docling 1.14.0__tar.gz → 1.16.0__tar.gz

{docling-1.14.0 → docling-1.16.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: docling
-Version: 1.14.0
+Version: 1.16.0
 Summary: Docling PDF conversion package
 Home-page: https://github.com/DS4SD/docling
 License: MIT
@@ -19,26 +19,18 @@ Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
-Provides-Extra: examples
 Requires-Dist: certifi (>=2024.7.4)
 Requires-Dist: deepsearch-glm (>=0.21.1,<0.22.0)
-Requires-Dist: docling-core (>=1.5.0,<2.0.0)
+Requires-Dist: docling-core (>=1.6.2,<2.0.0)
 Requires-Dist: docling-ibm-models (>=1.2.0,<2.0.0)
 Requires-Dist: docling-parse (>=1.2.0,<2.0.0)
 Requires-Dist: easyocr (>=1.7,<2.0)
 Requires-Dist: filetype (>=1.2.0,<2.0.0)
 Requires-Dist: huggingface_hub (>=0.23,<1)
-Requires-Dist: langchain-huggingface (>=0.0.3,<0.0.4) ; extra == "examples"
-Requires-Dist: langchain-milvus (>=0.1.4,<0.2.0) ; extra == "examples"
-Requires-Dist: langchain-text-splitters (>=0.2.4,<0.3.0) ; extra == "examples"
-Requires-Dist: llama-index-embeddings-huggingface (>=0.3.1,<0.4.0) ; extra == "examples"
-Requires-Dist: llama-index-llms-huggingface-api (>=0.2.0,<0.3.0) ; extra == "examples"
-Requires-Dist: llama-index-vector-stores-milvus (>=0.2.1,<0.3.0) ; extra == "examples"
 Requires-Dist: pyarrow (>=16.1.0,<17.0.0)
 Requires-Dist: pydantic (>=2.0.0,<3.0.0)
 Requires-Dist: pydantic-settings (>=2.3.0,<3.0.0)
 Requires-Dist: pypdfium2 (>=4.30.0,<5.0.0)
-Requires-Dist: python-dotenv (>=1.0.1,<2.0.0) ; extra == "examples"
 Requires-Dist: requests (>=2.32.3,<3.0.0)
 Requires-Dist: rtree (>=1.3.0,<2.0.0)
 Requires-Dist: scipy (>=1.14.1,<2.0.0)
@@ -211,6 +203,8 @@ This can improve output quality if you find that multiple columns in extracted t
 
 
 ```python
+from docling.datamodel.pipeline_options import PipelineOptions
+
 pipeline_options = PipelineOptions(do_table_structure=True)
 pipeline_options.table_structure_options.do_cell_matching = False  # uses text cells predicted from table structure model
 
@@ -220,6 +214,20 @@ doc_converter = DocumentConverter(
 )
 ```
 
+Since docling 1.16.0: You can control which TableFormer mode you want to use. Choose between `TableFormerMode.FAST` (default) and `TableFormerMode.ACCURATE` (better, but slower) to receive better quality with difficult table structures.
+
+```python
+from docling.datamodel.pipeline_options import PipelineOptions, TableFormerMode
+
+pipeline_options = PipelineOptions(do_table_structure=True)
+pipeline_options.table_structure_options.mode = TableFormerMode.ACCURATE  # use more accurate TableFormer model
+
+doc_converter = DocumentConverter(
+    artifacts_path=artifacts_path,
+    pipeline_options=pipeline_options,
+)
+```
+
 ### Impose limits on the document size
 
 You can limit the file size and number of pages which should be allowed to process per document:

{docling-1.14.0 → docling-1.16.0}/README.md

@@ -159,6 +159,8 @@ This can improve output quality if you find that multiple columns in extracted t
 
 
 ```python
+from docling.datamodel.pipeline_options import PipelineOptions
+
 pipeline_options = PipelineOptions(do_table_structure=True)
 pipeline_options.table_structure_options.do_cell_matching = False  # uses text cells predicted from table structure model
 
@@ -168,6 +170,20 @@ doc_converter = DocumentConverter(
 )
 ```
 
