feat: add governed knowledge lifecycle

2026-07-27 19:01:19 +08:00
parent 1a49ce4eae
commit 07c9ea1cc4
32 changed files with 5893 additions and 178 deletions
--- a/backend/apps/knowledge/tests/test_knowledge_api.py
+++ b/backend/apps/knowledge/tests/test_knowledge_api.py
@@ -1,4 +1,15 @@
-from apps.knowledge.models import KnowledgeBase, KnowledgeChunk, KnowledgeDocument, KnowledgeItem
+from django.core.files.uploadedfile import SimpleUploadedFile
+from django.test import override_settings
+
+from apps.knowledge.documents import process_knowledge_document
+from apps.knowledge.models import (
+    GovernedAsset,
+    KnowledgeBase,
+    KnowledgeChunk,
+    KnowledgeDocument,
+    KnowledgeItem,
+)
+from apps.tickets.materials import MaterialScanError


 def test_create_and_list_knowledge_base(admin_client, db):
@@ -23,7 +34,7 @@ def test_create_and_list_knowledge_base(admin_client, db):
    assert response.json()[0]["name"] == "税务采集知识库"


-def test_create_knowledge_item_builds_chunks(admin_client, db):
+def test_create_knowledge_item_always_creates_governed_draft(admin_client, db):
    base = KnowledgeBase.objects.create(name="税务采集知识库", scope="数据采集")

    response = admin_client.post(
@@ -45,8 +56,33 @@ def test_create_knowledge_item_builds_chunks(admin_client, db):

    assert response.status_code == 200
    item = KnowledgeItem.objects.get(id=response.json()["id"])
-    assert item.status == "published"
-    assert KnowledgeChunk.objects.filter(item=item).count() >= 1
+    assert item.status == "draft"
+    assert KnowledgeChunk.objects.filter(item=item).count() == 0
+    assert GovernedAsset.objects.filter(
+        asset_type="knowledge_item",
+        object_id=item.id,
+        versions__status="draft",
+    ).exists()
+
+
+def test_create_knowledge_item_rejects_unknown_category(admin_client, db):
+    base = KnowledgeBase.objects.create(name="税务采集知识库", scope="数据采集")
+
+    response = admin_client.post(
+        "/api/knowledge/items",
+        data={
+            "knowledge_base_id": base.id,
+            "title": "未知分类知识",
+            "problem": "分类编码不存在",
+            "category_code": "unknown_category",
+            "solution": "不应写入知识库",
+        },
+        content_type="application/json",
+    )
+
+    assert response.status_code == 400
+    assert response.json()["detail"] == "知识适用分类不存在或已停用"
+    assert not KnowledgeItem.objects.filter(title="未知分类知识").exists()


