feat: add hierarchical chunker (#18)

Signed-off-by: Panos Vagenas <[email protected]>

docling_core/transforms/__init__.py

@@ -0,0 +1,6 @@
+#
+# Copyright IBM Corp. 2024 - 2024
+# SPDX-License-Identifier: MIT
+#
+
+"""Data transformations package."""

docling_core/transforms/chunker/__init__.py

@@ -0,0 +1,15 @@
+#
+# Copyright IBM Corp. 2024 - 2024
+# SPDX-License-Identifier: MIT
+#
+
+"""Define the chunker types."""
+
+from docling_core.transforms.chunker.base import (  # noqa
+    BaseChunker,
+    Chunk,
+    ChunkWithMetadata,
+)
+from docling_core.transforms.chunker.hierarchical_chunker import (  # noqa
+    HierarchicalChunker,
+)

docling_core/transforms/chunker/base.py

@@ -0,0 +1,45 @@
+#
+# Copyright IBM Corp. 2024 - 2024
+# SPDX-License-Identifier: MIT
+#
+
+"""Define base classes for chunking."""
+from abc import ABC, abstractmethod
+from typing import Iterator, Optional
+
+from pydantic import BaseModel
+
+from docling_core.types import BoundingBox, Document
+
+
+class Chunk(BaseModel):
+    """Data model for Chunk."""
+
+    path: str
+    text: str
+
+
+class ChunkWithMetadata(Chunk):
+    """Data model for Chunk including metadata."""
+
+    page: Optional[int]
+    bbox: Optional[BoundingBox]
+
+
+class BaseChunker(BaseModel, ABC):
+    """Base class for Chunker."""
+
+    @abstractmethod
+    def chunk(self, dl_doc: Document, **kwargs) -> Iterator[Chunk]:
+        """Chunk the provided document.
+
+        Args:
+            dl_doc (Document): document to chunk
+
+        Raises:
+            NotImplementedError: in this abstract implementation
+
+        Yields:
+            Iterator[Chunk]: iterator over extracted chunks
+        """
+        raise NotImplementedError()