feat: add natural elder story companion flow

2026-09-06 20:16:01 +08:00
parent 029a6097a2
commit 8c834c0aa7
20 changed files with 582 additions and 149 deletions
@@ -57,17 +57,19 @@ func (DeterministicGenerator) Generate(_ context.Context, input GenerationInput)
 		}, nil
 	}
 	storyTitle, storyContent, factContent := "", "", ""
+	openTopic := ""
 	if storyWorthy {
 		storyTitle = "关于“" + truncateRunes(content, 24) + "”的回忆"
 		storyContent = content
 		factContent = content
+		openTopic = "如果长者愿意，下次可以从刚才经历里一个尚未讲清的具体细节自然接着聊。"
 	}
 	return GenerationResult{
 		Meaningful:     true,
 		StoryWorthy:    storyWorthy,
 		DecisionReason: reason,
 		Summary:        "长者在本次通话中谈到：" + content,
-		OpenTopic:      "下次可以继续询问长者刚才提到的经历，并让长者自行补充人物、时间和地点。",
+		OpenTopic:      openTopic,
 		StoryTitle:     storyTitle,
 		StoryContent:   storyContent,
 		FactContent:    factContent,
@@ -90,7 +92,7 @@ func sensitiveGeneration(decision contentpolicy.Decision) GenerationResult {
 		StoryWorthy:    false,
 		DecisionReason: "safety_policy_blocked_story",
 		Summary:        "本次通话包含需要人工谨慎处理的敏感信息，系统未生成普通故事卡片。",
-		OpenTopic:      "如确有必要处理本次敏感内容，应由有权限人员在受控流程中单独核对。",
+		OpenTopic:      "",
 		Confidence:     1,
 		ModelRef:       modelRef,
 		PromptVersion:  "summary-sensitive-filter-v1:" + reason,
@@ -291,7 +293,7 @@ func validGeneratedContent(result GenerationResult) bool {
 		return !result.StoryWorthy && result.Summary == "" && result.OpenTopic == "" && result.StoryTitle == "" && result.StoryContent == "" && result.FactContent == "" && len(result.Keywords) == 0
 	}
 	if result.Summary == "" || len([]rune(result.Summary)) > 2000 ||
-		result.OpenTopic == "" || len([]rune(result.OpenTopic)) > 1000 ||
+		len([]rune(result.OpenTopic)) > 1000 ||
 		len(result.Keywords) < 1 || len(result.Keywords) > 8 {
 		return false
 	}
@@ -313,7 +315,7 @@ const contentValuePrompt = `你是长者聊天内容筛选与整理助手。只
 必须只返回JSON对象，字段为meaningful、story_worthy、decision_reason、summary、keywords、open_topic、story_title、story_content、fact_content、confidence。
 - decision_reason用一句简短中文说明判断依据。
 - meaningful=false时，除decision_reason和confidence外，其余文本必须为空字符串，keywords必须为空数组，story_worthy必须为false。
- meaningful=true时，summary概括有意义内容，keywords为1至8个简短关键词，open_topic给出一个温和且不重复追问的后续话题。
+- meaningful=true时，summary概括有意义内容，keywords为1至8个简短关键词。只有存在明确、自然且尚未聊完的线索时，open_topic才给出一个温和的后续方向；没有合适线索时必须为空字符串，不要为了填字段编造问题。
 - story_worthy=false时，story_title、story_content、fact_content必须为空字符串。
 - story_worthy=true时，story_title和story_content忠实整理原话，fact_content只写对话中可核对的事实。
 - confidence取0到1。`
@@ -63,6 +63,21 @@ func TestCompatibleGeneratorKeepsMeaningfulNonStoryWithoutDraft(t *testing.T) {
 	}
 }

+func TestCompatibleGeneratorAllowsMeaningfulContentWithoutForcedOpenTopic(t *testing.T) {
+	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
+		_ = json.NewEncoder(w).Encode(map[string]any{"choices": []any{map[string]any{"message": map[string]any{"content": `{"meaningful":true,"story_worthy":false,"decision_reason":"记录了明确的生活偏好，但没有自然的未完话题","summary":"长者喜欢安静地听越剧","keywords":["越剧","偏好"],"open_topic":"","story_title":"","story_content":"","fact_content":"","confidence":0.88}`}}}})
+	}))
+	defer server.Close()
+	generator, err := NewCompatibleGenerator("openai_compatible", server.URL+"/v1", "local-key", "qwen-test", 0.1, 512, time.Second)
+	if err != nil {
+		t.Fatal(err)
+	}
+	result, err := generator.Generate(context.Background(), GenerationInput{Content: "长者：我平时喜欢安静地听越剧。"})
+	if err != nil || !result.Meaningful || result.OpenTopic != "" {
+		t.Fatalf("result=%+v err=%v", result, err)
+	}
+}
+
 func TestCompatibleGeneratorClearsContradictoryStoryFields(t *testing.T) {
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
 		_ = json.NewEncoder(w).Encode(map[string]any{"choices": []any{map[string]any{"message": map[string]any{"content": `{"meaningful":true,"story_worthy":false,"decision_reason":"只是近况","summary":"今天胃口不错","keywords":["胃口"],"open_topic":"最近喜欢吃什么？","story_title":"模型误填标题","story_content":"模型误填故事","fact_content":"模型误填事实","confidence":0.8}`}}}})
@@ -96,7 +96,10 @@ func (p *Processor) Handle(ctx context.Context, event outbox.Event) error {
 		// not cycle through retry and dead-letter queues forever.
 		return nil
 	}
-	content := joinedConversationContent(snapshot.turns)
+	// Only the subject's final turns can become a summary, story, fact, or
+	// continuation cue. Agent turns may contain questions or generated wording
+	// and must never be promoted into the subject's personal history.
+	content := joinedSubjectContent(sourceTurns)
 	generated, err := p.generateSafely(ctx, content, snapshot.tenantID)
 	if err != nil {
 		return err
@@ -250,14 +253,13 @@ func subjectTurns(turns []turn) []turn {
 	return items
 }

