docling-core 1.6.1__py3-none-any.whl → 1.6.2__py3-none-any.whl

docling_core/types/doc/document.py

@@ -477,6 +477,26 @@ class ExportedCCSDocument(
         md_texts: list[str] = []
 
         if self.main_text is not None:
+            # collect all captions embedded in table and figure objects
+            # to avoid repeating them
+            embedded_captions = set()
+            for orig_item in self.main_text[main_text_start:main_text_stop]:
+                item = (
+                    self._resolve_ref(orig_item)
+                    if isinstance(orig_item, Ref)
+                    else orig_item
+                )
+                if item is None:
+                    continue
+
+                if (
+                    isinstance(item, (Table, Figure))
+                    and item.text
+                    and item.obj_type in main_text_labels
+                ):
+                    embedded_captions.add(item.text)
+
+            # serialize document to markdown
             for orig_item in self.main_text[main_text_start:main_text_stop]:
                 markdown_text = ""
 
@@ -492,6 +512,11 @@ class ExportedCCSDocument(
                 if isinstance(item, BaseText) and item_type in main_text_labels:
                     text = item.text
 
+                    # skip captions of they are embedded in the actual
+                    # floating object
+                    if item_type == "caption" and text in embedded_captions:
+                        continue
+
                     # ignore repeated text
                     if prev_text == text or text is None:
                         continue
@@ -523,8 +548,9 @@ class ExportedCCSDocument(
                     isinstance(item, Table)
                     and item.data
                     and item_type in main_text_labels
-                    and not strict_text
                 ):
+
+                    md_table = ""
                     table = []
                     for row in item.data:
                         tmp = []
@@ -545,15 +571,19 @@ class ExportedCCSDocument(
                                 disable_numparse=True,
                             )
 
-                        markdown_text = md_table
+                    markdown_text = ""
+                    if item.text:
+                        markdown_text = item.text
+                    if not strict_text:
+                        markdown_text += "\n" + md_table
 
                 elif isinstance(item, Figure) and item_type in main_text_labels:
 
                     markdown_text = ""
-                    if not strict_text:
-                        markdown_text = f"{image_placeholder}"
                     if item.text:
-                        markdown_text += "\n" + item.text
+                        markdown_text = item.text
+                    if not strict_text:
+                        markdown_text += f"\n{image_placeholder}"
 
                 if markdown_text:
                     md_texts.append(markdown_text)

{docling_core-1.6.1.dist-info → docling_core-1.6.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: docling-core
-Version: 1.6.1
+Version: 1.6.2
 Summary: A python library to define and validate data types in Docling.
 Home-page: https://ds4sd.github.io/
 License: MIT

{docling_core-1.6.1.dist-info → docling_core-1.6.2.dist-info}/RECORD

@@ -24,7 +24,7 @@ docling_core/types/doc/base.py,sha256=QQC8KzQeYWnHFPY2_BNGcbTp6J2_rPbnLjsnbehICn
 docling_core/types/doc/doc_ann.py,sha256=8pV2efUglw19jxl4_oqB__mSxjWvtGIcllyCdqA-b2s,1196
 docling_core/types/doc/doc_ocr.py,sha256=6PC0C-OczF-MyfgRxEI1xs3PWgNOzi7i2yEQbTqZz0I,1387
 docling_core/types/doc/doc_raw.py,sha256=Y69G6IiauNDaoT-5el4xo1ypWpnBJQ75akGGkCMTZSc,3888
-docling_core/types/doc/document.py,sha256=qTVgmDX1geSRucJJeSV1G_sNjs6hL-kqns8rulbpyGw,22973
+docling_core/types/doc/document.py,sha256=AKp1kOo0tncf9FX3q7qRWQ2Jz_hZE44smZpyrtsRzY4,24104
 docling_core/types/doc/tokens.py,sha256=uU_MYW_p7ypf7eYICFBvxdnVaPZ7CQnvZmbJ6oPrtEA,6134
 docling_core/types/gen/__init__.py,sha256=C6TuCfvpSnSL5XDOFMcYHUY2-i08vvfOGRcdu6Af0pI,124
 docling_core/types/gen/generic.py,sha256=l4CZ4_Lb8ONG36WNJWbKX5hGKvTh_yU-hXp5hsm7uVU,844
@@ -45,8 +45,8 @@ docling_core/utils/ds_generate_jsonschema.py,sha256=EhNQutqWJFWuN-yl9UUPFZ7DJTvG
 docling_core/utils/file.py,sha256=VQgzjyvmJnAIHB6ex7ikcmbDAR4GA1ALreuO7Ubrp50,1895
 docling_core/utils/validate.py,sha256=3FmnxnKTDZC5J9OGxCL3U3DGRl0t0bBV1NcySXswdas,2031
 docling_core/utils/validators.py,sha256=fBdyWX4PvFh7o_d25ZTs4iwmeo75QTbrxsvXv2kXkTg,2777
-docling_core-1.6.1.dist-info/LICENSE,sha256=2M9-6EoQ1sxFztTOkXGAtwUDJvnWaAHdB9BYWVwGkIw,1087
-docling_core-1.6.1.dist-info/METADATA,sha256=JjVY-bxco9p4I2Rq-PY5LJePO5RBCVIIv-ZNK-F8EKA,5383
-docling_core-1.6.1.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-docling_core-1.6.1.dist-info/entry_points.txt,sha256=XHhtJEkdUuLxXSNxLdFIzx_siQ3z2UFQEKp-P8VYAE4,189
-docling_core-1.6.1.dist-info/RECORD,,
+docling_core-1.6.2.dist-info/LICENSE,sha256=2M9-6EoQ1sxFztTOkXGAtwUDJvnWaAHdB9BYWVwGkIw,1087
+docling_core-1.6.2.dist-info/METADATA,sha256=tYHPA7hs0aFYMr3W7FkkzsuwSySO5RbWrGN34neN2Mc,5383
+docling_core-1.6.2.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+docling_core-1.6.2.dist-info/entry_points.txt,sha256=XHhtJEkdUuLxXSNxLdFIzx_siQ3z2UFQEKp-P8VYAE4,189
+docling_core-1.6.2.dist-info/RECORD,,