luciano/markitdown
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

update: use smaller test-pdf size

Browse source
This commit is contained in:
tungsten106 2024-12-26 17:16:38 +08:00
parent bd95fb06c7
commit 84824778f6
1 changed files with 10 additions and 7 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

17
tests/test_markitdown.py
View file

@ -307,7 +307,8 @@ def test_markitdown_pdf() -> None:
result = markitdown.convert( result = markitdown.convert(
os.path.join(TEST_FILES_DIR, "2308.08155v2.pdf"), os.path.join(TEST_FILES_DIR, "2308.08155v2.pdf"),
engine="pymupdf4llm", engine="pymupdf4llm",
engine_kwargs={"show_progress": False}, # additional kwargs
engine_kwargs={"show_progress": False, "pages": [i for i in range(10)],}, # additional kwargs
) )
for test_string in PDF_TEST_STRINGS: for test_string in PDF_TEST_STRINGS:
assert test_string in result.text_content assert test_string in result.text_content
@ -320,6 +321,7 @@ def test_markitdown_pdf() -> None:
"show_progress": False, "show_progress": False,
"write_images": True, "write_images": True,
"image_path": "tests/out", "image_path": "tests/out",
"pages": [i for i in range(10)],
}, # `write_images` must be True, setting `image_path` for images saving dir. }, # `write_images` must be True, setting `image_path` for images saving dir.
) )
for test_string in PDF_TEST_STRINGS: for test_string in PDF_TEST_STRINGS:
@ -327,16 +329,17 @@ def test_markitdown_pdf() -> None:
# By pdfminer # By pdfminer
result = markitdown.convert( result = markitdown.convert(
os.path.join(TEST_FILES_DIR, "2308.08155v2.pdf"), engine="pdfminer" os.path.join(TEST_FILES_DIR, "2308.08155v2.pdf"), engine="pdfminer",
enging_kwargs={"page_numbers": [i for i in range(10)],}
) )
for test_string in PDF_TEST_STRINGS: for test_string in PDF_TEST_STRINGS:
assert test_string in result.text_content assert test_string in result.text_content
if __name__ == "__main__": if __name__ == "__main__":
"""Runs this file's tests from the command line.""" """Runs this file's tests from the command line."""
test_markitdown_remote() # test_markitdown_remote()
test_markitdown_local() # test_markitdown_local()
test_markitdown_exiftool() # test_markitdown_exiftool()
test_markitdown_deprecation() # test_markitdown_deprecation()
test_markitdown_llm() # test_markitdown_llm()
test_markitdown_pdf() test_markitdown_pdf()