| 123456789101112131415161718192021222324252627282930313233343536373839404142 |
- from pathlib import Path
- import sys
- sys.path.append(str(Path(__file__).resolve().parents[1]))
- from _chapter2_config import print_dependency_hint
- DATA_FILE = Path(__file__).with_name("data") / "衣服属性.txt"
- def main() -> None:
- try:
- from langchain_community.document_loaders import TextLoader
- except ModuleNotFoundError as exc:
- print_dependency_hint(exc)
- raise SystemExit(1)
- try:
- from langchain_unstructured import UnstructuredLoader
- except ModuleNotFoundError:
- UnstructuredLoader = None
- if UnstructuredLoader is not None:
- loader = UnstructuredLoader(str(DATA_FILE), encoding="utf8")
- docs = loader.load()
- print(f"unstructured docs-->{docs}")
- print(f"len-->{len(docs)}")
- print(f"第一行数据-->{docs[0].page_content}")
- print("*" * 100)
- else:
- print("未安装 langchain-unstructured,跳过 UnstructuredLoader 示例。")
- loader = TextLoader(str(DATA_FILE), encoding="utf8")
- docs = loader.load()
- print(f"TextLoader docs-->{docs}")
- print(f"len-->{len(docs)}")
- print("第一行数据-->{}".format(docs[0].page_content.split("\n")[0]))
- if __name__ == "__main__":
- main()
|