-func joinedConversationContent(turns []turn) string {
+func joinedSubjectContent(turns []turn) string {
 	parts := make([]string, 0, len(turns))
 	for _, item := range turns {
-		label := "豆包"
-		if item.speaker == "subject" {
-			label = "长者"
+		if item.speaker != "subject" {
+			continue
 		}
-		parts = append(parts, label+"："+item.text)
+		parts = append(parts, "长者："+item.text)
 	}
 	return strings.Join(parts, "\n")
 }
@@ -284,11 +286,13 @@ func storeGeneration(ctx context.Context, tx pgx.Tx, conversationID string, snap
 		ON CONFLICT (conversation_id,summary_type,generation_version) DO NOTHING`, snapshot.tenantID, snapshot.subjectID, conversationID, summaryWithKeywords(generated.Summary, generated.Keywords), sourceJSON, generated.ModelRef, generated.PromptVersion, generationVersion, generated.Confidence); err != nil {
 		return fmt.Errorf("store conversation summary: %w", err)
 	}
-	if _, err := tx.Exec(ctx, `
-		INSERT INTO conversation_summary (tenant_id,subject_id,conversation_id,summary_type,content,source_turn_ids,model_ref,prompt_version,generation_version,confidence,status)
-		VALUES ($1::uuid,$2::uuid,$3::uuid,'open_topic',$4,$5::jsonb,$6,$7,$8,$9,'READY')
-		ON CONFLICT (conversation_id,summary_type,generation_version) DO NOTHING`, snapshot.tenantID, snapshot.subjectID, conversationID, generated.OpenTopic, sourceJSON, generated.ModelRef, generated.PromptVersion, generationVersion, generated.Confidence); err != nil {
-		return fmt.Errorf("store open topic: %w", err)
+	if strings.TrimSpace(generated.OpenTopic) != "" {
+		if _, err := tx.Exec(ctx, `
+			INSERT INTO conversation_summary (tenant_id,subject_id,conversation_id,summary_type,content,source_turn_ids,model_ref,prompt_version,generation_version,confidence,status)
+			VALUES ($1::uuid,$2::uuid,$3::uuid,'open_topic',$4,$5::jsonb,$6,$7,$8,$9,'READY')
+			ON CONFLICT (conversation_id,summary_type,generation_version) DO NOTHING`, snapshot.tenantID, snapshot.subjectID, conversationID, generated.OpenTopic, sourceJSON, generated.ModelRef, generated.PromptVersion, generationVersion, generated.Confidence); err != nil {
+			return fmt.Errorf("store open topic: %w", err)
+		}
 	}
 	if !generated.AllowStory {
 		return tx.Commit(ctx)
@@ -2,6 +2,7 @@ package extraction

 import (
 	"context"
+	"strings"
 	"testing"

 	"data-terminal-cloud/internal/contentpolicy"
@@ -41,3 +42,17 @@ func TestGenerateSafelyBlocksModelOutputBeforePersistence(t *testing.T) {
 		t.Fatalf("result=%+v generator_calls=%d classifier_calls=%d", result, generator.calls, classifier.calls)
 	}
 }
+
+func TestJoinedSubjectContentCannotPromoteAgentClaims(t *testing.T) {
+	content := joinedSubjectContent([]turn{
+		{speaker: "agent", text: "您年轻时一定在铁路上工作过。"},
+		{speaker: "subject", text: "没有，我以前在学校教书。"},
+		{speaker: "agent", text: "那您一定教了四十年。"},
+	})
+	if strings.Contains(content, "铁路") || strings.Contains(content, "四十年") {
+		t.Fatalf("agent claims entered extraction content: %q", content)
+	}
+	if content != "长者：没有，我以前在学校教书。" {
+		t.Fatalf("subject extraction content=%q", content)
+	}
+}