change to sentence-level OCR

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

Author: cau-git

docling/datamodel/pipeline_options.py

@@ -351,7 +351,7 @@ class NemotronOcrOptions(OcrOptions):
                 "because it maps most directly to Docling OCR cells."
             )
         ),
-    ] = "word"
+    ] = "sentence"
     model_config = ConfigDict(
         extra="forbid",
     )

docs/examples/full_page_ocr.py

@@ -36,6 +36,7 @@
     TesseractCliOcrOptions,
 )
 from docling.document_converter import DocumentConverter, PdfFormatOption
+from docling.pipeline.legacy_standard_pdf_pipeline import LegacyStandardPdfPipeline
 
 
 def main():
@@ -53,17 +54,18 @@ def main():
     # TesseractCliOcrOptions, OcrMacOptions (macOS only), RapidOcrOptions,
     # NemotronOcrOptions (Linux x86_64, Python 3.12, CUDA 13.x only)
     # ocr_options = EasyOcrOptions(force_full_page_ocr=True)
-    # ocr_options = NemotronOcrOptions(force_full_page_ocr=True)
+    ocr_options = NemotronOcrOptions(force_full_page_ocr=True)
     # ocr_options = TesseractOcrOptions(force_full_page_ocr=True)
     # ocr_options = OcrMacOptions(force_full_page_ocr=True)
     # ocr_options = RapidOcrOptions(force_full_page_ocr=True)
-    ocr_options = TesseractCliOcrOptions(force_full_page_ocr=True)
+    # ocr_options = TesseractCliOcrOptions(force_full_page_ocr=True)
     pipeline_options.ocr_options = ocr_options
 
     converter = DocumentConverter(
         format_options={
             InputFormat.PDF: PdfFormatOption(
                 pipeline_options=pipeline_options,
+                # pipeline_cls=LegacyStandardPdfPipeline,
             )
         }
     )