mirror of
https://github.com/infiniflow/ragflow.git
synced 2026-07-08 04:22:20 +08:00
Replaces the Python agent canvas runtime with a Go implementation that runs inside `cmd/server_main`. The canvas compiles into an eino Workflow that pauses on wait-for-user via native Interrupt/Resume (no sentinel flag) and resumes from a Redis-backed CheckPointStore. All 21 Python agent components and ~35 tools are ported with functional parity. Sandbox providers now read their JSON config from the admin-panel system_settings table with env fallback. 234 files / +35,413 / -6,111. All Go files are gofmt-clean (CI gate added); drops the v2 DSL E2E step and the gap-analysis plan (both redundant after the port ships). ## Type of change - [x] Refactoring - [x] New feature - [x] Bug fix 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude <noreply@anthropic.com>
177 lines
6.0 KiB
Go
177 lines
6.0 KiB
Go
//
|
|
// Copyright 2026 The InfiniFlow Authors. All Rights Reserved.
|
|
//
|
|
// Licensed under the Apache License, Version 2.0 (the "License");
|
|
// you may not use this file except in compliance with the License.
|
|
// You may obtain a copy of the License at
|
|
//
|
|
// http://www.apache.org/licenses/LICENSE-2.0
|
|
//
|
|
// Unless required by applicable law or agreed to in writing, software
|
|
// distributed under the License is distributed on an "AS IS" BASIS,
|
|
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
// See the License for the specific language governing permissions and
|
|
// limitations under the License.
|
|
//
|
|
|
|
package component
|
|
|
|
import (
|
|
"context"
|
|
"testing"
|
|
|
|
"github.com/cloudwego/eino/schema"
|
|
|
|
"ragflow/internal/agent/runtime"
|
|
)
|
|
|
|
// TestPrependHistory_EmptyHistory: no history → no prepend.
|
|
func TestPrependHistory_EmptyHistory(t *testing.T) {
|
|
current := []schema.Message{{Role: schema.User, Content: "hi"}}
|
|
out := prependHistory(current, nil, 5)
|
|
if len(out) != 1 {
|
|
t.Errorf("expected 1 message, got %d", len(out))
|
|
}
|
|
}
|
|
|
|
// TestPrependHistory_WindowZero: window=0 → no prepend.
|
|
func TestPrependHistory_WindowZero(t *testing.T) {
|
|
current := []schema.Message{{Role: schema.User, Content: "hi"}}
|
|
hist := []map[string]any{
|
|
{"role": "user", "content": "older"},
|
|
}
|
|
out := prependHistory(current, hist, 0)
|
|
if len(out) != 1 {
|
|
t.Errorf("expected 1 message, got %d", len(out))
|
|
}
|
|
}
|
|
|
|
// TestPrependHistory_AllWithinWindow: history shorter than window → all kept.
|
|
func TestPrependHistory_AllWithinWindow(t *testing.T) {
|
|
current := []schema.Message{{Role: schema.User, Content: "now"}}
|
|
hist := []map[string]any{
|
|
{"role": "user", "content": "turn 1"},
|
|
{"role": "assistant", "content": "reply 1"},
|
|
{"role": "user", "content": "turn 2"},
|
|
}
|
|
out := prependHistory(current, hist, 10)
|
|
if len(out) != 4 {
|
|
t.Fatalf("expected 4 messages, got %d", len(out))
|
|
}
|
|
if out[0].Content != "turn 1" || out[0].Role != "user" {
|
|
t.Errorf("first history entry wrong: %+v", out[0])
|
|
}
|
|
if out[1].Role != "assistant" {
|
|
t.Errorf("second entry should be assistant: %+v", out[1])
|
|
}
|
|
if out[3].Content != "now" {
|
|
t.Errorf("current message should be last: %+v", out[3])
|
|
}
|
|
}
|
|
|
|
// TestPrependHistory_TruncatesToWindow: history longer than window → keep last N.
|
|
func TestPrependHistory_TruncatesToWindow(t *testing.T) {
|
|
current := []schema.Message{{Role: schema.User, Content: "now"}}
|
|
hist := []map[string]any{
|
|
{"role": "user", "content": "turn 1"},
|
|
{"role": "assistant", "content": "reply 1"},
|
|
{"role": "user", "content": "turn 2"},
|
|
{"role": "assistant", "content": "reply 2"},
|
|
{"role": "user", "content": "turn 3"},
|
|
}
|
|
out := prependHistory(current, hist, 2)
|
|
if len(out) != 3 {
|
|
t.Fatalf("expected 3 messages (2 history + current), got %d", len(out))
|
|
}
|
|
// Should keep the last 2: turn 2 (user) and turn 3 (user), plus current.
