Skip to content

Commit 2635cc7

Browse files
committed
add test files
1 parent 682cc3b commit 2635cc7

File tree

3 files changed

+26
-9
lines changed

3 files changed

+26
-9
lines changed
553 KB
Loading

_sample_docs/fake-power-point.ppt

594 KB
Binary file not shown.

_test_unstructured_client/integration/test_integration_freemium.py

Lines changed: 26 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -238,8 +238,15 @@ async def call_api():
238238
@pytest.mark.parametrize("split_pdf", [True, False])
239239
@pytest.mark.parametrize("vlm_model", ["gpt-4o"])
240240
@pytest.mark.parametrize("vlm_model_provider", ["openai"])
241-
def test_partition_strategy_vlm_openai(split_pdf, vlm_model, vlm_model_provider, client, doc_path):
242-
filename = "layout-parser-paper-fast.pdf"
241+
@pytest.mark.parametrize(
242+
"filename",
243+
[
244+
"layout-parser-paper-fast.pdf",
245+
"fake-power-point.ppt",
246+
"embedded-images-tables.jpg",
247+
]
248+
)
249+
def test_partition_strategy_vlm_openai(split_pdf, vlm_model, vlm_model_provider, client, doc_path, filename):
243250
with open(doc_path / filename, "rb") as f:
244251
files = shared.Files(
245252
content=f.read(),
@@ -279,8 +286,15 @@ def test_partition_strategy_vlm_openai(split_pdf, vlm_model, vlm_model_provider,
279286
]
280287
)
281288
@pytest.mark.parametrize("vlm_model_provider", ["bedrock"])
282-
def test_partition_strategy_vlm_bedrock(split_pdf, vlm_model, vlm_model_provider, client, doc_path):
283-
filename = "layout-parser-paper-fast.pdf"
289+
@pytest.mark.parametrize(
290+
"filename",
291+
[
292+
"layout-parser-paper-fast.pdf",
293+
"fake-power-point.ppt",
294+
"embedded-images-tables.jpg",
295+
]
296+
)
297+
def test_partition_strategy_vlm_bedrock(split_pdf, vlm_model, vlm_model_provider, client, doc_path, filename):
284298
with open(doc_path / filename, "rb") as f:
285299
files = shared.Files(
286300
content=f.read(),
@@ -307,14 +321,17 @@ def test_partition_strategy_vlm_bedrock(split_pdf, vlm_model, vlm_model_provider
307321

308322

309323
@pytest.mark.parametrize("split_pdf", [True, False])
310-
@pytest.mark.parametrize("vlm_model",
324+
@pytest.mark.parametrize("vlm_model", ["claude-3-5-sonnet-20241022",])
325+
@pytest.mark.parametrize("vlm_model_provider", ["anthropic"])
326+
@pytest.mark.parametrize(
327+
"filename",
311328
[
312-
"claude-3-5-sonnet-20241022",
329+
"layout-parser-paper-fast.pdf",
330+
"fake-power-point.ppt",
331+
"embedded-images-tables.jpg",
313332
]
314333
)
315-
@pytest.mark.parametrize("vlm_model_provider", ["anthropic"])
316-
def test_partition_strategy_vlm_anthropic(split_pdf, vlm_model, vlm_model_provider, client, doc_path):
317-
filename = "layout-parser-paper-fast.pdf"
334+
def test_partition_strategy_vlm_anthropic(split_pdf, vlm_model, vlm_model_provider, client, doc_path, filename):
318335
with open(doc_path / filename, "rb") as f:
319336
files = shared.Files(
320337
content=f.read(),

0 commit comments

Comments
 (0)