mirrors/docling
1
0
Fork 0
mirror of https://github.com/DS4SD/docling.git synced 2025-12-08 12:48:28 +00:00
Code Issues Packages Projects Releases Wiki Activity

ci: add coverage and ruff (#1383)

Browse Source 
* add coverage calculation and push

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* new codecov version and usage of token

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* enable ruff formatter instead of black and isort

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* apply ruff lint fixes

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* apply ruff unsafe fixes

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* add removed imports

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* runs 1 on linter issues

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* finalize linter fixes

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* Update pyproject.toml

Co-authored-by: Cesar Berrospi Ramis <75900930+ceberam@users.noreply.github.com>
Signed-off-by: Michele Dolfi <97102151+dolfim-ibm@users.noreply.github.com>

---------

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
Signed-off-by: Michele Dolfi <97102151+dolfim-ibm@users.noreply.github.com>
Co-authored-by: Cesar Berrospi Ramis <75900930+ceberam@users.noreply.github.com>
This commit is contained in:
Michele Dolfi
2025-04-14 18:01:26 +02:00
committed by GitHub
parent 293c28ca7c
commit 5458a88464
104 changed files with 665 additions and 633 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
tests/test_backend_asciidoc.py
View File

@@ -19,7 +19,6 @@ def _get_backend(fname):
def test_asciidocs_examples():
fnames = sorted(glob.glob("./tests/data/asciidoc/*.asciidoc"))
for fname in fnames:
@@ -38,8 +37,8 @@ def test_asciidocs_examples():
print("\n\n", pred_mddoc)
if os.path.exists(gname):
with open(gname, "r") as fr:
true_mddoc = fr.read()
with open(gname) as fr:
fr.read()
# assert pred_mddoc == true_mddoc, "pred_mddoc!=true_mddoc for asciidoc"
else:

13
tests/test_backend_csv.py
View File

@@ -1,5 +1,3 @@
import json
import os
from pathlib import Path
from pytest import warns
@@ -15,22 +13,19 @@ GENERATE = GEN_TEST_DATA
def get_csv_paths():
# Define the directory you want to search
directory = Path(f"./tests/data/csv/")
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
@@ -55,9 +50,9 @@ def test_e2e_valid_csv_conversions():
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_export(pred_itxt, str(gt_path) + ".itxt"), (
"export to indented-text"
)
assert verify_document(
pred_doc=doc,

10
tests/test_backend_docling_parse.py
View File

@@ -32,7 +32,7 @@ def test_text_cell_counts():
doc_backend = _get_backend(pdf_doc)
for page_index in range(0, doc_backend.page_count()):
for page_index in range(doc_backend.page_count()):
last_cell_count = None
for i in range(10):
page_backend: DoclingParsePageBackend = doc_backend.load_page(0)
@@ -42,9 +42,9 @@ def test_text_cell_counts():
last_cell_count = len(cells)
if len(cells) != last_cell_count:
assert (
False
), "Loading page multiple times yielded non-identical text cell counts"
assert False, (
"Loading page multiple times yielded non-identical text cell counts"
)
last_cell_count = len(cells)
@@ -66,7 +66,7 @@ def test_crop_page_image(test_doc_path):
page_backend: DoclingParsePageBackend = doc_backend.load_page(0)
# Crop out "Figure 1" from the DocLayNet paper
im = page_backend.get_page_image(
page_backend.get_page_image(
scale=2, cropbox=BoundingBox(l=317, t=246, r=574, b=527)
)
# im.show()

10
tests/test_backend_docling_parse_v2.py
View File

@@ -31,7 +31,7 @@ def test_text_cell_counts():
doc_backend = _get_backend(pdf_doc)
for page_index in range(0, doc_backend.page_count()):
for page_index in range(doc_backend.page_count()):
last_cell_count = None
for i in range(10):
page_backend: DoclingParseV2PageBackend = doc_backend.load_page(0)
@@ -41,9 +41,9 @@ def test_text_cell_counts():
last_cell_count = len(cells)
if len(cells) != last_cell_count:
assert (
False
), "Loading page multiple times yielded non-identical text cell counts"
assert False, (
"Loading page multiple times yielded non-identical text cell counts"
)
last_cell_count = len(cells)
@@ -65,7 +65,7 @@ def test_crop_page_image(test_doc_path):
page_backend: DoclingParseV2PageBackend = doc_backend.load_page(0)
# Crop out "Figure 1" from the DocLayNet paper
im = page_backend.get_page_image(
page_backend.get_page_image(
scale=2, cropbox=BoundingBox(l=317, t=246, r=574, b=527)
)
# im.show()

