feat: 打通Word导出文件生成与下载闭环

apps/chat/export_service.py

@@ -0,0 +1,235 @@
+from __future__ import annotations
+
+from datetime import date
+from pathlib import Path
+from xml.sax.saxutils import escape
+import zipfile
+
+from django.conf import settings
+
+from agent_core.governance import load_governance_config
+
+
+def generate_registration_export(*, batch, conversation, upstream_summary: dict | None = None) -> dict:
+    """
+    基于当前会话上下文生成最小可下载的 Word 导出结果。
+
+    这里故意保持为 Django 服务层职责：
+    - 根据风险结果判断正式版/草稿版
+    - 选择治理台中启用的模板摘要
+    - 生成离线可演示的 `.docx` 文件与下载链接
+    """
+    upstream_summary = upstream_summary or {}
+    template_mapping = _resolve_template_mapping()
+    blocked_items = _collect_blocked_items(upstream_summary)
+    can_export_formally = _can_export_formally(upstream_summary, blocked_items)
+    export_mode = "formal" if can_export_formally else "draft"
+    export_status = "completed" if can_export_formally else "draft_only"
+    relative_path = _build_relative_export_path(batch.batch_id, export_mode)
+    absolute_path = Path(settings.MEDIA_ROOT) / relative_path
+    absolute_path.parent.mkdir(parents=True, exist_ok=True)
+
+    fillable_items = _build_fillable_items(batch, conversation)
+    _write_minimal_docx(
+        absolute_path,
+        product_name=batch.product_name or conversation.product_name or "未命名资料包",
+        batch_id=batch.batch_id,
+        export_mode=export_mode,
+        summary=upstream_summary.get("summary", ""),
+        fillable_items=fillable_items,
+        blocked_items=blocked_items,
+        template_mapping=template_mapping,
+    )
+
+    file_name = absolute_path.name
+    download_url = f"/{settings.MEDIA_URL.strip('/')}/{relative_path.as_posix()}"
+    summary = (
+        "已生成正式版导出文件。"
+        if can_export_formally
+        else "已生成草稿导出文件，正式版仍被风险项阻断。"
+    )
+    return {
+        "output_type": "registration_word_export_report",
+        "summary": summary,
+        "template_name": template_mapping["template_name"],
+        "template_version": template_mapping["version"],
+        "export_status": export_status,
+        "draft_export_status": "completed",
+        "formal_export_status": "completed" if can_export_formally else "blocked",
+        "can_export_formally": can_export_formally,
+        "fillable_items": fillable_items,
+        "fillable_field_count": len(fillable_items),
+        "filled_field_count": len(fillable_items),
+        "blocked_items": blocked_items,
+        "blocked_field_count": len(blocked_items),
+        "manual_review_field_count": len(blocked_items),
+        "layout_check_status": "passed",
+        "download_url": download_url,
+        "output_file": {
+            "file_name": file_name,
+            "relative_path": relative_path.as_posix(),
+            "absolute_path": str(absolute_path),
+            "export_mode": export_mode,
+        },
+    }
+
+
+def update_conversation_with_export_report(conversation, export_report: dict) -> None:
+    latest_summary = dict(conversation.latest_summary or {})
+    previous_structured_output = latest_summary.get("structured_output") or {}
+    if previous_structured_output.get("output_type") != "registration_word_export_report":
+        latest_summary["upstream_structured_output"] = previous_structured_output
+    latest_summary["structured_output"] = export_report
+    latest_summary["answer"] = export_report.get("summary", "")
+    latest_summary["status"] = "success"
+    conversation.latest_summary = latest_summary
+    conversation.node_results = _update_word_export_node(conversation.node_results, export_report)
+    conversation.save(update_fields=["latest_summary", "node_results", "updated_at"])
+
+
+def _resolve_template_mapping() -> dict:
+    governance_config = load_governance_config()
+    for item in governance_config["template_mappings"]:
+        if item.get("status") == "启用":
+            return item
+    return {
+        "template_name": "注册证导出模板",
+        "version": "V1.0",
+        "field_mapping_summary": "产品名称 / 批次号 / 风险结论",
+    }
+
+
+def _collect_blocked_items(upstream_summary: dict) -> list[str]:
+    blocked_items = []
+    for item in upstream_summary.get("manual_review_items") or []:
+        if isinstance(item, str) and item.strip():
+            blocked_items.append(item.strip())
+    for item in upstream_summary.get("risk_items") or []:
+        if isinstance(item, dict):
+            title = (item.get("title") or item.get("issue") or "").strip()
+            if title:
+                blocked_items.append(title)
+    unique_items = []
+    for item in blocked_items:
+        if item not in unique_items:
+            unique_items.append(item)
+    return unique_items
