Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.
Python agentagentic-aiagentic-frameworkagentic-workflowaiautomationclaudedeepseekgenaigptllmlow-codemcpnextjsno-codeopenaipythonskillsworkflow
feat: support ODT document extraction (#39973)
蓝破碎半圆/破碎蓝调ccfb47d2c5fb
diff --git a/api/constants/__init__.py b/api/constants/__init__.pyindex 8698fb855de..095cfbb8610 100644--- a/api/constants/__init__.py+++ b/api/constants/__init__.py@@ -36,6 +36,7 @@ _UNSTRUCTURED_DOCUMENT_EXTENSION_BASE: frozenset[str] = frozenset("pptx","xml","epub",+ "odt",))_DEFAULT_DOCUMENT_EXTENSION_BASE: frozenset[str] = frozenset(@@ -53,6 +54,7 @@ _DEFAULT_DOCUMENT_EXTENSION_BASE: frozenset[str] = frozenset("csv","vtt","properties",+ "odt",))diff --git a/api/tests/unit_tests/core/workflow/nodes/test_document_extractor_node.py b/api/tests/unit_tests/core/workflow/nodes/test_document_extractor_node.pyindex d8d3c1ecc2e..ee094732b65 100644--- a/api/tests/unit_tests/core/workflow/nodes/test_document_extractor_node.py+++ b/api/tests/unit_tests/core/workflow/nodes/test_document_extractor_node.py@@ -441,6 +441,48 @@ def test_extract_text_from_excel_numeric_type_column(mock_excel_file):assert expected_manual == result+ ("extension", "mime_type", "route_label"),+ [+ (".odt", "text/plain", "extension"),+ (None, "application/vnd.oasis.opendocument.text", "mime_type"),+ ],+)+def test_extract_text_from_file_routes_odt_inputs_to_graphon_odt_extractor(+ document_extractor_node,+ extension,+ mime_type,+ route_label,+):+ file = Mock(spec=File)+ file.extension = extension+ file.mime_type = mime_type++ def fake_partition(file_content, *, suffix, unstructured_api_config, load_local_partition, render_element):+ assert file_content == b"odt content"+ assert suffix == ".odt"+ assert unstructured_api_config == document_extractor_node._unstructured_api_config+ assert load_local_partition.__name__ == "_load_partition_odt"+ assert render_element is not None+ return f"extracted through {route_label}"++ with (+ patch(+ "graphon.nodes.document_extractor.node._download_file_content",+ return_value=b"odt content",+ ) as mock_download,+ patch("graphon.nodes.document_extractor.node._partition_unstructured_file", side_effect=fake_partition),+ ):+ text = _extract_text_from_file(+ document_extractor_node.http_client,+ file,+ unstructured_api_config=document_extractor_node._unstructured_api_config,+ )++ assert text == f"extracted through {route_label}"+ mock_download.assert_called_once_with(document_extractor_node.http_client, file)++@pytest.mark.parametrize(("extension", "mime_type"),[diff --git a/api/tests/unit_tests/test_constants.py b/api/tests/unit_tests/test_constants.pynew file mode 100644index 00000000000..e40744a894a--- /dev/null+++ b/api/tests/unit_tests/test_constants.py@@ -0,0 +1,25 @@+import importlib++import pytest++import constants+from configs import dify_config++[email protected]("etl_type", ["SelfHosted", "Unstructured"])+def test_document_extensions_include_odt_for_document_etl_modes(monkeypatch: pytest.MonkeyPatch, etl_type: str) -> None:+ original_etl_type = dify_config.ETL_TYPE+ original_unstructured_api_url = dify_config.UNSTRUCTURED_API_URL++ try:+ monkeypatch.setattr(dify_config, "ETL_TYPE", etl_type)+ monkeypatch.setattr(dify_config, "UNSTRUCTURED_API_URL", None)++ reloaded_constants = importlib.reload(constants)++ assert "odt" in reloaded_constants.DOCUMENT_EXTENSIONS+ assert "ODT" in reloaded_constants.DOCUMENT_EXTENSIONS+ finally:+ monkeypatch.setattr(dify_config, "ETL_TYPE", original_etl_type)+ monkeypatch.setattr(dify_config, "UNSTRUCTURED_API_URL", original_unstructured_api_url)+ importlib.reload(constants)