feat(kb): 截图内容视觉提取并入库(36张图→21条知识),import_kb 支持 --include-vision

kb_build/import_kb.py

@@ -21,6 +21,7 @@ nonebot.init()
 sys.path.insert(0, str(Path(__file__).parent.parent))
 
 PAGES_DIR = Path(__file__).parent / "pages"
+VISION_DIR = Path(__file__).parent / "vision_pages"
 
 # 每条知识的目标长度(字符),按标题切块后超过再按段落切
 CHUNK_SIZE = 600
@@ -86,10 +87,12 @@ def parse_page(path: Path) -> list[dict]:
     return chunks
 
 
-async def main(agent_id: str, dry_run: bool) -> None:
+async def main(agent_id: str, dry_run: bool, include_vision: bool = False) -> None:
     entries_spec: list[dict] = []
-    for f in sorted(PAGES_DIR.glob("*.txt")):
-        entries_spec.extend(parse_page(f))
+    dirs = [PAGES_DIR] + ([VISION_DIR] if include_vision else [])
+    for d in dirs:
+        for f in sorted(d.glob("*.txt")):
+            entries_spec.extend(parse_page(f))
     print(f"共解析出 {len(entries_spec)} 条知识条目")
     if dry_run:
         for e in entries_spec[:10]:
@@ -134,5 +137,6 @@ if __name__ == "__main__":
     parser = argparse.ArgumentParser()
     parser.add_argument("--agent-id", required=True, help="机器人 QQ 号(wtfllm 的 agent_id)")
     parser.add_argument("--dry-run", action="store_true", help="只解析不写入")
+    parser.add_argument("--include-vision", action="store_true", help="包含截图提取的知识")
     args = parser.parse_args()
-    asyncio.run(main(args.agent_id, args.dry_run))
+    asyncio.run(main(args.agent_id, args.dry_run, args.include_vision))