|
|
if out[0].Content != "reply 2" {
|
|
t.Errorf("expected first kept entry to be 'reply 2' (the 4th of 5 with window=2), got %q", out[0].Content)
|
|
}
|
|
if out[1].Content != "turn 3" {
|
|
t.Errorf("expected second kept entry to be 'turn 3', got %q", out[1].Content)
|
|
}
|
|
if out[2].Content != "now" {
|
|
t.Errorf("current should be last: %+v", out[2])
|
|
}
|
|
}
|
|
|
|
// TestPrependHistory_SkipsInvalidEntries: entries missing role or content are skipped.
|
|
func TestPrependHistory_SkipsInvalidEntries(t *testing.T) {
|
|
current := []schema.Message{{Role: schema.User, Content: "now"}}
|
|
hist := []map[string]any{
|
|
{"role": "user"}, // missing content
|
|
{"content": "no role"}, // missing role
|
|
{"role": "user", "content": "valid"},
|
|
}
|
|
out := prependHistory(current, hist, 10)
|
|
if len(out) != 2 {
|
|
t.Fatalf("expected 2 (1 valid history + current), got %d", len(out))
|
|
}
|
|
if out[0].Content != "valid" {
|
|
t.Errorf("expected valid entry to be kept; got %+v", out[0])
|
|
}
|
|
}
|
|
|
|
// TestLLM_Invoke_HistoryWindow_PrependsFromState: end-to-end — when a
|
|
// canvas state carries history and the LLM is configured with a
|
|
// non-zero window size, the prior turns are prepended.
|
|
func TestLLM_Invoke_HistoryWindow_PrependsFromState(t *testing.T) {
|
|
stub := &stubInvoker{resp: &ChatInvokeResponse{Content: "ok", Model: "echo"}}
|
|
withStubInvoker(t, stub)
|
|
|
|
state := runtime.NewCanvasState("rid", "tid")
|
|
state.History = []map[string]any{
|
|
{"role": "user", "content": "earlier 1"},
|
|
{"role": "assistant", "content": "earlier reply 1"},
|
|
}
|
|
window := 5
|
|
c := NewLLMComponent(LLMParam{
|
|
ModelID: "echo",
|
|
UserPrompt: "now",
|
|
MessageHistoryWindowSize: window,
|
|
})
|
|
ctx := runtime.WithState(context.Background(), state)
|
|
_, err := c.Invoke(ctx, map[string]any{})
|
|
if err != nil {
|
|
t.Fatalf("Invoke: %v", err)
|
|
}
|
|
if stub.captured == nil {
|
|
t.Fatal("invoker captured no request")
|
|
}
|
|
if len(stub.captured.Messages) != 3 {
|
|
t.Fatalf("expected 3 messages (2 history + 1 current), got %d",
|
|
len(stub.captured.Messages))
|
|
}
|
|
if stub.captured.Messages[0].Content != "earlier 1" {
|
|
t.Errorf("first msg should be 'earlier 1', got %q", stub.captured.Messages[0].Content)
|
|
}
|
|
if stub.captured.Messages[1].Content != "earlier reply 1" {
|
|
t.Errorf("second msg should be 'earlier reply 1', got %q", stub.captured.Messages[1].Content)
|
|
}
|
|
if stub.captured.Messages[2].Content != "now" {
|
|
t.Errorf("last msg should be current 'now', got %q", stub.captured.Messages[2].Content)
|
|
}
|
|
}
|
|
|
|
// TestLLM_Invoke_HistoryWindow_ZeroIsNoop: when window is 0, history is
|
|
// not prepended even if present in state.
|
|
func TestLLM_Invoke_HistoryWindow_ZeroIsNoop(t *testing.T) {
|
|
stub := &stubInvoker{resp: &ChatInvokeResponse{Content: "ok", Model: "echo"}}
|
|
withStubInvoker(t, stub)
|
|
|
|
state := runtime.NewCanvasState("rid", "tid")
|
|
state.History = []map[string]any{
|
|
{"role": "user", "content": "should be ignored"},
|
|
}
|
|
c := NewLLMComponent(LLMParam{
|
|
ModelID: "echo",
|
|
UserPrompt: "now",
|
|
// MessageHistoryWindowSize: 0 (default)
|
|
})
|
|
ctx := runtime.WithState(context.Background(), state)
|
|
_, _ = c.Invoke(ctx, map[string]any{})
|
|
if len(stub.captured.Messages) != 1 {
|
|
t.Errorf("expected only 1 message (no history), got %d", len(stub.captured.Messages))
|
|
}
|
|
}
|