yujunjun/unstructured
1
0
Fork 0
mirror of https://github.com/Unstructured-IO/unstructured.git synced 2025-09-27 09:25:59 +00:00
Code Issues Packages Projects Releases Wiki Activity

chore: fix tests breaking on main (#3603)

Browse Source 
Fix API tests (really more like integration tests) that run only on
main. Also use less compute intensive files to speedup test time and
remove a useless test.

Tests in `test_unstructured/partition/test_api.py` pass, temporarily
running outside of main per per screenshot:

![image](https://github.com/user-attachments/assets/f15d440a-2574-40f2-98b4-adf57fbae704)


https://github.com/Unstructured-IO/unstructured/actions/runs/10754098974/job/29824415513
This commit is contained in:
cragwolfe 2024-09-08 14:25:52 -07:00 committed by GitHub
parent c060467018
commit 3bb0ee1e79
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 35 additions and 32 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

67
test_unstructured/partition/test_api.py
View File

@ -18,6 +18,9 @@ from ..unit_utils import ANY, FixtureRequest, example_doc_path, method_mock
DIRECTORY = pathlib.Path(__file__).parent.resolve() DIRECTORY = pathlib.Path(__file__).parent.resolve()
# NOTE(crag): point to freemium API for now
API_URL = "https://api.unstructured.io/general/v0/general"
is_in_ci = os.getenv("CI", "").lower() not in {"", "false", "f", "0"} is_in_ci = os.getenv("CI", "").lower() not in {"", "false", "f", "0"}
skip_not_on_main = os.getenv("GITHUB_REF_NAME", "").lower() != "main" skip_not_on_main = os.getenv("GITHUB_REF_NAME", "").lower() != "main"
@ -105,20 +108,29 @@ def test_partition_via_api_raises_with_bad_response(request: FixtureRequest):
@pytest.mark.skipif(not is_in_ci, reason="Skipping test run outside of CI") @pytest.mark.skipif(not is_in_ci, reason="Skipping test run outside of CI")
@pytest.mark.skipif(skip_not_on_main, reason="Skipping test run outside of main branch") @pytest.mark.skipif(skip_not_on_main, reason="Skipping test run outside of main branch")
def test_partition_via_api_with_no_strategy(): def test_partition_via_api_with_no_strategy():
test_file = example_doc_path("pdf/loremipsum-flat.pdf")
elements_no_strategy = partition_via_api( elements_no_strategy = partition_via_api(
filename=example_doc_path("layout-parser-paper-fast.pdf"), filename=test_file,
strategy="auto", strategy="auto",
api_key=get_api_key(), api_key=get_api_key(),
# The url has changed since the 06/24 API release while the sdk defaults to the old url # The url has changed since the 06/24 API release while the sdk defaults to the old url
api_url="https://api.unstructuredapp.io/general/v0/general", api_url=API_URL,
skip_infer_table_types=["pdf"], skip_infer_table_types=["pdf"],
) )
elements_hi_res = partition_via_api( elements_hi_res = partition_via_api(
filename=example_doc_path("layout-parser-paper-fast.pdf"), filename=test_file,
strategy="hi_res", strategy="hi_res",
api_key=get_api_key(), api_key=get_api_key(),
# The url has changed since the 06/24 API release while the sdk defaults to the old url # The url has changed since the 06/24 API release while the sdk defaults to the old url
api_url="https://api.unstructuredapp.io/general/v0/general", api_url=API_URL,
skip_infer_table_types=["pdf"],
)
elements_fast_res = partition_via_api(
filename=test_file,
strategy="fast",
api_key=get_api_key(),
# The url has changed since the 06/24 API release while the sdk defaults to the old url
api_url=API_URL,
skip_infer_table_types=["pdf"], skip_infer_table_types=["pdf"],
) )
@ -126,7 +138,11 @@ def test_partition_via_api_with_no_strategy():
# elements_hi_res[3].text = # elements_hi_res[3].text =
# 'LayoutParser: A Unified Toolkit for Deep Learning Based Document Image Analysis' # 'LayoutParser: A Unified Toolkit for Deep Learning Based Document Image Analysis'
# while elements_no_strategy[3].text = ']' (as of this writing) # while elements_no_strategy[3].text = ']' (as of this writing)
assert elements_no_strategy[3].text != elements_hi_res[3].text assert len(elements_no_strategy) == len(elements_hi_res)
assert len(elements_hi_res) != len(elements_fast_res)
# NOTE(crag): slightly out scope assertion, but avoid extra API call
assert elements_hi_res[0].metadata.coordinates is None
@pytest.mark.skipif(not is_in_ci, reason="Skipping test run outside of CI") @pytest.mark.skipif(not is_in_ci, reason="Skipping test run outside of CI")
@ -134,41 +150,26 @@ def test_partition_via_api_with_no_strategy():
