test_json.py 1.0 KB

12345678910111213141516171819202122232425262728293031
  1. import hashlib
  2. from unittest.mock import patch
  3. from langchain.docstore.document import Document
  4. from langchain.document_loaders.json_loader import JSONLoader as LcJSONLoader
  5. from embedchain.loaders.json import JSONLoader
  6. def test_load_data():
  7. mock_document = [
  8. Document(page_content="content1", metadata={"seq_num": 1}),
  9. Document(page_content="content2", metadata={"seq_num": 2}),
  10. ]
  11. with patch.object(LcJSONLoader, "load", return_value=mock_document):
  12. content = "temp.json"
  13. result = JsonLoader.load_data(content)
  14. assert "doc_id" in result
  15. assert "data" in result
  16. expected_data = [
  17. {"content": "content1", "meta_data": {"url": content, "row": 1}},
  18. {"content": "content2", "meta_data": {"url": content, "row": 2}},
  19. ]
  20. assert result["data"] == expected_data
  21. expected_doc_id = hashlib.sha256((content + ", ".join(["content1", "content2"])).encode()).hexdigest()
  22. assert result["doc_id"] == expected_doc_id