+Since docling 1.16.0: You can control which TableFormer mode you want to use. Choose between `TableFormerMode.FAST` (default) and `TableFormerMode.ACCURATE` (better, but slower) to receive better quality with difficult table structures.
+
+```python
+from docling.datamodel.pipeline_options import PipelineOptions, TableFormerMode
+
+pipeline_options = PipelineOptions(do_table_structure=True)
+pipeline_options.table_structure_options.mode = TableFormerMode.ACCURATE  # use more accurate TableFormer model
+
+doc_converter = DocumentConverter(
+    artifacts_path=artifacts_path,
+    pipeline_options=pipeline_options,
+)
+```
+
 ### Impose limits on the document size
 
 You can limit the file size and number of pages which should be allowed to process per document:

{docling-1.14.0 → docling-1.16.0}/docling/cli/main.py

@@ -12,8 +12,9 @@ from docling_core.utils.file import resolve_file_source
 
 from docling.backend.docling_parse_backend import DoclingParseDocumentBackend
 from docling.backend.pypdfium2_backend import PyPdfiumDocumentBackend
-from docling.datamodel.base_models import ConversionStatus, PipelineOptions
+from docling.datamodel.base_models import ConversionStatus
 from docling.datamodel.document import ConversionResult, DocumentConversionInput
+from docling.datamodel.pipeline_options import PipelineOptions
 from docling.document_converter import DocumentConverter
 
 warnings.filterwarnings(action="ignore", category=UserWarning, module="pydantic|torch")

{docling-1.14.0 → docling-1.16.0}/docling/datamodel/base_models.py

@@ -9,6 +9,10 @@ from pydantic import BaseModel, ConfigDict, Field, model_validator
 from typing_extensions import Self
 
 from docling.backend.abstract_backend import PdfPageBackend
+from docling.datamodel.pipeline_options import (  # Must be imported here for backward compatibility.
+    PipelineOptions,
+    TableStructureOptions,
+)
 
 
 class ConversionStatus(str, Enum):
@@ -298,22 +302,6 @@ class DocumentStream(BaseModel):
     stream: BytesIO
 
 
-class TableStructureOptions(BaseModel):
-    do_cell_matching: bool = (
-        True
-        # True:  Matches predictions back to PDF cells. Can break table output if PDF cells
-        #        are merged across table columns.
-        # False: Let table structure model define the text cells, ignore PDF cells.
-    )
-
-
-class PipelineOptions(BaseModel):
-    do_table_structure: bool = True  # True: perform table structure extraction
-    do_ocr: bool = True  # True: perform OCR, replace programmatic PDF text
-
-    table_structure_options: TableStructureOptions = TableStructureOptions()
-
-
 class AssembleOptions(BaseModel):
     keep_page_images: Annotated[
         bool,

{docling-1.14.0 → docling-1.16.0}/docling/datamodel/document.py

@@ -4,13 +4,13 @@ from pathlib import Path, PurePath
 from typing import ClassVar, Dict, Iterable, List, Optional, Tuple, Type, Union
 
 from docling_core.types import BaseCell, BaseText
-from docling_core.types import BoundingBox as DsBoundingBox
 from docling_core.types import Document as DsDocument
 from docling_core.types import DocumentDescription as DsDocumentDescription
 from docling_core.types import FileInfoObject as DsFileInfoObject
 from docling_core.types import PageDimensions, PageReference, Prov, Ref
 from docling_core.types import Table as DsSchemaTable
 from docling_core.types import TableCell
+from docling_core.types.doc.base import BoundingBox as DsBoundingBox
 from docling_core.types.doc.base import Figure
 from pydantic import BaseModel
 from typing_extensions import deprecated
@@ -324,8 +324,10 @@ class ConvertedDocument(BaseModel):
             "paragraph",
             "caption",
             "table",
+            "figure",
         ],
         strict_text: bool = False,
+        image_placeholder: str = "<!-- image -->",
     ):
         return self.output.export_to_markdown(
             delim=delim,
@@ -333,6 +335,7 @@ class ConvertedDocument(BaseModel):
             main_text_stop=main_text_stop,
             main_text_labels=main_text_labels,
             strict_text=strict_text,
+            image_placeholder=image_placeholder,
         )
 
