import (
"context"
"encoding/json"
- "errors"
"net/http"
"path/filepath"
"strings"
"echo: hi\n",
"reply: It said hi.",
}, test.Log(t))
+
require.Equal(t, []string{"user", "busy", "tool", "echo", "reply", "idle"}, test.events.Kinds())
var turns []string
+
for _, record := range modelReadUsage(t, test.usagePath) {
turns = append(turns, record.Turn)
}
require.NoError(t, err)
<-arrived
+
_, err = test.loop.Submit(terminalSource, "second, during the turn")
require.NoError(t, err)
close(release)
test.events.WaitFor(t, "idle", 1)
+
input := server.Call(1).body["input"].([]any)
require.Equal(t, map[string]any{
"role": "user",
require.NoError(t, err)
<-arrived
+
_, err = test.loop.Submit(terminalSource, "second")
require.NoError(t, err)
close(release)
require.NoError(t, err)
test.events.WaitFor(t, "idle", 2)
require.Equal(t, "reply: answer", test.Log(t)[2])
+ require.Empty(t, test.logs.String())
}
func TestFailedCallEndsTheTurnWithoutARetry(t *testing.T) {
server := newModelFake(t)
- server.reply = func(w http.ResponseWriter, call ModelFakeCall) {
+ server.reply = func(w http.ResponseWriter, _ ModelFakeCall) {
http.Error(w, "busy", http.StatusServiceUnavailable)
}
time.Sleep(50 * time.Millisecond)
require.Equal(t, 1, server.Calls())
require.Contains(t, test.events.Last("error"), "HTTP 503")
+ require.Equal(t, 1, strings.Count(test.logs.String(), "\n"))
+ require.Contains(t, test.logs.String(), "<3>turn failed: ")
+ require.Contains(t, test.logs.String(), "HTTP 503")
+ require.NotContains(t, test.logs.String(), "hello")
}
func TestViewMessagesOnlyGrowBetweenTurns(t *testing.T) {
first := server.Call(1).body["input"].([]any)
second := server.Call(2).body["input"].([]any)
+
require.Len(t, first, 4)
require.Len(t, second, 5)
require.Equal(t, first[:2], second[:2])
memory *Memory
loop *TurnLoop
events *TurnEventRecorder
+ logs *SyncBuffer
usagePath string
}
+// TurnEventRecorder keeps the events of a loop.
+type TurnEventRecorder struct {
+ mutex sync.Mutex
+ events []Event
+}
+
+func newTestTurnLoop(t *testing.T, server *ModelFake) *TurnTestLoop {
+ t.Helper()
+
+ return newTestTurnLoopWith(t, server, newFakeSummarizer())
+}
+
+// newTestTurnLoopWith returns a running loop on a new agent whose builds use
+// summarizer and whose commands run in a temporary directory.
+func newTestTurnLoopWith(t *testing.T, server *ModelFake, summarizer Summarizer) *TurnTestLoop {
+ t.Helper()
+
+ agent, err := openAgent(filepath.Join(t.TempDir(), "agent"))
+ require.NoError(t, err)
+
+ agent.config.Workdir = t.TempDir()
+ clock := newTestClock()
+ memory, err := openMemory(t.Context(), t.TempDir(), clock.Now, summarizer, discardLog())
+ require.NoError(t, err)
+
+ test := &TurnTestLoop{
+ agent: agent,
+ memory: memory,
+ events: &TurnEventRecorder{},
+ logs: &SyncBuffer{},
+ usagePath: filepath.Join(t.TempDir(), "usage.jsonl"),
+ }
+
+ client := newModelClient(server.URL, ModelAPIKey(modelTestKey), agent.CacheKey(), test.usagePath)
+ whatsapp, _, _ := whatsappStartTest(t, t.TempDir(), clock)
+ test.loop, err = newTurnLoop(agent, memory, newTestSubagents(t), whatsapp, newTestSchedules(t, clock.Now), client, clock.Now, daemonTestLogger(test.logs), test.events.Add)
+ require.NoError(t, err)
+
+ ctx, cancel := context.WithCancel(t.Context())
+ done := make(chan error)
+ go func() {
+ done <- test.loop.Run(ctx)
+ }()
+
+ t.Cleanup(func() {
+ cancel()
+ require.ErrorIs(t, <-done, context.Canceled)
+ require.NoError(t, memory.Close())
+ })
+
+ return test
+}
+
// Log returns the logged messages as "kind: text".
func (l *TurnTestLoop) Log(t *testing.T) []string {
t.Helper()
var lines []string
+
for id := range l.memory.Len() {
message, err := l.memory.Message(id)
require.NoError(t, err)
return lines
}
-// TurnEventRecorder keeps the events of a loop.
-type TurnEventRecorder struct {
- mutex sync.Mutex
- events []Event
-}
-
// WaitFor waits until count events of kind came.
func (r *TurnEventRecorder) WaitFor(t *testing.T, kind string, count int) {
t.Helper()
r.events = append(r.events, event)
}
-func newTestTurnLoop(t *testing.T, server *ModelFake) *TurnTestLoop {
- t.Helper()
-
- return newTestTurnLoopWith(t, server, newFakeSummarizer())
-}
-
// turnInputTexts returns the text of each user message of the input of call.
func turnInputTexts(t *testing.T, call ModelFakeCall) []string {
t.Helper()
require.NoError(t, json.Unmarshal(call.raw, &body))
var texts []string
+
for _, item := range body.Input {
if item.Role != "user" {
continue
}
var text strings.Builder
+
for _, part := range item.Content {
text.WriteString(part.Text)
}
return texts
}
-// newTestTurnLoopWith returns a running loop on a new agent whose builds use
-// summarizer and whose commands run in a temporary directory.
-func newTestTurnLoopWith(t *testing.T, server *ModelFake, summarizer Summarizer) *TurnTestLoop {
- t.Helper()
-
- agent, err := openAgent(t.TempDir())
- require.NoError(t, err)
-
- agent.config.Workdir = t.TempDir()
- clock := newTestClock()
- memory, err := openMemory(t.Context(), t.TempDir(), clock.Now, summarizer)
- require.NoError(t, err)
-
- test := &TurnTestLoop{
- agent: agent,
- memory: memory,
- events: &TurnEventRecorder{},
- usagePath: filepath.Join(t.TempDir(), "usage.jsonl"),
+func modelReasoningItem(id, encrypted string) map[string]any {
+ return map[string]any{
+ "type": "reasoning",
+ "id": id,
+ "encrypted_content": encrypted,
+ "summary": []any{},
}
-
- client := newModelClient(server.URL, ModelAPIKey(modelTestKey), agent.CacheKey(), test.usagePath)
- whatsapp, _, _ := whatsappStartTest(t, t.TempDir(), clock)
- test.loop, err = newTurnLoop(agent, memory, newTestSubagents(t), whatsapp, newTestSchedules(t, clock.Now), client, clock.Now, test.events.Add)
- require.NoError(t, err)
- ctx, cancel := context.WithCancel(t.Context())
- done := make(chan error)
- go func() {
- done <- test.loop.Run(ctx)
- }()
-
- t.Cleanup(func() {
- cancel()
- require.True(t, errors.Is(<-done, context.Canceled))
- require.NoError(t, memory.Close())
- })
-
- return test
}