mirror of
https://github.com/docling-project/docling.git
synced 2026-05-17 13:10:38 +00:00
Eugenio-BAYE
f2834848ae
* fix: handle empty CSV file without crashing Fixes #3195 Accessing self.csv_data[0] on an empty CSV raised an IndexError. Moved the column uniformity check inside the existing if/else on csv_data so empty files log a warning and return an empty document. Signed-off-by: Eugenio-BAYE <baye.eugenio.egnb@gmail.com> * Apply suggestions from code review Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Signed-off-by: Peter W. J. Staar <91719829+PeterStaar-IBM@users.noreply.github.com> --------- Signed-off-by: Eugenio-BAYE <baye.eugenio.egnb@gmail.com> Signed-off-by: Peter W. J. Staar <91719829+PeterStaar-IBM@users.noreply.github.com> Co-authored-by: Peter W. J. Staar <91719829+PeterStaar-IBM@users.noreply.github.com> Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
102 lines
3.3 KiB
Python
102 lines
3.3 KiB
Python
from io import BytesIO
|
|
from pathlib import Path
|
|
|
|
from pytest import warns
|
|
|
|
from docling.datamodel.base_models import DocumentStream, InputFormat
|
|
from docling.datamodel.document import ConversionResult, DoclingDocument
|
|
from docling.document_converter import DocumentConverter
|
|
|
|
from .test_data_gen_flag import GEN_TEST_DATA
|
|
from .verify_utils import verify_document, verify_export
|
|
|
|
GENERATE = GEN_TEST_DATA
|
|
|
|
|
|
def get_csv_paths():
|
|
# Define the directory you want to search
|
|
directory = Path("./tests/data/csv/")
|
|
|
|
# List all CSV files in the directory and its subdirectories
|
|
return sorted(directory.rglob("*.csv"))
|
|
|
|
|
|
def get_csv_path(name: str):
|
|
# Return the matching CSV file path
|
|
return Path(f"./tests/data/csv/{name}.csv")
|
|
|
|
|
|
def get_converter():
|
|
converter = DocumentConverter(allowed_formats=[InputFormat.CSV])
|
|
|
|
return converter
|
|
|
|
|
|
def test_e2e_valid_csv_conversions():
|
|
valid_csv_paths = get_csv_paths()
|
|
converter = get_converter()
|
|
|
|
for csv_path in valid_csv_paths:
|
|
print(f"converting {csv_path}")
|
|
|
|
gt_path = csv_path.parent.parent / "groundtruth" / "docling_v2" / csv_path.name
|
|
if csv_path.stem in (
|
|
"csv-too-few-columns",
|
|
"csv-too-many-columns",
|
|
"csv-inconsistent-header",
|
|
):
|
|
with warns(UserWarning, match="Inconsistent column lengths"):
|
|
conv_result: ConversionResult = converter.convert(csv_path)
|
|
else:
|
|
conv_result: ConversionResult = converter.convert(csv_path)
|
|
|
|
doc: DoclingDocument = conv_result.document
|
|
|
|
pred_md: str = doc.export_to_markdown()
|
|
assert verify_export(pred_md, str(gt_path) + ".md"), "export to md"
|
|
|
|
pred_itxt: str = doc._export_to_indented_text(
|
|
max_text_len=70, explicit_tables=False
|
|
)
|
|
assert verify_export(pred_itxt, str(gt_path) + ".itxt"), (
|
|
"export to indented-text"
|
|
)
|
|
|
|
assert verify_document(
|
|
pred_doc=doc,
|
|
gtfile=str(gt_path) + ".json",
|
|
generate=GENERATE,
|
|
), "export to json"
|
|
|
|
|
|
def test_e2e_invalid_csv_conversions():
|
|
csv_too_few_columns = get_csv_path("csv-too-few-columns")
|
|
csv_too_many_columns = get_csv_path("csv-too-many-columns")
|
|
csv_inconsistent_header = get_csv_path("csv-inconsistent-header")
|
|
converter = get_converter()
|
|
|
|
print(f"converting {csv_too_few_columns}")
|
|
with warns(UserWarning, match="Inconsistent column lengths"):
|
|
converter.convert(csv_too_few_columns)
|
|
|
|
print(f"converting {csv_too_many_columns}")
|
|
with warns(UserWarning, match="Inconsistent column lengths"):
|
|
converter.convert(csv_too_many_columns)
|
|
|
|
print(f"converting {csv_inconsistent_header}")
|
|
with warns(UserWarning, match="Inconsistent column lengths"):
|
|
converter.convert(csv_inconsistent_header)
|
|
|
|
|
|
def test_empty_csv():
|
|
"""Regression test: converting an empty CSV file should not raise an IndexError."""
|
|
conv_result = get_converter().convert(
|
|
DocumentStream(name="empty.csv", stream=BytesIO(b"")),
|
|
raises_on_error=True,
|
|
)
|
|
doc = conv_result.document
|
|
assert doc is not None
|
|
# The empty CSV should result in an empty document (no tables and no texts).
|
|
assert len(getattr(doc, "tables", [])) == 0
|
|
assert len(getattr(doc, "texts", [])) == 0
|