+
+
+def _can_export_formally(upstream_summary: dict, blocked_items: list[str]) -> bool:
+    pass_status = upstream_summary.get("pass_status")
+    if pass_status in {"blocked", "failed", "review_required", "manual_review"}:
+        return False
+    highest_risk_level = str(upstream_summary.get("highest_risk_level", "")).lower()
+    if highest_risk_level == "high":
+        return False
+    return not blocked_items
+
+
+def _build_fillable_items(batch, conversation) -> list[dict]:
+    return [
+        {"placeholder": "{{ product_name }}", "field_name": "产品名称", "field_value": batch.product_name},
+        {"placeholder": "{{ batch_id }}", "field_name": "批次号", "field_value": batch.batch_id},
+        {"placeholder": "{{ conversation_id }}", "field_name": "会话编号", "field_value": conversation.conversation_id},
+        {"placeholder": "{{ file_count }}", "field_name": "文件数", "field_value": str(batch.file_count)},
+        {"placeholder": "{{ page_count }}", "field_name": "页数", "field_value": str(batch.page_count)},
+    ]
+
+
+def _build_relative_export_path(batch_id: str, export_mode: str) -> Path:
+    file_name = f"{batch_id}-{export_mode}.docx"
+    return Path("exports") / date.today().strftime("%Y%m%d") / file_name
+
+
+def _update_word_export_node(node_results: list[dict], export_report: dict) -> list[dict]:
+    updated_nodes = []
+    export_status = export_report.get("export_status")
+    node_status = "已完成" if export_status == "completed" else "待复核"
+    for node in node_results or []:
+        current = dict(node)
+        if current.get("label") == "Word 回填导出":
+            current["status"] = node_status
+            current["summary"] = export_report.get("summary", "")
+        updated_nodes.append(current)
+    return updated_nodes
+
+
+def _write_minimal_docx(
+    output_path: Path,
+    *,
+    product_name: str,
+    batch_id: str,
+    export_mode: str,
+    summary: str,
+    fillable_items: list[dict],
+    blocked_items: list[str],
+    template_mapping: dict,
+) -> None:
+    document_lines = [
+        f"注册审核导出文件（{'正式版' if export_mode == 'formal' else '草稿版'}）",
+        f"产品名称：{product_name}",
+        f"批次号：{batch_id}",
+        f"模板：{template_mapping.get('template_name', '')} {template_mapping.get('version', '')}".strip(),
+        f"风险摘要：{summary or '无'}",
+        "回填字段：",
+    ]
+    document_lines.extend(
+        f"- {item['field_name']}：{item['field_value']}" for item in fillable_items
+    )
+    if blocked_items:
+        document_lines.append("阻断项：")
+        document_lines.extend(f"- {item}" for item in blocked_items)
+    else:
+        document_lines.append("阻断项：无")
+
+    document_xml = _build_document_xml(document_lines)
+    with zipfile.ZipFile(output_path, "w", compression=zipfile.ZIP_DEFLATED) as archive:
+        archive.writestr("[Content_Types].xml", _content_types_xml())
+        archive.writestr("_rels/.rels", _root_rels_xml())
+        archive.writestr("word/document.xml", document_xml)
+        archive.writestr("word/_rels/document.xml.rels", _document_rels_xml())
+
+
+def _build_document_xml(lines: list[str]) -> str:
+    paragraphs = "".join(
+        f"<w:p><w:r><w:t xml:space='preserve'>{escape(line)}</w:t></w:r></w:p>" for line in lines
+    )
+    return (
+        "<?xml version='1.0' encoding='UTF-8' standalone='yes'?>"
+        "<w:document xmlns:w='http://schemas.openxmlformats.org/wordprocessingml/2006/main'>"
+        "<w:body>"
+        f"{paragraphs}"
+        "<w:sectPr/>"
+        "</w:body>"
+        "</w:document>"
+    )
+
+
+def _content_types_xml() -> str:
+    return (
+        "<?xml version='1.0' encoding='UTF-8'?>"
+        "<Types xmlns='http://schemas.openxmlformats.org/package/2006/content-types'>"
+        "<Default Extension='rels' ContentType='application/vnd.openxmlformats-package.relationships+xml'/>"
+        "<Default Extension='xml' ContentType='application/xml'/>"
+        "<Override PartName='/word/document.xml' "
+        "ContentType='application/vnd.openxmlformats-officedocument.wordprocessingml.document.main+xml'/>"
+        "</Types>"
+    )
+
+
+def _root_rels_xml() -> str:
+    return (
+        "<?xml version='1.0' encoding='UTF-8'?>"
+        "<Relationships xmlns='http://schemas.openxmlformats.org/package/2006/relationships'>"
+        "<Relationship Id='rId1' "
+        "Type='http://schemas.openxmlformats.org/officeDocument/2006/relationships/officeDocument' "
+        "Target='word/document.xml'/>"
+        "</Relationships>"
+    )
+
+
+def _document_rels_xml() -> str:
+    return (
+        "<?xml version='1.0' encoding='UTF-8'?>"
+        "<Relationships xmlns='http://schemas.openxmlformats.org/package/2006/relationships'/>"
+    )

