bugfix: fixing test_markitdown.py; updating exception messages
This commit is contained in:
parent
84824778f6
commit
f07ea3e09b
2 changed files with 10 additions and 10 deletions
|
|
@ -702,8 +702,8 @@ class PdfConverter(DocumentConverter):
|
||||||
return None
|
return None
|
||||||
if engine is not None and engine not in self._engines:
|
if engine is not None and engine not in self._engines:
|
||||||
raise FileConversionException(
|
raise FileConversionException(
|
||||||
"'pdf_engine' not valid. Please choose between {}.".format(
|
"'engine' not valid for {} files. Please choose between {}.".format(
|
||||||
list(self._engines.keys())
|
extension, list(self._engines.keys())
|
||||||
)
|
)
|
||||||
)
|
)
|
||||||
else:
|
else:
|
||||||
|
|
|
||||||
|
|
@ -308,7 +308,7 @@ def test_markitdown_pdf() -> None:
|
||||||
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, "pages": [i for i in range(10)],}, # additional kwargs
|
engine_kwargs={"show_progress": False, "pages": 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
|
||||||
|
|
@ -321,7 +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)],
|
"pages": 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:
|
||||||
|
|
@ -330,16 +330,16 @@ 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)],}
|
enging_kwargs={"page_numbers": 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()
|
||||||
|
|
|
||||||
Loading…
Reference in a new issue