chore: fix deprecation warnings (#303)

* chore: fix deprecation warnings

Signed-off-by: Cesar Berrospi Ramis <[email protected]>

* chore: disregard deprecated captions from hierarchical chunker in hybrid chunker

Signed-off-by: Cesar Berrospi Ramis <[email protected]>

* chore: update poetry lock file

Signed-off-by: Cesar Berrospi Ramis <[email protected]>

---------

Signed-off-by: Cesar Berrospi Ramis <[email protected]>

Author: ceberam

docling_core/transforms/chunker/hybrid_chunker.py

@@ -156,7 +156,6 @@ def _make_chunk_from_doc_items(
         meta = DocMeta(
             doc_items=doc_items,
             headings=doc_chunk.meta.headings,
-            captions=doc_chunk.meta.captions,
             origin=doc_chunk.meta.origin,
         )
         window_text = (
@@ -235,7 +234,9 @@ def _split_using_plain_text(
             )
             if available_length <= 0:
                 warnings.warn(
-                    f"Headers and captions for this chunk are longer than the total amount of size for the chunk, chunk will be ignored: {doc_chunk.text=}"  # noqa
+                    "Headers and captions for this chunk are longer than the total "
+                    "amount of size for the chunk, chunk will be ignored: "
+                    f"{doc_chunk.text=}"
                 )
                 return []
             text = doc_chunk.text
@@ -250,10 +251,10 @@ def _merge_chunks_with_matching_metadata(self, chunks: list[DocChunk]):
         num_chunks = len(chunks)
         while window_end < num_chunks:
             chunk = chunks[window_end]
-            headings_and_captions = (chunk.meta.headings, chunk.meta.captions)
+            headings = chunk.meta.headings
             ready_to_append = False
             if window_start == window_end:
-                current_headings_and_captions = headings_and_captions
+                current_headings = headings
                 window_end += 1
                 first_chunk_of_window = chunk
             else:
@@ -264,13 +265,12 @@ def _merge_chunks_with_matching_metadata(self, chunks: list[DocChunk]):
                     text=self.delim.join([chk.text for chk in chks]),
                     meta=DocMeta(
                         doc_items=doc_items,
-                        headings=current_headings_and_captions[0],
-                        captions=current_headings_and_captions[1],
+                        headings=current_headings,
                         origin=chunk.meta.origin,
                     ),
                 )
                 if (
-                    headings_and_captions == current_headings_and_captions
+                    headings == current_headings
                     and self._count_chunk_tokens(doc_chunk=candidate) <= self.max_tokens
                 ):
                     # there is room to include the new chunk so add it to the window and

docling_core/transforms/serializer/common.py

@@ -11,7 +11,7 @@
 from pathlib import Path
 from typing import Any, Iterable, Optional, Tuple, Union
 
-from pydantic import AnyUrl, BaseModel, NonNegativeInt, computed_field
+from pydantic import AnyUrl, BaseModel, ConfigDict, NonNegativeInt, computed_field
 from typing_extensions import Self, override
 
 from docling_core.transforms.serializer.base import (
@@ -176,11 +176,7 @@ def merge_with_patch(self, patch: dict[str, Any]) -> Self:
 class DocSerializer(BaseModel, BaseDocSerializer):
     """Class for document serializers."""
 
-    class Config:
-        """Pydantic config."""
-
-        arbitrary_types_allowed = True
-        extra = "forbid"
+    model_config = ConfigDict(arbitrary_types_allowed=True, extra="forbid")
 
     doc: DoclingDocument
 

docling_core/types/doc/document.py

@@ -11,6 +11,7 @@
 import re
 import sys
 import typing
+import warnings
 from enum import Enum
 from io import BytesIO
 from pathlib import Path
@@ -4109,7 +4110,10 @@ def check_version_is_compatible(cls, v: str) -> str:
     @classmethod
     def validate_document(cls, d: "DoclingDocument"):
         """validate_document."""
-        if not d.validate_tree(d.body) or not d.validate_tree(d.furniture):
-            raise ValueError("Document hierachy is inconsistent.")
+        with warnings.catch_warnings():
+            # ignore warning from deprecated furniture
+            warnings.filterwarnings("ignore", category=DeprecationWarning)
+            if not d.validate_tree(d.body) or not d.validate_tree(d.furniture):
+                raise ValueError("Document hierachy is inconsistent.")
 
         return d

docs/DoclingDocument.json

@@ -1475,6 +1475,7 @@
           "type": "string"
         },
         "content": {
+          "additionalProperties": true,
           "title": "Content",
           "type": "object"
         }