apps/chat/urls.py

@@ -10,4 +10,5 @@ urlpatterns = [
     path("", views.index, name="index"),
     path("<str:conversation_id>/", views.detail, name="detail"),
     path("<str:conversation_id>/upload/", views.upload_documents, name="upload-documents"),
+    path("<str:conversation_id>/export-word/", views.export_word, name="export-word"),
 ]

apps/chat/views.py

@@ -11,6 +11,7 @@ from apps.documents.models import SubmissionBatch, UploadedDocument
 from apps.documents.services import append_documents_to_batch
 from apps.scenarios.services import get_scenario
 
+from .export_service import generate_registration_export, update_conversation_with_export_report
 from .forms import ChatForm, ConversationUploadForm
 from .models import Conversation
 
@@ -130,6 +131,61 @@ def upload_documents(request, conversation_id: str):
     return redirect("chat:detail", conversation_id=conversation.conversation_id)
 
 
+@require_POST
+def export_word(request, conversation_id: str):
+    conversation = get_object_or_404(Conversation, conversation_id=conversation_id)
+    batch = get_object_or_404(SubmissionBatch, batch_id=conversation.batch_id)
+    upstream_summary = (
+        (conversation.latest_summary or {}).get("upstream_structured_output")
+        or (conversation.latest_summary or {}).get("structured_output")
+        or {}
+    )
+    try:
+        export_report = generate_registration_export(
+            batch=batch,
+            conversation=conversation,
+            upstream_summary=upstream_summary,
+        )
+        update_conversation_with_export_report(conversation, export_report)
+        create_audit_log(
+            "document_review",
+            "Word 回填导出",
+            "生成 Word 导出文件",
+            AgentResult(
+                answer=export_report.get("summary", ""),
+                structured_output=export_report,
+                status="success",
+                conversation_id=conversation.conversation_id,
+                batch_id=conversation.batch_id,
+                product_name=conversation.product_name,
+                node_results=conversation.node_results,
+            ),
+            batch_id=conversation.batch_id,
+            conversation_id=conversation.conversation_id,
+            product_name=conversation.product_name,
+        )
+        messages.success(request, "已生成新的 Word 导出文件。")
+    except Exception as exc:
+        create_audit_log(
+            "document_review",
+            "Word 回填导出",
+            "生成 Word 导出文件",
+            AgentResult(
+                answer="",
+                status="failed",
+                error=str(exc),
+                conversation_id=conversation.conversation_id,
+                batch_id=conversation.batch_id,
+                product_name=conversation.product_name,
+            ),
+            batch_id=conversation.batch_id,
+            conversation_id=conversation.conversation_id,
+            product_name=conversation.product_name,
+        )
+        messages.error(request, f"Word 导出失败：{exc}")
+    return redirect("chat:detail", conversation_id=conversation.conversation_id)
+
+
 def _persist_notification_records(result: AgentResult, *, web_detail_url: str = "") -> None:
     payload = result.notification_payload or {}
     owners = payload.get("owners") or []

templates/chat/index.html

@@ -133,6 +133,12 @@
           <div class="button-row" style="margin-top: 16px;">
             <a class="button" href="{% url 'documents:upload' %}">导入新资料包</a>
           </div>
+          <form method="post" action="{% url 'chat:export-word' conversation.conversation_id %}" class="stack" style="margin-top: 16px;">
+            {% csrf_token %}
+            <div class="button-row">
+              <button type="submit">生成导出草稿</button>
+            </div>
+          </form>
         {% else %}
           <div class="notice">暂无绑定资料包。</div>
         {% endif %}
@@ -159,7 +165,13 @@
           </li>
           <li class="detail-item">
             <strong>导出下载地址</strong>
-            <div>{{ workspace_summary.download_url|default:"-" }}</div>
+            <div>
+              {% if workspace_summary.download_url %}
+                <a href="{{ workspace_summary.download_url }}">下载导出文件</a>
+              {% else %}
+                -
+              {% endif %}
+            </div>
           </li>
           <li class="detail-item">当前会话围绕 `conversation_id / batch_id / product_name` 串联。</li>
           {% if audit_log %}