10
tests/test_backend_docling_parse_v4.py
View File

@@ -31,7 +31,7 @@ def test_text_cell_counts():
doc_backend = _get_backend(pdf_doc)
for page_index in range(0, doc_backend.page_count()):
for page_index in range(doc_backend.page_count()):
last_cell_count = None
for i in range(10):
page_backend: DoclingParseV4PageBackend = doc_backend.load_page(0)
@@ -41,9 +41,9 @@ def test_text_cell_counts():
last_cell_count = len(cells)
if len(cells) != last_cell_count:
assert (
False
), "Loading page multiple times yielded non-identical text cell counts"
assert False, (
"Loading page multiple times yielded non-identical text cell counts"
)
last_cell_count = len(cells)
@@ -65,7 +65,7 @@ def test_crop_page_image(test_doc_path):
page_backend: DoclingParseV4PageBackend = doc_backend.load_page(0)
# Crop out "Figure 1" from the DocLayNet paper
im = page_backend.get_page_image(
page_backend.get_page_image(
scale=2, cropbox=BoundingBox(l=317, t=246, r=574, b=527)
)
# im.show()

15
tests/test_backend_html.py
View File

@@ -105,7 +105,6 @@ def test_ordered_lists():
def get_html_paths():
# Define the directory you want to search
directory = Path("./tests/data/html/")
@@ -115,14 +114,12 @@ def get_html_paths():
def get_converter():
converter = DocumentConverter(allowed_formats=[InputFormat.HTML])
return converter
def test_e2e_html_conversions():
html_paths = get_html_paths()
converter = get_converter()
@@ -138,15 +135,15 @@ def test_e2e_html_conversions():
doc: DoclingDocument = conv_result.document
pred_md: str = doc.export_to_markdown()
assert verify_export(
pred_md, str(gt_path) + ".md", generate=GENERATE
), "export to md"
assert verify_export(pred_md, str(gt_path) + ".md", generate=GENERATE), (
"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", generate=GENERATE
), "export to indented-text"
assert verify_export(pred_itxt, str(gt_path) + ".itxt", generate=GENERATE), (
"export to indented-text"
)
assert verify_document(doc, str(gt_path) + ".json", GENERATE)

8
tests/test_backend_jats.py
View File

@@ -15,7 +15,7 @@ GENERATE = GEN_TEST_DATA
def get_pubmed_paths():
directory = Path(os.path.dirname(__file__) + f"/data/pubmed/")
directory = Path(os.path.dirname(__file__) + "/data/pubmed/")
xml_files = sorted(directory.rglob("*.xml"))
return xml_files
@@ -47,9 +47,9 @@ def test_e2e_pubmed_conversions(use_stream=False):
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_export(pred_itxt, str(gt_path) + ".itxt"), (
"export to indented-text"
)
assert verify_document(doc, str(gt_path) + ".json", GENERATE), "export to json"

18
tests/test_backend_msexcel.py
View File

@@ -17,7 +17,6 @@ GENERATE = GEN_TEST_DATA
def get_xlsx_paths():
# Define the directory you want to search
directory = Path("./tests/data/xlsx/")
@@ -27,7 +26,6 @@ def get_xlsx_paths():
def get_converter():
converter = DocumentConverter(allowed_formats=[InputFormat.XLSX])
return converter
@@ -65,13 +63,13 @@ def test_e2e_xlsx_conversions(documents) -> None:
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_export(pred_itxt, str(gt_path) + ".itxt"), (
"export to indented-text"
)
assert verify_document(
doc, str(gt_path) + ".json", GENERATE
), "document document"
assert verify_document(doc, str(gt_path) + ".json", GENERATE), (
"document document"
)
def test_pages(documents) -> None:
@@ -81,7 +79,7 @@ def test_pages(documents) -> None:
documents: The paths and converted documents.
"""
# number of pages from the backend method
path = [item for item in get_xlsx_paths() if item.stem == "test-01"][0]
path = next(item for item in get_xlsx_paths() if item.stem == "test-01")
in_doc = InputDocument(
path_or_stream=path,
format=InputFormat.XLSX,
@@ -92,7 +90,7 @@ def test_pages(documents) -> None:
assert backend.page_count() == 3
# number of pages from the converted document
doc = [item for path, item in documents if path.stem == "test-01"][0]
doc = next(item for path, item in documents if path.stem == "test-01")
assert len(doc.pages) == 3
# page sizes as number of cells

22
tests/test_backend_msword.py
View File

@@ -1,4 +1,3 @@
import os
from pathlib import Path
from docling.backend.msword_backend import MsWordDocumentBackend
@@ -43,7 +42,6 @@ def test_heading_levels():
def get_docx_paths():
# Define the directory you want to search
directory = Path("./tests/data/docx/")
@@ -53,14 +51,12 @@ def get_docx_paths():
def get_converter():
converter = DocumentConverter(allowed_formats=[InputFormat.DOCX])
return converter
def test_e2e_docx_conversions():
docx_paths = get_docx_paths()
converter = get_converter()
@@ -76,20 +72,20 @@ def test_e2e_docx_conversions():
doc: DoclingDocument = conv_result.document
pred_md: str = doc.export_to_markdown()
assert verify_export(
pred_md, str(gt_path) + ".md", generate=GENERATE
), "export to md"
assert verify_export(pred_md, str(gt_path) + ".md", generate=GENERATE), (
"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", generate=GENERATE
), "export to indented-text"
assert verify_export(pred_itxt, str(gt_path) + ".itxt", generate=GENERATE), (
"export to indented-text"
)
assert verify_document(
doc, str(gt_path) + ".json", generate=GENERATE
), "document document"
assert verify_document(doc, str(gt_path) + ".json", generate=GENERATE), (
"document document"
)
if docx_path.name == "word_tables.docx":
pred_html: str = doc.export_to_html()

