 @@ func TestReceiverLeavesCaptureUnassignedWithoutCurrentWorkspace(t *testing.T) {
}
}
func TestReceiverAcceptsDomainActivityBatchAfterDurablePersistence(t *testing.T) {
bus := events.NewBus()
received := make(chan events.Event, 1)
bus.Subscribe("browser.activity.batch", func(event events.Event) {
received <- event
})
persisted := 0
receiver := NewWithOptions(bus, Options{
ActivityAvailable: func() bool { return true },
PersistActivity: func(event events.Event) error {
persisted++
return nil
},
})
body := `{
"schemaVersion": 1,
"batchId": "batch-123",
"createdAt": "2026-07-12T10:05:00.000Z",
"source": "verstak-browser-extension",
"entries": [{
"hostname": "пример.рф",
"startedAt": "2026-07-12T10:00:00.000Z",
"endedAt": "2026-07-12T10:05:00.000Z",
"durationSeconds": 300
}]
}`
req := httptest.NewRequest(http.MethodPost, "/api/browser-activity/v1/batches", bytes.NewBufferString(body))
rec := httptest.NewRecorder()
receiver.ServeHTTP(rec, req)
if rec.Code != http.StatusAccepted {
t.Fatalf("status = %d, want %d; body=%s", rec.Code, http.StatusAccepted, rec.Body.String())
}
if persisted != 1 {
t.Fatalf("persisted = %d, want 1 before acknowledgement", persisted)
}
var response map[string]string
if err := json.Unmarshal(rec.Body.Bytes(), &response); err != nil {
t.Fatalf("response json: %v", err)
}
if response["batchId"] != "batch-123" || response["status"] != "accepted" {
t.Fatalf("response = %+v, want accepted batch-123", response)
}
event := <-received
payload, ok := event.Payload.(map[string]interface{})
if !ok {
t.Fatalf("event payload type = %T, want map[string]interface{}", event.Payload)
}
if payload["batchId"] != "batch-123" {
t.Fatalf("event payload = %+v, want batch id", payload)
}
entries, ok := payload["entries"].([]map[string]interface{})
if !ok || len(entries) != 1 || entries[0]["hostname"] != "xn--e1afmkfd.xn--p1ai" {
t.Fatalf("event entries = %+v, want one canonical hostname", payload["entries"])
}
if _, ok := payload["url"]; ok {
t.Fatalf("activity payload must not contain URL: %+v", payload)
}
}
func TestReceiverRejectsDomainActivityWithoutConsumerOrValidInterval(t *testing.T) {
for _, tc := range []struct {
name string
receiver *Receiver
body string
wantCode int
}{
{
name: "consumer unavailable",
receiver: New(events.NewBus()),
body: `{"schemaVersion":1,"batchId":"batch-a","createdAt":"2026-07-12T10:05:00Z","source":"verstak-browser-extension","entries":[{"hostname":"example.com","startedAt":"2026-07-12T10:00:00Z","endedAt":"2026-07-12T10:05:00Z","durationSeconds":300}]}`,
wantCode: http.StatusServiceUnavailable,
},
{
name: "duration too long",
receiver: NewWithOptions(events.NewBus(), Options{
ActivityAvailable: func() bool { return true },
PersistActivity: func(events.Event) error { return nil },
}),
body: `{"schemaVersion":1,"batchId":"batch-b","createdAt":"2026-07-12T10:05:00Z","source":"verstak-browser-extension","entries":[{"hostname":"example.com","startedAt":"2026-07-12T10:00:00Z","endedAt":"2026-07-12T10:20:01Z","durationSeconds":1201}]}`,
wantCode: http.StatusBadRequest,
},
} {
t.Run(tc.name, func(t *testing.T) {
req := httptest.NewRequest(http.MethodPost, "/api/browser-activity/v1/batches", strings.NewReader(tc.body))
rec := httptest.NewRecorder()
tc.receiver.ServeHTTP(rec, req)
if rec.Code != tc.wantCode {
t.Fatalf("status = %d, want %d; body=%s", rec.Code, tc.wantCode, rec.Body.String())
}
})
}
}
func TestReceiverRequiresTokenWhenPaired(t *testing.T) {
bus := events.NewBus()
received := make(chan events.Event, 1)