DS4SD · vagenas · Nov 4, 2024 · Nov 4, 2024 · Nov 4, 2024
diff --git a/.github/workflows/checks.yml b/.github/workflows/checks.yml
@@ -28,7 +28,7 @@ jobs:
         run: |
           for file in docs/examples/*.py; do
             # Skip batch_convert.py
-            if [[ "$(basename "$file")" == "batch_convert.py" ]]; then
+            if [[ "$(basename "$file")" =~ ^(batch_convert|minimal|export_multimodal|custom_convert|develop_picture_enrichment).py ]]; then
                 echo "Skipping $file"
                 continue
             fi

diff --git a/docs/examples/batch_convert.py b/docs/examples/batch_convert.py
@@ -106,8 +106,7 @@ def main():
         Path("./tests/data/2206.01062.pdf"),
         Path("./tests/data/2203.01017v2.pdf"),
         Path("./tests/data/2305.03393v1.pdf"),
-        Path("./tests/data/redp5110.pdf"),
-        Path("./tests/data/redp5695.pdf"),
+        Path("./tests/data/redp5110_sampled.pdf"),
     ]
 
     # buf = BytesIO(Path("./test/data/2206.01062.pdf").open("rb").read())

diff --git a/tests/data/groundtruth/docling_v1/redp5110.doctags.txt b/tests/data/groundtruth/docling_v1/redp5110.doctags.txt
diff --git a/tests/data/groundtruth/docling_v1/redp5110.json b/tests/data/groundtruth/docling_v1/redp5110.json
diff --git a/tests/data/groundtruth/docling_v1/redp5110.md b/tests/data/groundtruth/docling_v1/redp5110.md
diff --git a/tests/data/groundtruth/docling_v1/redp5110.pages.json b/tests/data/groundtruth/docling_v1/redp5110.pages.json
diff --git a/tests/data/groundtruth/docling_v1/redp5110_sampled.doctags.txt b/tests/data/groundtruth/docling_v1/redp5110_sampled.doctags.txt
diff --git a/tests/data/groundtruth/docling_v1/redp5110_sampled.json b/tests/data/groundtruth/docling_v1/redp5110_sampled.json
diff --git a/tests/data/groundtruth/docling_v1/redp5110_sampled.md b/tests/data/groundtruth/docling_v1/redp5110_sampled.md
diff --git a/tests/data/groundtruth/docling_v1/redp5110_sampled.pages.json b/tests/data/groundtruth/docling_v1/redp5110_sampled.pages.json
diff --git a/tests/data/groundtruth/docling_v1/redp5695.doctags.txt b/tests/data/groundtruth/docling_v1/redp5695.doctags.txt
diff --git a/tests/data/groundtruth/docling_v1/redp5695.json b/tests/data/groundtruth/docling_v1/redp5695.json
diff --git a/tests/data/groundtruth/docling_v1/redp5695.md b/tests/data/groundtruth/docling_v1/redp5695.md
diff --git a/tests/data/groundtruth/docling_v1/redp5695.pages.json b/tests/data/groundtruth/docling_v1/redp5695.pages.json
diff --git a/tests/data/groundtruth/docling_v2/redp5110.doctags.txt b/tests/data/groundtruth/docling_v2/redp5110.doctags.txt
diff --git a/tests/data/groundtruth/docling_v2/redp5110.json b/tests/data/groundtruth/docling_v2/redp5110.json
diff --git a/tests/data/groundtruth/docling_v2/redp5110.md b/tests/data/groundtruth/docling_v2/redp5110.md
diff --git a/tests/data/groundtruth/docling_v2/redp5110.pages.json b/tests/data/groundtruth/docling_v2/redp5110.pages.json
diff --git a/tests/data/groundtruth/docling_v2/redp5110_sampled.doctags.txt b/tests/data/groundtruth/docling_v2/redp5110_sampled.doctags.txt
diff --git a/tests/data/groundtruth/docling_v2/redp5110_sampled.json b/tests/data/groundtruth/docling_v2/redp5110_sampled.json
diff --git a/tests/data/groundtruth/docling_v2/redp5110_sampled.md b/tests/data/groundtruth/docling_v2/redp5110_sampled.md
diff --git a/tests/data/groundtruth/docling_v2/redp5110_sampled.pages.json b/tests/data/groundtruth/docling_v2/redp5110_sampled.pages.json
diff --git a/tests/data/groundtruth/docling_v2/redp5695.doctags.txt b/tests/data/groundtruth/docling_v2/redp5695.doctags.txt
diff --git a/tests/data/groundtruth/docling_v2/redp5695.json b/tests/data/groundtruth/docling_v2/redp5695.json
diff --git a/tests/data/groundtruth/docling_v2/redp5695.md b/tests/data/groundtruth/docling_v2/redp5695.md
diff --git a/tests/data/groundtruth/docling_v2/redp5695.pages.json b/tests/data/groundtruth/docling_v2/redp5695.pages.json
diff --git a/tests/data/redp5110.pdf b/tests/data/redp5110.pdf
diff --git a/tests/data/redp5110_sampled.pdf b/tests/data/redp5110_sampled.pdf
diff --git a/tests/data/redp5695.pdf b/tests/data/redp5695.pdf
diff --git a/tests/test_backend_docling_parse.py b/tests/test_backend_docling_parse.py
@@ -28,7 +28,7 @@ def _get_backend(pdf_doc):
 
 
 def test_text_cell_counts():
-    pdf_doc = Path("./tests/data/redp5695.pdf")
+    pdf_doc = Path("./tests/data/redp5110_sampled.pdf")
 
     doc_backend = _get_backend(pdf_doc)
 

diff --git a/tests/test_backend_docling_parse_v2.py b/tests/test_backend_docling_parse_v2.py
@@ -27,7 +27,7 @@ def _get_backend(pdf_doc):
 
 
 def test_text_cell_counts():
-    pdf_doc = Path("./tests/data/redp5695.pdf")
+    pdf_doc = Path("./tests/data/redp5110_sampled.pdf")
 
     doc_backend = _get_backend(pdf_doc)
 

diff --git a/tests/test_backend_pdfium.py b/tests/test_backend_pdfium.py
@@ -28,7 +28,7 @@ def _get_backend(pdf_doc):
 
 
 def test_text_cell_counts():
-    pdf_doc = Path("./tests/data/redp5695.pdf")
+    pdf_doc = Path("./tests/data/redp5110_sampled.pdf")
 
     doc_backend = _get_backend(pdf_doc)