38
tests/test_backend_patent_uspto.py
View File

@@ -109,27 +109,27 @@ def test_patent_groundtruth(patents, groundtruth):
md_name = path.stem + ".md"
if md_name in gt_names:
pred_md = doc.export_to_markdown()
assert (
pred_md == gt_names[md_name]
), f"Markdown file mismatch against groundtruth {md_name}"
assert pred_md == gt_names[md_name], (
f"Markdown file mismatch against groundtruth {md_name}"
)
json_path = path.with_suffix(".json")
if json_path.stem in gt_names:
assert verify_document(
doc, str(json_path), GENERATE
), f"JSON file mismatch against groundtruth {json_path}"
assert verify_document(doc, str(json_path), GENERATE), (
f"JSON file mismatch against groundtruth {json_path}"
)
itxt_name = path.stem + ".itxt"
if itxt_name in gt_names:
pred_itxt = doc._export_to_indented_text()
assert (
pred_itxt == gt_names[itxt_name]
), f"Indented text file mismatch against groundtruth {itxt_name}"
assert pred_itxt == gt_names[itxt_name], (
f"Indented text file mismatch against groundtruth {itxt_name}"
)
def test_tables(tables):
"""Test the table parser."""
# CHECK table in file tables_20180000016.xml
file_name = "tables_ipa20180000016.xml"
file_table = [item[1] for item in tables if item[0].name == file_name][0]
file_table = next(item[1] for item in tables if item[0].name == file_name)
assert file_table.num_rows == 13
assert file_table.num_cols == 10
assert len(file_table.table_cells) == 130
@@ -140,7 +140,7 @@ def test_patent_uspto_ice(patents):
# CHECK application doc number 20200022300
file_name = "ipa20200022300.xml"
doc = [item[1] for item in patents if item[0].name == file_name][0]
doc = next(item[1] for item in patents if item[0].name == file_name)
if GENERATE:
_generate_groundtruth(doc, Path(file_name).stem)
@@ -278,7 +278,7 @@ def test_patent_uspto_ice(patents):
# CHECK application doc number 20180000016 for HTML entities, level 2 headings, tables
file_name = "ipa20180000016.xml"
doc = [item[1] for item in patents if item[0].name == file_name][0]
doc = next(item[1] for item in patents if item[0].name == file_name)
if GENERATE:
_generate_groundtruth(doc, Path(file_name).stem)
@@ -348,7 +348,7 @@ def test_patent_uspto_ice(patents):
# CHECK application doc number 20110039701 for complex long tables
file_name = "ipa20110039701.xml"
doc = [item[1] for item in patents if item[0].name == file_name][0]
doc = next(item[1] for item in patents if item[0].name == file_name)
assert doc.name == file_name
assert len(doc.tables) == 17
@@ -358,7 +358,7 @@ def test_patent_uspto_grant_v2(patents):
# CHECK application doc number 06442728
file_name = "pg06442728.xml"
doc = [item[1] for item in patents if item[0].name == file_name][0]
doc = next(item[1] for item in patents if item[0].name == file_name)
if GENERATE:
_generate_groundtruth(doc, Path(file_name).stem)
@@ -376,12 +376,12 @@ def test_patent_uspto_grant_v2(patents):
assert isinstance(texts[2], TextItem)
assert texts[2].text == (
"An interleaver receives incoming data frames of size N. The interleaver "
"indexes the elements of the frame with an N₁×N₂ index array. The interleaver "
"indexes the elements of the frame with an N₁×N₂ index array. The interleaver " # noqa: RUF001
"then effectively rearranges (permutes) the data by permuting the rows of the "
"index array. The interleaver employs the equation I(j,k)=I(j,αjk+βj)modP) to "
"index array. The interleaver employs the equation I(j,k)=I(j,αjk+βj)modP) to " # noqa: RUF001
"permute the columns (indexed by k) of each row (indexed by j). P is at least "
"equal to N₂, βj is a constant which may be different for each row, and each "
"αj is a relative prime number relative to P. After permuting, the "
"αj is a relative prime number relative to P. After permuting, the " # noqa: RUF001
"interleaver outputs the data in a different order than received (e.g., "
"receives sequentially row by row, outputs sequentially each column by column)."
)
@@ -402,7 +402,7 @@ def test_patent_uspto_app_v1(patents):
# CHECK application doc number 20010031492
file_name = "pa20010031492.xml"
doc = [item[1] for item in patents if item[0].name == file_name][0]
doc = next(item[1] for item in patents if item[0].name == file_name)
if GENERATE:
_generate_groundtruth(doc, Path(file_name).stem)
@@ -432,7 +432,7 @@ def test_patent_uspto_grant_aps(patents):
# CHECK application doc number 057006474
file_name = "pftaps057006474.txt"
doc = [item[1] for item in patents if item[0].name == file_name][0]
doc = next(item[1] for item in patents if item[0].name == file_name)
if GENERATE:
_generate_groundtruth(doc, Path(file_name).stem)

