From 925ef0238d33f24799f2f0b5524add618f2732eb Mon Sep 17 00:00:00 2001
From: Nelson <1475262689@qq.com>
Date: Wed, 8 Jul 2026 12:29:21 +0800
Subject: [PATCH] Support Feishu file node retrieval

---
 .../backend/app/services/feishu_service.py    | 69 ++++++++++++++++++-
 1 file changed, 66 insertions(+), 3 deletions(-)

diff --git a/ai_knowledge_base_v2/apps/backend/app/services/feishu_service.py b/ai_knowledge_base_v2/apps/backend/app/services/feishu_service.py
index eba5be7..9ae0bc6 100644
--- a/ai_knowledge_base_v2/apps/backend/app/services/feishu_service.py
+++ b/ai_knowledge_base_v2/apps/backend/app/services/feishu_service.py
@@ -128,6 +128,24 @@ def _feishu_post(path: str, payload: dict | None = None, config: FeishuRetrieval
     return data
 
 
+def _feishu_download(path: str, config: FeishuRetrievalConfig) -> bytes:
+    token = _get_tenant_token(config)
+    url = f"{FEISHU_OPEN_API}{path}"
+    headers = {"Authorization": f"Bearer {token}"}
+
+    try:
+        resp = httpx.get(url, headers=headers, timeout=config.timeout_seconds)
+        resp.raise_for_status()
+        return resp.content
+    except httpx.HTTPStatusError as exc:
+        raise ExternalServiceError(
+            f"飞书文件下载失败 [{path}]：{exc.response.status_code} {_response_error_detail(exc.response)}",
+            provider="feishu",
+        ) from exc
+    except httpx.HTTPError as exc:
+        raise ExternalServiceError(f"飞书文件下载网络错误 [{path}]：{exc}", provider="feishu") from exc
+
+
 def _response_error_detail(response: httpx.Response) -> str:
     try:
         body = response.json()
@@ -220,6 +238,17 @@ def _get_docx_content(doc_token: str, config: FeishuRetrievalConfig) -> str:
     return _parse_docx_content(items)
 
 
+def _get_file_content(file_token: str, config: FeishuRetrievalConfig) -> str:
+    """下载普通文件节点内容，主要用于 Markdown/TXT 知识库文件。"""
+    content = _feishu_download(f"/open-apis/drive/v1/files/{file_token}/download", config)
+    for encoding in ("utf-8-sig", "utf-8", "gb18030"):
+        try:
+            return content.decode(encoding).strip()
+        except UnicodeDecodeError:
+            continue
+    raise ExternalServiceError("飞书文件不是可解析的文本内容", provider="feishu")
+
+
 def _get_wiki_node_info(node_token: str, config: FeishuRetrievalConfig) -> dict:
     """获取 Wiki 节点信息，返回 space_id 和 obj_token"""
     data = _feishu_get("/open-apis/wiki/v2/spaces/get_node", params={"token": node_token}, config=config)
@@ -378,6 +407,11 @@ class FeishuKnowledgeService:
                     provider="feishu",
                 )
             return []
+        if load_errors:
+            raise ExternalServiceError(
+                "部分飞书知识库读取失败：" + "；".join(load_errors[:3]),
+                provider="feishu",
+            )
 
         # 关键词匹配评分（简单但够用的检索方式）
         scored_chunks = cls._score_and_rank(question, all_chunks)
@@ -410,14 +444,17 @@ class FeishuKnowledgeService:
         # 1. 判断 node 类型：先查 wiki 节点信息
         try:
             node_info = cls._get_node_info_with_cache(node_id, config)
-        except ExternalServiceError:
+        except ExternalServiceError as node_exc:
             # 可能是普通 docx token，直接尝试读取
             try:
                 content = cls._get_document_content_with_cache(node_id, config)
                 title = node_id
                 return [(title, content, f"{base_url}/docx/{node_id}")]
-            except ExternalServiceError:
-                return []
+            except ExternalServiceError as docx_exc:
+                raise ExternalServiceError(
+                    f"飞书节点读取失败：{node_exc}；按普通 docx token 读取也失败：{docx_exc}",
+                    provider="feishu",
+                ) from docx_exc
 
         obj_type = node_info.get("obj_type", "")
         obj_token = node_info.get("obj_token", "")
@@ -429,6 +466,10 @@ class FeishuKnowledgeService:
             content = cls._get_document_content_with_cache(obj_token, config)
             source_url = f"{base_url}/wiki/{node_id}"
             return [(title, content, source_url)]
+        if obj_type == "file":
+            content = cls._get_file_content_with_cache(obj_token, config)
+            source_url = f"{base_url}/wiki/{node_id}"
+            return [(title, content, source_url)]
 
         # 如果节点是文件夹，获取子节点
         children = cls._get_wiki_children_with_cache(space_id, node_id, config)
@@ -447,6 +488,13 @@ class FeishuKnowledgeService:
                     documents.append((child_title, content, source_url))
                 except ExternalServiceError:
                     continue
+            elif child_obj_type == "file" and child_obj_token:
+                try:
+                    content = cls._get_file_content_with_cache(child_obj_token, config)
+                    source_url = f"{base_url}/wiki/{child_node_token}"
+                    documents.append((child_title, content, source_url))
+                except ExternalServiceError:
+                    continue
 
         # 也读取节点本身的文档（如果有）
         if obj_type == "docx" and obj_token:
@@ -489,6 +537,21 @@ class FeishuKnowledgeService:
         cls._content_cache[cache_key] = (content, now)
         return content
 
+    @classmethod
+    def _get_file_content_with_cache(cls, file_token: str, config: FeishuRetrievalConfig) -> str:
+        """带缓存读取普通文件内容"""
+        cache_key = f"{config.app_id}:file:{file_token}"
+        now = time.time()
+
+        if cache_key in cls._content_cache:
+            content, ts = cls._content_cache[cache_key]
+            if now - ts < cls._CACHE_TTL:
+                return content
+
+        content = _get_file_content(file_token, config)
+        cls._content_cache[cache_key] = (content, now)
+        return content
+
     @classmethod
     def _get_wiki_children_with_cache(
         cls,