def test_partition_via_api_with_image_hi_res_strategy_includes_coordinates(): def test_partition_via_api_with_image_hi_res_strategy_includes_coordinates():
# coordinates not included by default to limit payload size # coordinates not included by default to limit payload size
elements = partition_via_api( elements = partition_via_api(
filename=example_doc_path("layout-parser-paper-fast.pdf"), filename=example_doc_path("pdf/fake-memo.pdf"),
strategy="hi_res", strategy="hi_res",
coordinates="true", coordinates="true",
api_key=get_api_key(), api_key=get_api_key(),
# The url has changed since the 06/24 API release while the sdk defaults to the old url api_url=API_URL,
api_url="https://api.unstructuredapp.io/general/v0/general",
) )
assert elements[0].metadata.coordinates is not None assert elements[0].metadata.coordinates is not None
@pytest.mark.skipif(not is_in_ci, reason="Skipping test run outside of CI")
@pytest.mark.skipif(skip_not_on_main, reason="Skipping test run outside of main branch")
def test_partition_via_api_valid_request_data_kwargs():
elements = partition_via_api(
filename=example_doc_path("layout-parser-paper-fast.pdf"),
strategy="fast",
api_key=get_api_key(),
# The url has changed since the 06/24 API release while the sdk defaults to the old url
api_url="https://api.unstructuredapp.io/general/v0/general",
)
assert isinstance(elements, list)
@pytest.mark.skipif(not is_in_ci, reason="Skipping test run outside of CI") @pytest.mark.skipif(not is_in_ci, reason="Skipping test run outside of CI")
@pytest.mark.skipif(skip_not_on_main, reason="Skipping test run outside of main branch") @pytest.mark.skipif(skip_not_on_main, reason="Skipping test run outside of main branch")
def test_partition_via_api_image_block_extraction(): def test_partition_via_api_image_block_extraction():
elements = partition_via_api( elements = partition_via_api(
filename=example_doc_path("embedded-images-tables.pdf"), filename=example_doc_path("pdf/embedded-images-tables.pdf"),
strategy="hi_res", strategy="hi_res",
extract_image_block_types=["image", "table"], extract_image_block_types=["image", "table"],
api_key=get_api_key(), api_key=get_api_key(),
# The url has changed since the 06/24 API release while the sdk defaults to the old url # The url has changed since the 06/24 API release while the sdk defaults to the old url
api_url="https://api.unstructuredapp.io/general/v0/general", api_url=API_URL,
) )
image_elements = [el for el in elements if el.category == ElementType.IMAGE] image_elements = [el for el in elements if el.category == ElementType.IMAGE]
for el in image_elements: for el in image_elements:
@ -357,18 +358,20 @@ def get_api_key():
@pytest.mark.skipif(skip_not_on_main, reason="Skipping test run outside of main branch") @pytest.mark.skipif(skip_not_on_main, reason="Skipping test run outside of main branch")
def test_partition_multiple_via_api_valid_request_data_kwargs(): def test_partition_multiple_via_api_valid_request_data_kwargs():
filenames = [ filenames = [
example_doc_path("layout-parser-paper-fast.pdf"), example_doc_path("fake-text.txt"),
example_doc_path("layout-parser-paper-fast.jpg"), example_doc_path("fake-email.txt"),
] ]
elements = partition_multiple_via_api( list_of_lists_of_elements = partition_multiple_via_api(
filenames=filenames, filenames=filenames,
strategy="auto", strategy="fast",
api_key=get_api_key(), api_key=get_api_key(),
# The url has changed since the 06/24 API release while the sdk defaults to the old url api_url=API_URL,
api_url="https://api.unstructuredapp.io/general/v0/general",
) )
assert isinstance(elements, list) # assert there is a list of elements for each file
assert len(list_of_lists_of_elements) == 2
assert isinstance(list_of_lists_of_elements[0], list)
assert isinstance(list_of_lists_of_elements[1], list)
@pytest.mark.skipif(not is_in_ci, reason="Skipping test run outside of CI") @pytest.mark.skipif(not is_in_ci, reason="Skipping test run outside of CI")
@ -383,7 +386,7 @@ def test_partition_multiple_via_api_invalid_request_data_kwargs():
strategy="not_a_strategy", strategy="not_a_strategy",
api_key=get_api_key(), api_key=get_api_key(),
# The url has changed since the 06/24 API release while the sdk defaults to the old url # The url has changed since the 06/24 API release while the sdk defaults to the old url
api_url="https://api.unstructuredapp.io/general/v0/general", api_url=API_URL,
) )