10
tests/test_backend_pdfium.py
View File

@@ -32,7 +32,7 @@ def test_text_cell_counts():
doc_backend = _get_backend(pdf_doc)
for page_index in range(0, doc_backend.page_count()):
for page_index in range(doc_backend.page_count()):
last_cell_count = None
for i in range(10):
page_backend: PyPdfiumPageBackend = doc_backend.load_page(0)
@@ -42,9 +42,9 @@ def test_text_cell_counts():
last_cell_count = len(cells)
if len(cells) != last_cell_count:
assert (
False
), "Loading page multiple times yielded non-identical text cell counts"
assert False, (
"Loading page multiple times yielded non-identical text cell counts"
)
last_cell_count = len(cells)
@@ -66,7 +66,7 @@ def test_crop_page_image(test_doc_path):
page_backend: PyPdfiumPageBackend = doc_backend.load_page(0)
# Crop out "Figure 1" from the DocLayNet paper
im = page_backend.get_page_image(
page_backend.get_page_image(
scale=2, cropbox=BoundingBox(l=317, t=246, r=574, b=527)
)
# im.show()

16
tests/test_backend_pptx.py
View File

@@ -1,4 +1,3 @@
import os
from pathlib import Path
from docling.datamodel.base_models import InputFormat
@@ -12,7 +11,6 @@ GENERATE = GEN_TEST_DATA
def get_pptx_paths():
# Define the directory you want to search
directory = Path("./tests/data/pptx/")
@@ -22,14 +20,12 @@ def get_pptx_paths():
def get_converter():
converter = DocumentConverter(allowed_formats=[InputFormat.PPTX])
return converter
def test_e2e_pptx_conversions():
pptx_paths = get_pptx_paths()
converter = get_converter()
@@ -50,10 +46,10 @@ def test_e2e_pptx_conversions():
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_export(pred_itxt, str(gt_path) + ".itxt"), (
"export to indented-text"
)
assert verify_document(
doc, str(gt_path) + ".json", GENERATE
), "document document"
assert verify_document(doc, str(gt_path) + ".json", GENERATE), (
"document document"
)

2
tests/test_code_formula.py
View File

@@ -3,7 +3,6 @@ from pathlib import Path
from docling_core.types.doc import CodeItem, TextItem
from docling_core.types.doc.labels import CodeLanguageLabel, DocItemLabel
from docling.backend.docling_parse_backend import DoclingParseDocumentBackend
from docling.datamodel.base_models import InputFormat
from docling.datamodel.document import ConversionResult
from docling.datamodel.pipeline_options import PdfPipelineOptions
@@ -12,7 +11,6 @@ from docling.pipeline.standard_pdf_pipeline import StandardPdfPipeline
def get_converter():
pipeline_options = PdfPipelineOptions()
pipeline_options.generate_page_images = True

42
tests/test_document_picture_classifier.py
View File

@@ -2,7 +2,6 @@ from pathlib import Path
from docling_core.types.doc import PictureClassificationData
from docling.backend.docling_parse_backend import DoclingParseDocumentBackend
from docling.datamodel.base_models import InputFormat
from docling.datamodel.document import ConversionResult
from docling.datamodel.pipeline_options import PdfPipelineOptions
@@ -11,7 +10,6 @@ from docling.pipeline.standard_pdf_pipeline import StandardPdfPipeline
def get_converter():
pipeline_options = PdfPipelineOptions()
pipeline_options.generate_page_images = True
@@ -49,32 +47,32 @@ def test_picture_classifier():
res = results[0]
assert len(res.annotations) == 1
assert type(res.annotations[0]) == PictureClassificationData
assert isinstance(res.annotations[0], PictureClassificationData)
classification_data = res.annotations[0]
assert classification_data.provenance == "DocumentPictureClassifier"
assert (
len(classification_data.predicted_classes) == 16
), "Number of predicted classes is not equal to 16"
assert len(classification_data.predicted_classes) == 16, (
"Number of predicted classes is not equal to 16"
)
confidences = [pred.confidence for pred in classification_data.predicted_classes]
assert confidences == sorted(
confidences, reverse=True
), "Predictions are not sorted in descending order of confidence"
assert (
classification_data.predicted_classes[0].class_name == "bar_chart"
), "The prediction is wrong for the bar chart image."
assert confidences == sorted(confidences, reverse=True), (
"Predictions are not sorted in descending order of confidence"
)
assert classification_data.predicted_classes[0].class_name == "bar_chart", (
"The prediction is wrong for the bar chart image."
)
res = results[1]
assert len(res.annotations) == 1
assert type(res.annotations[0]) == PictureClassificationData
assert isinstance(res.annotations[0], PictureClassificationData)
classification_data = res.annotations[0]
assert classification_data.provenance == "DocumentPictureClassifier"
assert (
len(classification_data.predicted_classes) == 16
), "Number of predicted classes is not equal to 16"
assert len(classification_data.predicted_classes) == 16, (
"Number of predicted classes is not equal to 16"
)
confidences = [pred.confidence for pred in classification_data.predicted_classes]
assert confidences == sorted(
confidences, reverse=True
), "Predictions are not sorted in descending order of confidence"
assert (
classification_data.predicted_classes[0].class_name == "map"
), "The prediction is wrong for the bar chart image."
assert confidences == sorted(confidences, reverse=True), (
"Predictions are not sorted in descending order of confidence"
)
assert classification_data.predicted_classes[0].class_name == "map", (
"The prediction is wrong for the bar chart image."
)