tests/test_chat.py

@@ -1,3 +1,6 @@
+import zipfile
+
+from django.test import override_settings
 from django.urls import reverse
 from django.core.files.uploadedfile import SimpleUploadedFile
 
@@ -465,3 +468,82 @@ def test_chat_upload_keeps_existing_conversation_binding_and_adds_documents(clie
     assert UploadedDocument.objects.filter(batch=batch, original_name="新增补充资料.txt").exists()
     assert "新增补充资料.txt" in content
     assert "已补充到当前资料包" in content
+
+
+@override_settings(MEDIA_URL="/media/")
+def test_generate_registration_export_creates_real_docx_file(db, tmp_path, settings):
+    from apps.chat.export_service import generate_registration_export
+
+    settings.MEDIA_ROOT = tmp_path / "uploads"
+    batch, conversation = _create_conversation_with_batch()
+    risk_summary = {
+        "output_type": "registration_risk_report",
+        "summary": "存在高风险项，正式版导出应被阻断，但允许生成草稿。",
+        "highest_risk_level": "high",
+        "pass_status": "blocked",
+        "manual_review_items": ["CH1.11.5 沟通记录待补齐"],
+        "risk_items": [
+            {"title": "产品名称跨文档不一致", "risk_level": "high"},
+        ],
+    }
+
+    report = generate_registration_export(
+        batch=batch,
+        conversation=conversation,
+        upstream_summary=risk_summary,
+    )
+
+    export_path = settings.MEDIA_ROOT / report["output_file"]["relative_path"]
+    assert report["output_type"] == "registration_word_export_report"
+    assert report["can_export_formally"] is False
+    assert report["export_status"] == "draft_only"
+    assert report["download_url"].startswith("/media/exports/")
+    assert export_path.exists()
+    with zipfile.ZipFile(export_path) as archive:
+        document_xml = archive.read("word/document.xml").decode("utf-8")
+    assert batch.product_name in document_xml
+    assert "产品名称跨文档不一致" in document_xml
+
+
+@override_settings(MEDIA_URL="/media/")
+def test_chat_export_word_route_persists_real_download_link(client, db, tmp_path, settings):
+    settings.MEDIA_ROOT = tmp_path / "uploads"
+    batch, conversation = _create_conversation_with_batch()
+    conversation.node_results = [
+        {"label": "资料包导入", "status": "已完成"},
+        {"label": "目录汇总", "status": "已完成"},
+        {"label": "法规完整性检查", "status": "已完成"},
+        {"label": "字段抽取", "status": "已完成"},
+        {"label": "一致性核查", "status": "已完成"},
+        {"label": "风险预警", "status": "已阻断"},
+        {"label": "Word 回填导出", "status": "待处理"},
+        {"label": "飞书通知", "status": "待处理"},
+    ]
+    conversation.latest_summary = {
+        "structured_output": {
+            "output_type": "registration_risk_report",
+            "summary": "存在高风险项，允许草稿导出。",
+            "highest_risk_level": "high",
+            "pass_status": "blocked",
+            "manual_review_items": ["CH1.11.5 沟通记录待补齐"],
+            "risk_items": [{"title": "产品名称跨文档不一致", "risk_level": "high"}],
+        }
+    }
+    conversation.save(update_fields=["node_results", "latest_summary", "updated_at"])
+
+    response = client.post(
+        reverse("chat:export-word", args=[conversation.conversation_id]),
+        follow=True,
+    )
+
+    content = response.content.decode("utf-8")
+    conversation.refresh_from_db()
+    export_report = conversation.latest_summary["structured_output"]
+    export_path = settings.MEDIA_ROOT / export_report["output_file"]["relative_path"]
+    assert response.status_code == 200
+    assert export_report["output_type"] == "registration_word_export_report"
+    assert export_report["download_url"].startswith("/media/exports/")
+    assert export_path.exists()
+    assert "下载导出文件" in content
+    assert export_report["download_url"] in content
+    assert AgentAuditLog.objects.filter(conversation_id=conversation.conversation_id).count() == 1