 def test_ingest_document_text_creates_document_and_chunks(admin_client, db):
@@ -67,3 +103,223 @@ def test_ingest_document_text_creates_document_and_chunks(admin_client, db):
    document = KnowledgeDocument.objects.get(id=response.json()["id"])
    assert document.parse_status == "success"
    assert KnowledgeChunk.objects.filter(document=document).count() >= 1
+
+
+def test_upload_document_preserves_original_and_indexes_text(
+    admin_client,
+    db,
+    monkeypatch,
+):
+    base = KnowledgeBase.objects.create(name="文件知识库", scope="技术支持")
+    objects = {}
+    monkeypatch.setattr(
+        "apps.knowledge.documents.put_object",
+        lambda path, raw, mime_type: objects.update({path: raw}),
+    )
+    monkeypatch.setattr(
+        "apps.knowledge.documents.get_object",
+        lambda path: objects[path],
+    )
+    monkeypatch.setattr(
+        "apps.knowledge.api.dispatch_knowledge_document_processing",
+        lambda document_id: process_knowledge_document(document_id),
+    )
+
+    response = admin_client.post(
+        f"/api/knowledge/documents/upload?knowledge_base_id={base.id}",
+        data={
+            "file": SimpleUploadedFile(
+                "support.log",
+                "接口超时，请检查连接池。".encode(),
+                content_type="text/plain",
+            )
+        },
+    )
+
+    assert response.status_code == 200
+    payload = response.json()
+    document = KnowledgeDocument.objects.get(public_id=payload["public_id"])
+    assert document.scan_status == "clean"
+    assert document.parse_status == "success"
+    assert document.storage_path in objects
+    assert document.content_text == "接口超时，请检查连接池。"
+    assert document.sha256
+    assert document.file_size == len(objects[document.storage_path])
+    assert document.processing_attempts == 1
+    assert document.indexed_at is not None
+    assert document.chunks.count() >= 1
+    assert payload["has_original"] is True
+
+    listed = admin_client.get(f"/api/knowledge/documents?knowledge_base_id={base.id}")
+    assert listed.status_code == 200
+    assert listed.json()[0]["public_id"] == str(document.public_id)
+
+
+def test_upload_document_rejects_disguised_file(admin_client, db):
+    base = KnowledgeBase.objects.create(name="文件知识库")
+
+    response = admin_client.post(
+        f"/api/knowledge/documents/upload?knowledge_base_id={base.id}",
+        data={
+            "file": SimpleUploadedFile(
+                "fake.pdf",
+                b"not a pdf",
+                content_type="application/pdf",
+            )
+        },
+    )
+
+    assert response.status_code == 415
+    assert not KnowledgeDocument.objects.exists()
+
+
+def test_document_scan_failure_keeps_original_and_can_retry(
+    admin_client,
+    db,
+    monkeypatch,
+):
+    base = KnowledgeBase.objects.create(name="文件知识库")
+    objects = {}
+    monkeypatch.setattr(
+        "apps.knowledge.documents.put_object",
+        lambda path, raw, mime_type: objects.update({path: raw}),
+    )
+    monkeypatch.setattr(
+        "apps.knowledge.documents.get_object",
+        lambda path: objects[path],
+    )
+    monkeypatch.setattr(
+        "apps.knowledge.documents.scan_bytes",
+        lambda raw: (_ for _ in ()).throw(MaterialScanError("扫描服务不可用")),
+    )
+    monkeypatch.setattr(
+        "apps.knowledge.api.dispatch_knowledge_document_processing",
+        lambda document_id: process_knowledge_document(document_id),
+    )
+    response = admin_client.post(
+        f"/api/knowledge/documents/upload?knowledge_base_id={base.id}",
+        data={
+            "file": SimpleUploadedFile(
+                "support.txt",
+                "重试后可解析".encode(),
+                content_type="text/plain",
+            )
+        },
+    )
+    document = KnowledgeDocument.objects.get(public_id=response.json()["public_id"])
+
+    assert response.status_code == 200
+    assert document.scan_status == "failed"
+    assert document.parse_status == "failed"
+    assert document.storage_path in objects
+    assert document.chunks.count() == 0
+
+    monkeypatch.setattr(
+        "apps.knowledge.documents.scan_bytes",
+        lambda raw: {"status": "clean", "engine": "test"},
+    )
+    retried = admin_client.post(f"/api/knowledge/documents/{document.public_id}/retry")
+
+    assert retried.status_code == 200
+    document.refresh_from_db()
+    assert document.parse_status == "success"
+    assert document.processing_attempts == 2
+    assert document.chunks.count() >= 1
+    process_knowledge_document(document.id)
+    document.refresh_from_db()
+    assert document.processing_attempts == 2
+
+
+def test_upload_document_storage_failure_does_not_create_record(
+    admin_client,
+    db,
+    monkeypatch,
+):
+    base = KnowledgeBase.objects.create(name="文件知识库")
+    monkeypatch.setattr(
+        "apps.knowledge.documents.put_object",
+        lambda *args: (_ for _ in ()).throw(OSError("storage down")),
+    )
+
+    response = admin_client.post(
+        f"/api/knowledge/documents/upload?knowledge_base_id={base.id}",
+        data={
+            "file": SimpleUploadedFile(
+                "support.txt",
+                b"content",
+                content_type="text/plain",
+            )
+        },
+    )
+
+    assert response.status_code == 503
+    assert not KnowledgeDocument.objects.exists()
+
+
+def test_download_uploaded_document_returns_original(
+    admin_client,
+    db,
+    monkeypatch,
+):
+    base = KnowledgeBase.objects.create(name="文件知识库")
+    document = KnowledgeDocument.objects.create(
+        knowledge_base=base,
+        file_name="支持手册.txt",
+        file_type="text",
+        extension="txt",
+        mime_type="text/plain",
+        file_size=7,
+        sha256="a" * 64,
+        storage_path="knowledge-documents/manual.txt",
+        scan_status="clean",
+        parse_status="success",
+    )
+    monkeypatch.setattr("apps.knowledge.api.get_object", lambda path: b"content")
+
+    response = admin_client.get(f"/api/knowledge/documents/{document.public_id}/download")
+
+    assert response.status_code == 200
+    assert response.content == b"content"
+    assert response["Content-Type"] == "text/plain"
+    assert "attachment" in response["Content-Disposition"]
+
+
+def test_download_infected_document_is_blocked(admin_client, db, monkeypatch):
+    base = KnowledgeBase.objects.create(name="文件知识库")
+    document = KnowledgeDocument.objects.create(
+        knowledge_base=base,
+        file_name="可疑文件.txt",
+        storage_path="knowledge-documents/infected.txt",
+        scan_status="infected",
+        parse_status="failed",
+    )
+    monkeypatch.setattr(
+        "apps.knowledge.api.get_object",
+        lambda path: (_ for _ in ()).throw(AssertionError("隔离文件不应从对象存储读取")),
+    )
+
+    response = admin_client.get(f"/api/knowledge/documents/{document.public_id}/download")
+
+    assert response.status_code == 423
+
+
+@override_settings(KNOWLEDGE_EMBEDDING_ALLOW_HASH_FALLBACK=False)
+def test_knowledge_draft_creation_does_not_require_embedding_deployment(admin_client, db):
+    base = KnowledgeBase.objects.create(name="正式知识库")
+
+    response = admin_client.post(
+        "/api/knowledge/items",
+        data={
+            "knowledge_base_id": base.id,
+            "title": "订单超时",
+            "problem": "订单接口超时",
+            "solution": "检查下游连接池",
+            "status": "published",
+        },
+        content_type="application/json",
+    )
+
+    assert response.status_code == 200
+    item = KnowledgeItem.objects.get(title="订单超时")
+    assert item.status == "draft"
+    assert item.chunks.count() == 0