4
tests/test_e2e_conversion.py
View File

@@ -1,7 +1,6 @@
from pathlib import Path
from docling.backend.docling_parse_backend import DoclingParseDocumentBackend
from docling.backend.docling_parse_v4_backend import DoclingParseV4DocumentBackend
from docling.datamodel.base_models import InputFormat
from docling.datamodel.document import ConversionResult
from docling.datamodel.pipeline_options import AcceleratorDevice, PdfPipelineOptions
@@ -15,7 +14,6 @@ GENERATE_V2 = GEN_TEST_DATA
def get_pdf_paths():
# Define the directory you want to search
directory = Path("./tests/data/pdf/")
@@ -25,7 +23,6 @@ def get_pdf_paths():
def get_converter():
pipeline_options = PdfPipelineOptions()
pipeline_options.do_ocr = False
pipeline_options.do_table_structure = True
@@ -45,7 +42,6 @@ def get_converter():
def test_e2e_pdfs_conversions():
pdf_paths = get_pdf_paths()
converter = get_converter()

1
tests/test_e2e_ocr_conversion.py
View File

@@ -3,7 +3,6 @@ from pathlib import Path
from typing import List
from docling.backend.docling_parse_backend import DoclingParseDocumentBackend
from docling.backend.docling_parse_v4_backend import DoclingParseV4DocumentBackend
from docling.datamodel.base_models import InputFormat
from docling.datamodel.document import ConversionResult
from docling.datamodel.pipeline_options import (

27
tests/test_input_doc.py
View File

@@ -12,10 +12,9 @@ from docling.document_converter import PdfFormatOption
def test_in_doc_from_valid_path():
test_doc_path = Path("./tests/data/pdf/2206.01062.pdf")
doc = _make_input_doc(test_doc_path)
assert doc.valid == True
assert doc.valid is True
def test_in_doc_from_invalid_path():
@@ -23,29 +22,26 @@ def test_in_doc_from_invalid_path():
doc = _make_input_doc(test_doc_path)
assert doc.valid == False
assert doc.valid is False
def test_in_doc_from_valid_buf():
buf = BytesIO(Path("./tests/data/pdf/2206.01062.pdf").open("rb").read())
stream = DocumentStream(name="my_doc.pdf", stream=buf)
doc = _make_input_doc_from_stream(stream)
assert doc.valid == True
assert doc.valid is True
def test_in_doc_from_invalid_buf():
buf = BytesIO(b"")
stream = DocumentStream(name="my_doc.pdf", stream=buf)
doc = _make_input_doc_from_stream(stream)
assert doc.valid == False
assert doc.valid is False
def test_image_in_pdf_backend():
in_doc = InputDocument(
path_or_stream=Path("tests/data/2305.03393v1-pg9-img.png"),
format=InputFormat.IMAGE,
@@ -76,7 +72,6 @@ def test_image_in_pdf_backend():
def test_in_doc_with_page_range():
test_doc_path = Path("./tests/data/pdf/2206.01062.pdf")
limits = DocumentLimits()
limits.page_range = (1, 10)
@@ -87,7 +82,7 @@ def test_in_doc_with_page_range():
backend=PyPdfiumDocumentBackend,
limits=limits,
)
assert doc.valid == True
assert doc.valid is True
limits.page_range = (9, 9)
@@ -97,7 +92,7 @@ def test_in_doc_with_page_range():
backend=PyPdfiumDocumentBackend,
limits=limits,
)
assert doc.valid == True
assert doc.valid is True
limits.page_range = (11, 12)
@@ -107,7 +102,7 @@ def test_in_doc_with_page_range():
backend=PyPdfiumDocumentBackend,
limits=limits,
)
assert doc.valid == False
assert doc.valid is False
def test_guess_format(tmp_path):
@@ -192,17 +187,17 @@ def test_guess_format(tmp_path):
)
doc_path = temp_dir / "docling_test.xml"
doc_path.write_text(xml_content, encoding="utf-8")
assert dci._guess_format(doc_path) == None
assert dci._guess_format(doc_path) is None
buf = BytesIO(Path(doc_path).open("rb").read())
stream = DocumentStream(name="docling_test.xml", stream=buf)
assert dci._guess_format(stream) == None
assert dci._guess_format(stream) is None
# Invalid USPTO patent (as plain text)
stream = DocumentStream(name="pftaps057006474.txt", stream=BytesIO(b"xyz"))
assert dci._guess_format(stream) == None
assert dci._guess_format(stream) is None
doc_path = temp_dir / "pftaps_wrong.txt"
doc_path.write_text("xyz", encoding="utf-8")
assert dci._guess_format(doc_path) == None
assert dci._guess_format(doc_path) is None
# Valid Docling JSON
test_str = '{"name": ""}'

