Remove more unused code

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

Author: cau-git

docling_eval/benchmarks/dpbench/create.py

@@ -336,14 +336,6 @@ def create_dpbench_e2e_dataset(
     )
 
 
-def create_dpbench_layout_dataset(
-    dpbench_dir: Path, output_dir: Path, image_scale: float = 1.0
-):
-    create_dpbench_e2e_dataset(
-        dpbench_dir=dpbench_dir, output_dir=output_dir, image_scale=image_scale
-    )
-
-
 def create_dpbench_tableformer_dataset(
     dpbench_dir: Path, output_dir: Path, image_scale: float = 1.0
 ):
@@ -391,7 +383,7 @@ def create_dpbench_tableformer_dataset(
 
         # Create the updated Document
         updated, pred_doc = tf_updater.replace_tabledata(
-            pdf_path=pdf_path, true_doc=true_doc  # , true_page_images=true_page_images
+            pdf_path=pdf_path, true_doc=true_doc
         )
 
         if updated:

docling_eval/benchmarks/omnidocbench/create.py

@@ -360,16 +360,6 @@ def create_omnidocbench_e2e_dataset(
     )
 
 
-def create_omnidocbench_layout_dataset(
-    omnidocbench_dir: Path, output_dir: Path, image_scale: float = 1.0
-):
-    create_omnidocbench_e2e_dataset(
-        omnidocbench_dir=omnidocbench_dir,
-        output_dir=output_dir,
-        image_scale=image_scale,
-    )
-
-
 def create_omnidocbench_tableformer_dataset(
     omnidocbench_dir: Path, output_dir: Path, image_scale: float = 1.0
 ):
@@ -427,7 +417,7 @@ def create_omnidocbench_tableformer_dataset(
         )
 
         updated, pred_doc = tf_updater.replace_tabledata(
-            pdf_path=pdf_path, true_doc=true_doc, true_page_images=true_page_images
+            pdf_path=pdf_path, true_doc=true_doc
         )
 
         if updated:

docling_eval/benchmarks/utils.py

@@ -93,24 +93,25 @@ def add_pages_to_true_doc(
             page_width, page_height = page.size.width, page.size.height
 
             page_image = page.get_image(scale=image_scale)
-            page_images.append(page_image)
-            page._backend.unload()
+            if page_image is not None:
+                page_images.append(page_image)
+                image_ref = ImageRef(
+                    mimetype="image/png",
+                    dpi=round(72 * image_scale),
+                    size=Size(
+                        width=float(page_image.width), height=float(page_image.height)
+                    ),
+                    uri=Path(f"{BenchMarkColumns.PAGE_IMAGES}/{page_no}"),
+                )
+                page_item = PageItem(
+                    page_no=page_no + 1,
+                    size=Size(width=float(page_width), height=float(page_height)),
+                    image=image_ref,
+                )
 
-            image_ref = ImageRef(
-                mimetype="image/png",
-                dpi=round(72 * image_scale),
-                size=Size(
-                    width=float(page_image.width), height=float(page_image.height)
-                ),
-                uri=Path(f"{BenchMarkColumns.PAGE_IMAGES}/{page_no}"),
-            )
-            page_item = PageItem(
-                page_no=page_no + 1,
-                size=Size(width=float(page_width), height=float(page_height)),
-                image=image_ref,
-            )
+                true_doc.pages[page_no + 1] = page_item
 
-            true_doc.pages[page_no + 1] = page_item
+            page._backend.unload()
 
     return true_doc, page_images
 

docling_eval/cli/main.py

@@ -14,7 +14,7 @@
     create_dpbench_tableformer_dataset,
 )
 from docling_eval.benchmarks.omnidocbench.create import (
-    create_omnidocbench_layout_dataset,
+    create_omnidocbench_e2e_dataset,
     create_omnidocbench_tableformer_dataset,
 )
 from docling_eval.benchmarks.tableformer_huggingface_otsl.create import (
@@ -87,7 +87,7 @@ def create(
             modality == EvaluationModality.END2END
             or modality == EvaluationModality.LAYOUT
         ):
-            create_omnidocbench_layout_dataset(
+            create_omnidocbench_e2e_dataset(
                 omnidocbench_dir=idir, output_dir=odir, image_scale=image_scale
             )
         elif modality == EvaluationModality.TABLEFORMER:

docling_eval/docling/models/tableformer/tf_model_prediction.py

@@ -113,6 +113,7 @@ def to_np(pil_image: Image.Image):
     else:
         raise ValueError("Unsupported image format")
 
+
 # TODO: This method must be dropped.
 def tf_predict_with_page_tokens(
     config,
@@ -247,7 +248,6 @@ def replace_tabledata(
         self,
         pdf_path: Path,
         true_doc: DoclingDocument,
-        # true_page_images: List[Image.Image],
     ) -> Tuple[bool, DoclingDocument]:
 
         updated = False

docs/examples/benchmark_dpbench.py

@@ -6,7 +6,7 @@
 
 from docling_eval.benchmarks.constants import BenchMarkNames, EvaluationModality
 from docling_eval.benchmarks.dpbench.create import (
-    create_dpbench_layout_dataset,
+    create_dpbench_e2e_dataset,
     create_dpbench_tableformer_dataset,
 )
 from docling_eval.cli.main import evaluate, visualise
@@ -44,7 +44,7 @@ def main():
     image_scale = 1.0
 
     if True:
-        create_dpbench_layout_dataset(
+        create_dpbench_e2e_dataset(
             dpbench_dir=idir, output_dir=odir_lay, image_scale=image_scale
         )
 

docs/examples/benchmark_omnidocbench.py

@@ -6,7 +6,7 @@
 
 from docling_eval.benchmarks.constants import BenchMarkNames, EvaluationModality
 from docling_eval.benchmarks.omnidocbench.create import (
-    create_omnidocbench_layout_dataset,
+    create_omnidocbench_e2e_dataset,
     create_omnidocbench_tableformer_dataset,
 )
 from docling_eval.cli.main import evaluate, visualise
@@ -44,7 +44,8 @@ def main():
     image_scale = 1.0
 
     if True:
-        create_omnidocbench_layout_dataset(
+
+        create_omnidocbench_e2e_dataset(
             omnidocbench_dir=idir, output_dir=odir_lay, image_scale=image_scale
         )