fix(rag): retrieve complete course assignments

ai_knowledge_base_v2/apps/backend/tests/test_feishu_pagination.py

@@ -0,0 +1,54 @@
+from __future__ import annotations
+
+import asyncio
+from unittest.mock import AsyncMock, patch
+
+from app.services.feishu_async_service import _get_docx_content_async
+from app.services.feishu_service import FeishuRetrievalConfig, _get_docx_content
+
+
+def _config() -> FeishuRetrievalConfig:
+    return FeishuRetrievalConfig(
+        search_url="",
+        app_id="app-id",
+        app_secret="secret",
+        timeout_seconds=10,
+        retry_count=1,
+    )
+
+
+def _heading(text: str) -> dict:
+    return {
+        "block_type": 3,
+        "heading1": {"elements": [{"text_run": {"content": text}}]},
+    }
+
+
+def test_docx_content_reads_all_block_pages():
+    pages = [
+        {"data": {"items": [_heading("第一页")], "has_more": True, "page_token": "next-page"}},
+        {"data": {"items": [_heading("第二页作业")], "has_more": False, "page_token": ""}},
+    ]
+    with patch("app.services.feishu_service._feishu_get", side_effect=pages) as feishu_get:
+        content = _get_docx_content("doc-token", _config())
+
+    assert "第一页" in content
+    assert "第二页作业" in content
+    assert feishu_get.call_count == 2
+    assert feishu_get.call_args_list[0].kwargs["params"] == {"page_size": 500}
+    assert feishu_get.call_args_list[1].kwargs["params"] == {"page_size": 500, "page_token": "next-page"}
+
+
+def test_async_docx_content_reads_all_block_pages():
+    pages = [
+        {"data": {"items": [_heading("第一页")], "has_more": True, "page_token": "next-page"}},
+        {"data": {"items": [_heading("第二页作业")], "has_more": False, "page_token": ""}},
+    ]
+    feishu_get = AsyncMock(side_effect=pages)
+    with patch("app.services.feishu_async_service._feishu_get_async", feishu_get):
+        content = asyncio.run(_get_docx_content_async("doc-token", _config()))
+
+    assert "第一页" in content
+    assert "第二页作业" in content
+    assert feishu_get.await_count == 2
+    assert feishu_get.await_args_list[1].kwargs["params"] == {"page_size": 500, "page_token": "next-page"}

ai_knowledge_base_v2/apps/backend/tests/test_knowledge_agent.py

@@ -4,7 +4,7 @@ import asyncio
 import json
 from datetime import datetime
 
-from sqlalchemy import create_engine
+from sqlalchemy import create_engine, select
 from sqlalchemy.orm import Session
 from sqlalchemy.pool import StaticPool
 
@@ -141,6 +141,13 @@ def test_course_question_searches_chunk_and_reads_parent_section():
         assert len(result.chunks) == 1
         assert "先稳定自己的焦虑" in result.chunks[0].content
         assert any(item["tool"] == "read_knowledge_sections" for item in result.tool_trace)
+        knowledge_context = next(
+            message["content"] for message in result.messages if message["content"].startswith("[本轮可靠知识上下文]")
+        )
+        assert "不得再以‘没有资料’" in knowledge_context
+        assert "不要求知识库中必须存在一个" in knowledge_context
+        assert "不要把本轮已经取到的内容留到下一轮再问" in knowledge_context
+        assert "来源知识库：亲子课程" in knowledge_context
 
 
 def test_tool_trace_serializes_published_datetime():
@@ -173,3 +180,57 @@ def test_contextual_follow_up_is_rewritten_before_agent_decision():
 
         assert rewrite["rewrittenQuestion"] == "关于“课程退款条件有哪些？”的追问：那第二种情况呢？"
         assert decision["request"]["question"] == rewrite["rewrittenQuestion"]
+
+
+def test_homework_overview_expands_practice_terms_and_section_limit():
+    terms = KnowledgeAgentService._query_terms("合一的作业是什么？")
+
+    assert "功课" in terms
+    assert "练习" in terms
+    assert KnowledgeAgentService._selection_limit("合一的作业是什么？") == 8
+    assert KnowledgeAgentService._selection_limit("这个练习怎么做？") == 4
+    assert KnowledgeAgentService._title_intent_boost("十七、练习一：风铃式静心", terms) == 20.0
+    assert KnowledgeAgentService._title_intent_boost("完整练习的方向", terms) == 3.0
+    assert KnowledgeAgentService._title_intent_boost("课程定位", terms) == 0.0
+
+
+def test_complete_section_includes_child_headings_but_stops_at_next_peer():
+    with _database() as db:
+        knowledge = _add_published_knowledge(db, knowledge_id=1, name="合一课程")
+        version_id = knowledge.current_version_id
+        parent = db.scalar(select(KnowledgeSection).where(KnowledgeSection.version_id == version_id))
+        assert parent is not None
+        parent.title = "练习一：风铃式静心"
+        parent.content = "# 练习一：风铃式静心"
+        parent.sort_order = 1
+        db.add_all([
+            parent,
+            KnowledgeSection(
+                knowledge_id=knowledge.id,
+                version_id=version_id,
+                section_key="S0002",
+                title="基本操作",
+                content="## 基本操作\n鼻吸鼻呼，呼气时发出连续蜂鸣声。",
+                source_start=10,
+                source_end=40,
+                sort_order=2,
+                content_hash="child",
+            ),
+            KnowledgeSection(
+                knowledge_id=knowledge.id,
+                version_id=version_id,
+                section_key="S0003",
+                title="练习二：内感知建模",
+                content="# 练习二：内感知建模\n这是下一项作业。",
+                source_start=41,
+                source_end=70,
+                sort_order=3,
+                content_hash="next",
+            ),
+        ])
+        db.commit()
+
+        content = KnowledgeAgentService._read_complete_section(db, parent)
+
+        assert "鼻吸鼻呼" in content
+        assert "下一项作业" not in content