5
tests/test_interfaces.py
View File

@@ -4,7 +4,6 @@ from pathlib import Path
import pytest
from docling.backend.docling_parse_backend import DoclingParseDocumentBackend
from docling.backend.docling_parse_v4_backend import DoclingParseV4DocumentBackend
from docling.datamodel.base_models import DocumentStream, InputFormat
from docling.datamodel.pipeline_options import PdfPipelineOptions
from docling.document_converter import DocumentConverter, PdfFormatOption
@@ -16,14 +15,12 @@ GENERATE = GEN_TEST_DATA
def get_pdf_path():
pdf_path = Path("./tests/data/pdf/2305.03393v1-pg9.pdf")
return pdf_path
@pytest.fixture
def converter():
pipeline_options = PdfPipelineOptions()
pipeline_options.do_ocr = False
pipeline_options.do_table_structure = True
@@ -42,7 +39,6 @@ def converter():
def test_convert_path(converter: DocumentConverter):
pdf_path = get_pdf_path()
print(f"converting {pdf_path}")
@@ -56,7 +52,6 @@ def test_convert_path(converter: DocumentConverter):
def test_convert_stream(converter: DocumentConverter):
pdf_path = get_pdf_path()
print(f"converting {pdf_path}")

1
tests/test_invalid_input.py
View File

@@ -8,7 +8,6 @@ from docling.document_converter import ConversionError, DocumentConverter
def get_pdf_path():
pdf_path = Path("./tests/data/pdf/2305.03393v1-pg9.pdf")
return pdf_path

3
tests/test_legacy_format_transform.py
View File

@@ -3,8 +3,6 @@ from pathlib import Path
import pytest
from docling.backend.docling_parse_backend import DoclingParseDocumentBackend
from docling.backend.docling_parse_v4_backend import DoclingParseV4DocumentBackend
from docling.datamodel.base_models import InputFormat
from docling.datamodel.pipeline_options import PdfPipelineOptions
from docling.document_converter import DocumentConverter, PdfFormatOption
@@ -23,7 +21,6 @@ def test_doc_paths():
def get_converter():
pipeline_options = PdfPipelineOptions()
pipeline_options.do_ocr = False

