sync from accepted repo: init_knowledge.py

This commit is contained in:
2026-05-27 12:36:17 +00:00
parent 5f93cc403f
commit 808a916782
+21 -20
View File
@@ -4,28 +4,30 @@ from __future__ import annotations
import argparse import argparse
from pathlib import Path from pathlib import Path
try: from knowledge_base import KnowledgeBase
from knowledge_base import build_knowledge_base
except ModuleNotFoundError:
from .knowledge_base import build_knowledge_base
SUPPORTED_SUFFIXES = {".txt", ".md", ".markdown"} SUPPORTED_SUFFIXES = {".txt", ".md", ".markdown"}
def load_directory(directory: str | Path) -> int: def load_directory(knowledge_dir: Path, kb: KnowledgeBase | None = None) -> int:
root = Path(directory) kb = kb or KnowledgeBase()
if not root.exists(): if not knowledge_dir.exists():
raise FileNotFoundError(f"Директория не найдена: {root}") knowledge_dir.mkdir(parents=True, exist_ok=True)
print(f"Создана пустая директория: {knowledge_dir}")
return 0
kb = build_knowledge_base()
total_chunks = 0 total_chunks = 0
for path in sorted(root.rglob("*")): files = [p for p in knowledge_dir.rglob("*") if p.suffix.lower() in SUPPORTED_SUFFIXES]
if not path.is_file() or path.suffix.lower() not in SUPPORTED_SUFFIXES: if not files:
continue print(f"Нет файлов {SUPPORTED_SUFFIXES} в {knowledge_dir}")
content = path.read_text(encoding="utf-8") return 0
title = str(path.relative_to(root))
total_chunks += kb.add_document(content, title) for file_path in sorted(files):
print(f"Загружен: {title} ({len(content)} символов)") chunks = kb.add_file(file_path)
total_chunks += chunks
print(f" + {file_path.name}: {chunks} чанк(ов)")
print(f"Готово: {len(files)} файл(ов), {total_chunks} чанк(ов).")
return total_chunks return total_chunks
@@ -33,12 +35,11 @@ def main() -> None:
parser = argparse.ArgumentParser(description="Инициализация базы знаний из папки") parser = argparse.ArgumentParser(description="Инициализация базы знаний из папки")
parser.add_argument( parser.add_argument(
"--dir", "--dir",
default="./docs", default="./knowledge",
help="Путь к директории с .txt/.md файлами", help="Путь к директории с документами (.txt, .md)",
) )
args = parser.parse_args() args = parser.parse_args()
chunks = load_directory(args.dir) load_directory(Path(args.dir))
print(f"Готово. Всего чанков: {chunks}")
if __name__ == "__main__": if __name__ == "__main__":