docling-core 1.6.1__tar.gz → 1.6.3__tar.gz

{docling_core-1.6.1 → docling_core-1.6.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: docling-core
-Version: 1.6.1
+Version: 1.6.3
 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 → docling_core-1.6.3}/docling_core/search/json_schema_to_search_mapper.py

@@ -8,7 +8,7 @@ import re
 from copy import deepcopy
 from typing import Any, Optional, Pattern, Tuple, TypedDict
 
-from jsonref import JsonRef
+from jsonref import replace_refs
 
 
 class SearchIndexDefinition(TypedDict):
@@ -95,7 +95,11 @@ class JsonSchemaToSearchMapper:
         which define the fields, their data types, and other specifications to index
         JSON documents into a Lucene index.
         """
-        mapping = JsonRef.replace_refs(schema)
+        mapping = deepcopy(schema)
+
+        mapping = self._suppress(mapping, self._suppress_key)
+
+        mapping = replace_refs(mapping)
 
         mapping = self._merge_unions(mapping)
 
@@ -105,8 +109,6 @@ class JsonSchemaToSearchMapper:
 
         mapping = self._remove_keys(mapping, self._rm_keys)
 
-        mapping = self._suppress(mapping, self._suppress_key)
-
         mapping = self._translate_keys_re(mapping)
 
         mapping = self._clean(mapping)

{docling_core-1.6.1 → docling_core-1.6.3}/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 → docling_core-1.6.3}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "docling-core"
-version = "1.6.1"
+version = "1.6.3"
 description = "A python library to define and validate data types in Docling."
 license = "MIT"
 authors = [