180
tests/verify_utils.py
View File

@@ -21,7 +21,6 @@ from docling.datamodel.document import ConversionResult
def levenshtein(str1: str, str2: str) -> int:
# Ensure str1 is the shorter string to optimize memory usage
if len(str1) > len(str2):
str1, str2 = str2, str1
@@ -46,7 +45,6 @@ def levenshtein(str1: str, str2: str) -> int:
def verify_text(gt: str, pred: str, fuzzy: bool, fuzzy_threshold: float = 0.4):
if len(gt) == 0 or not fuzzy:
assert gt == pred, f"{gt}!={pred}"
else:
@@ -57,22 +55,19 @@ def verify_text(gt: str, pred: str, fuzzy: bool, fuzzy_threshold: float = 0.4):
def verify_cells(doc_pred_pages: List[Page], doc_true_pages: List[Page]):
assert len(doc_pred_pages) == len(
doc_true_pages
), "pred- and true-doc do not have the same number of pages"
assert len(doc_pred_pages) == len(doc_true_pages), (
"pred- and true-doc do not have the same number of pages"
)
for pid, page_true_item in enumerate(doc_true_pages):
num_true_cells = len(page_true_item.cells)
num_pred_cells = len(doc_pred_pages[pid].cells)
assert (
num_true_cells == num_pred_cells
), f"num_true_cells!=num_pred_cells {num_true_cells}!={num_pred_cells}"
assert num_true_cells == num_pred_cells, (
f"num_true_cells!=num_pred_cells {num_true_cells}!={num_pred_cells}"
)
for cid, cell_true_item in enumerate(page_true_item.cells):
cell_pred_item = doc_pred_pages[pid].cells[cid]
true_text = cell_true_item.text
@@ -81,9 +76,9 @@ def verify_cells(doc_pred_pages: List[Page], doc_true_pages: List[Page]):
true_bbox = cell_true_item.rect.to_bounding_box().as_tuple()
pred_bbox = cell_pred_item.rect.to_bounding_box().as_tuple()
assert (
true_bbox == pred_bbox
), f"bbox is not the same: {true_bbox} != {pred_bbox}"
assert true_bbox == pred_bbox, (
f"bbox is not the same: {true_bbox} != {pred_bbox}"
)
return True
@@ -123,19 +118,19 @@ def verify_tables_v1(doc_pred: DsDocument, doc_true: DsDocument, fuzzy: bool):
# print("Expected number of tables: {}, result: {}".format(len(doc_true.tables), len(doc_pred.tables)))
assert len(doc_true.tables) == len(
doc_pred.tables
), "document has different count of tables than expected."
assert len(doc_true.tables) == len(doc_pred.tables), (
"document has different count of tables than expected."
)
for l, true_item in enumerate(doc_true.tables):
pred_item = doc_pred.tables[l]
for ix, true_item in enumerate(doc_true.tables):
pred_item = doc_pred.tables[ix]
assert (
true_item.num_rows == pred_item.num_rows
), "table does not have the same #-rows"
assert (
true_item.num_cols == pred_item.num_cols
), "table does not have the same #-cols"
assert true_item.num_rows == pred_item.num_rows, (
"table does not have the same #-rows"
)
assert true_item.num_cols == pred_item.num_cols, (
"table does not have the same #-cols"
)
assert true_item.data is not None, "documents are expected to have table data"
assert pred_item.data is not None, "documents are expected to have table data"
@@ -145,7 +140,6 @@ def verify_tables_v1(doc_pred: DsDocument, doc_true: DsDocument, fuzzy: bool):
for i, row in enumerate(true_item.data):
for j, col in enumerate(true_item.data[i]):
# print("true: ", true_item.data[i][j].text)
# print("pred: ", pred_item.data[i][j].text)
# print("")
@@ -154,20 +148,20 @@ def verify_tables_v1(doc_pred: DsDocument, doc_true: DsDocument, fuzzy: bool):
true_item.data[i][j].text, pred_item.data[i][j].text, fuzzy=fuzzy
)
assert (
true_item.data[i][j].obj_type == pred_item.data[i][j].obj_type
), "table-cell does not have the same type"
assert true_item.data[i][j].obj_type == pred_item.data[i][j].obj_type, (
"table-cell does not have the same type"
)
return True
def verify_table_v2(true_item: TableItem, pred_item: TableItem, fuzzy: bool):
assert (
true_item.data.num_rows == pred_item.data.num_rows
), "table does not have the same #-rows"
assert (
true_item.data.num_cols == pred_item.data.num_cols
), "table does not have the same #-cols"
assert true_item.data.num_rows == pred_item.data.num_rows, (
"table does not have the same #-rows"
)
assert true_item.data.num_cols == pred_item.data.num_cols, (
"table does not have the same #-cols"
)
assert true_item.data is not None, "documents are expected to have table data"
assert pred_item.data is not None, "documents are expected to have table data"
@@ -177,7 +171,6 @@ def verify_table_v2(true_item: TableItem, pred_item: TableItem, fuzzy: bool):
for i, row in enumerate(true_item.data.grid):
for j, col in enumerate(true_item.data.grid[i]):
# print("true: ", true_item.data[i][j].text)
# print("pred: ", pred_item.data[i][j].text)
# print("")
@@ -223,11 +216,11 @@ def verify_picture_image_v2(
def verify_docitems(doc_pred: DoclingDocument, doc_true: DoclingDocument, fuzzy: bool):
assert len(doc_pred.texts) == len(doc_true.texts), f"Text lengths do not match."
assert len(doc_pred.texts) == len(doc_true.texts), "Text lengths do not match."
assert len(doc_true.tables) == len(
doc_pred.tables
), "document has different count of tables than expected."
assert len(doc_true.tables) == len(doc_pred.tables), (
"document has different count of tables than expected."
)
for (true_item, _true_level), (pred_item, _pred_level) in zip(
doc_true.iterate_items(), doc_pred.iterate_items()
@@ -237,7 +230,7 @@ def verify_docitems(doc_pred: DoclingDocument, doc_true: DoclingDocument, fuzzy:
assert isinstance(pred_item, DocItem), "Test item is not a DocItem"
# Validate type
assert true_item.label == pred_item.label, f"Object label does not match."
assert true_item.label == pred_item.label, "Object label does not match."
# Validate provenance
assert len(true_item.prov) == len(pred_item.prov), "Length of prov mismatch"
@@ -261,25 +254,25 @@ def verify_docitems(doc_pred: DoclingDocument, doc_true: DoclingDocument, fuzzy:
# Validate table content
if isinstance(true_item, TableItem):
assert isinstance(
pred_item, TableItem
), "Test item is not a TableItem as the expected one"
assert verify_table_v2(
true_item, pred_item, fuzzy=fuzzy
), "Tables not matching"
assert isinstance(pred_item, TableItem), (
"Test item is not a TableItem as the expected one"
)
assert verify_table_v2(true_item, pred_item, fuzzy=fuzzy), (
"Tables not matching"
)
# Validate picture content
if isinstance(true_item, PictureItem):
assert isinstance(
pred_item, PictureItem
), "Test item is not a PictureItem as the expected one"
assert isinstance(pred_item, PictureItem), (
"Test item is not a PictureItem as the expected one"
)
true_image = true_item.get_image(doc=doc_true)
pred_image = true_item.get_image(doc=doc_pred)
if true_image is not None:
assert verify_picture_image_v2(
true_image, pred_image
), "Picture image mismatch"
assert verify_picture_image_v2(true_image, pred_image), (
"Picture image mismatch"
)
# TODO: check picture annotations
@@ -298,14 +291,14 @@ def verify_conversion_result_v1(
input_path: Path,
doc_result: ConversionResult,
generate: bool = False,
ocr_engine: str = None,
ocr_engine: Optional[str] = None,
fuzzy: bool = False,
):
PageList = TypeAdapter(List[Page])
assert (
doc_result.status == ConversionStatus.SUCCESS
), f"Doc {input_path} did not convert successfully."
assert doc_result.status == ConversionStatus.SUCCESS, (
f"Doc {input_path} did not convert successfully."
)
doc_pred_pages: List[Page] = doc_result.pages
doc_pred: DsDocument = doc_result.legacy_document
@@ -344,52 +337,52 @@ def verify_conversion_result_v1(
with open(dt_path, "w") as fw:
fw.write(doc_pred_dt)
else: # default branch in test
with open(pages_path, "r") as fr:
with open(pages_path) as fr:
doc_true_pages = PageList.validate_json(fr.read())
with open(json_path, "r") as fr:
with open(json_path) as fr:
doc_true: DsDocument = DsDocument.model_validate_json(fr.read())
with open(md_path, "r") as fr:
with open(md_path) as fr:
doc_true_md = fr.read()
with open(dt_path, "r") as fr:
with open(dt_path) as fr:
doc_true_dt = fr.read()
if not fuzzy:
assert verify_cells(
doc_pred_pages, doc_true_pages
), f"Mismatch in PDF cell prediction for {input_path}"
assert verify_cells(doc_pred_pages, doc_true_pages), (
f"Mismatch in PDF cell prediction for {input_path}"
)
# assert verify_output(
# doc_pred, doc_true
# ), f"Mismatch in JSON prediction for {input_path}"
assert verify_tables_v1(
doc_pred, doc_true, fuzzy=fuzzy
), f"verify_tables(doc_pred, doc_true) mismatch for {input_path}"
assert verify_tables_v1(doc_pred, doc_true, fuzzy=fuzzy), (
f"verify_tables(doc_pred, doc_true) mismatch for {input_path}"
)
assert verify_md(
doc_pred_md, doc_true_md, fuzzy=fuzzy
), f"Mismatch in Markdown prediction for {input_path}"
assert verify_md(doc_pred_md, doc_true_md, fuzzy=fuzzy), (
f"Mismatch in Markdown prediction for {input_path}"
)
assert verify_dt(
doc_pred_dt, doc_true_dt, fuzzy=fuzzy
), f"Mismatch in DocTags prediction for {input_path}"
assert verify_dt(doc_pred_dt, doc_true_dt, fuzzy=fuzzy), (
f"Mismatch in DocTags prediction for {input_path}"
)
def verify_conversion_result_v2(
input_path: Path,
doc_result: ConversionResult,
generate: bool = False,
ocr_engine: str = None,
ocr_engine: Optional[str] = None,
fuzzy: bool = False,
):
PageList = TypeAdapter(List[Page])
assert (
doc_result.status == ConversionStatus.SUCCESS
), f"Doc {input_path} did not convert successfully."
assert doc_result.status == ConversionStatus.SUCCESS, (
f"Doc {input_path} did not convert successfully."
)
doc_pred_pages: List[Page] = doc_result.pages
doc_pred: DoclingDocument = doc_result.document
@@ -426,42 +419,41 @@ def verify_conversion_result_v2(
with open(dt_path, "w") as fw:
fw.write(doc_pred_dt)
else: # default branch in test
with open(pages_path, "r") as fr:
with open(pages_path) as fr:
doc_true_pages = PageList.validate_json(fr.read())
with open(json_path, "r") as fr:
with open(json_path) as fr:
doc_true: DoclingDocument = DoclingDocument.model_validate_json(fr.read())
with open(md_path, "r") as fr:
with open(md_path) as fr:
doc_true_md = fr.read()
with open(dt_path, "r") as fr:
with open(dt_path) as fr:
doc_true_dt = fr.read()
if not fuzzy:
assert verify_cells(
doc_pred_pages, doc_true_pages
), f"Mismatch in PDF cell prediction for {input_path}"
assert verify_cells(doc_pred_pages, doc_true_pages), (
f"Mismatch in PDF cell prediction for {input_path}"
)
# assert verify_output(
# doc_pred, doc_true
# ), f"Mismatch in JSON prediction for {input_path}"
assert verify_docitems(
doc_pred, doc_true, fuzzy=fuzzy
), f"verify_docling_document(doc_pred, doc_true) mismatch for {input_path}"
assert verify_docitems(doc_pred, doc_true, fuzzy=fuzzy), (
f"verify_docling_document(doc_pred, doc_true) mismatch for {input_path}"
)
assert verify_md(
doc_pred_md, doc_true_md, fuzzy=fuzzy
), f"Mismatch in Markdown prediction for {input_path}"
assert verify_md(doc_pred_md, doc_true_md, fuzzy=fuzzy), (
f"Mismatch in Markdown prediction for {input_path}"
)
assert verify_dt(
doc_pred_dt, doc_true_dt, fuzzy=fuzzy
), f"Mismatch in DocTags prediction for {input_path}"
assert verify_dt(doc_pred_dt, doc_true_dt, fuzzy=fuzzy), (
f"Mismatch in DocTags prediction for {input_path}"
)
def verify_document(pred_doc: DoclingDocument, gtfile: str, generate: bool = False):
if not os.path.exists(gtfile) or generate:
with open(gtfile, "w") as fw:
json.dump(pred_doc.export_to_dict(), fw, indent=2)