     def render_as_text(

docling-1.16.0/docling/datamodel/pipeline_options.py

@@ -0,0 +1,25 @@
+from enum import Enum, auto
+
+from pydantic import BaseModel
+
+
+class TableFormerMode(str, Enum):
+    FAST = auto()
+    ACCURATE = auto()
+
+
+class TableStructureOptions(BaseModel):
+    do_cell_matching: bool = (
+        True
+        # True:  Matches predictions back to PDF cells. Can break table output if PDF cells
+        #        are merged across table columns.
+        # False: Let table structure model define the text cells, ignore PDF cells.
+    )
+    mode: TableFormerMode = TableFormerMode.FAST
+
+
+class PipelineOptions(BaseModel):
+    do_table_structure: bool = True  # True: perform table structure extraction
+    do_ocr: bool = True  # True: perform OCR, replace programmatic PDF text
+
+    table_structure_options: TableStructureOptions = TableStructureOptions()

{docling-1.14.0 → docling-1.16.0}/docling/document_converter.py

@@ -18,13 +18,13 @@ from docling.datamodel.base_models import (
     DoclingComponentType,
     ErrorItem,
     Page,
-    PipelineOptions,
 )
 from docling.datamodel.document import (
     ConversionResult,
     DocumentConversionInput,
     InputDocument,
 )
+from docling.datamodel.pipeline_options import PipelineOptions
 from docling.datamodel.settings import settings
 from docling.models.ds_glm_model import GlmModel
 from docling.models.page_assemble_model import PageAssembleModel

{docling-1.14.0 → docling-1.16.0}/docling/models/table_structure_model.py

@@ -1,4 +1,5 @@
 import copy
+from pathlib import Path
 from typing import Iterable, List
 
 import numpy
@@ -12,16 +13,22 @@ from docling.datamodel.base_models import (
     TableElement,
     TableStructurePrediction,
 )
+from docling.datamodel.pipeline_options import TableFormerMode
 
 
 class TableStructureModel:
     def __init__(self, config):
         self.config = config
         self.do_cell_matching = config["do_cell_matching"]
+        self.mode = config["mode"]
 
         self.enabled = config["enabled"]
         if self.enabled:
-            artifacts_path = config["artifacts_path"]
+            artifacts_path: Path = config["artifacts_path"]
+
+            if self.mode == TableFormerMode.ACCURATE:
+                artifacts_path = artifacts_path / "fat"
+
             # Third Party
             import docling_ibm_models.tableformer.common as c
 

{docling-1.14.0 → docling-1.16.0}/docling/pipeline/base_model_pipeline.py

@@ -1,7 +1,8 @@
 from pathlib import Path
 from typing import Callable, Iterable, List
 
-from docling.datamodel.base_models import Page, PipelineOptions
+from docling.datamodel.base_models import Page
+from docling.datamodel.pipeline_options import PipelineOptions
 
 
 class BaseModelPipeline:

{docling-1.14.0 → docling-1.16.0}/docling/pipeline/standard_model_pipeline.py

@@ -1,6 +1,6 @@
 from pathlib import Path
 
-from docling.datamodel.base_models import PipelineOptions
+from docling.datamodel.pipeline_options import PipelineOptions
 from docling.models.easyocr_model import EasyOcrModel
 from docling.models.layout_model import LayoutModel
 from docling.models.table_structure_model import TableStructureModel
@@ -32,6 +32,7 @@ class StandardModelPipeline(BaseModelPipeline):
                     "artifacts_path": artifacts_path
                     / StandardModelPipeline._table_model_path,
                     "enabled": pipeline_options.do_table_structure,
+                    "mode": pipeline_options.table_structure_options.mode,
                     "do_cell_matching": pipeline_options.table_structure_options.do_cell_matching,
                 }
             ),

{docling-1.14.0 → docling-1.16.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "docling"
-version = "1.14.0"  # DO NOT EDIT, updated automatically
+version = "1.16.0"  # DO NOT EDIT, updated automatically
 description = "Docling PDF conversion package"
 authors = ["Christoph Auer <cau@zurich.ibm.com>", "Michele Dolfi <dol@zurich.ibm.com>", "Maxim Lysak <mly@zurich.ibm.com>", "Nikos Livathinos <nli@zurich.ibm.com>", "Ahmed Nassar <ahn@zurich.ibm.com>", "Peter Staar <taa@zurich.ibm.com>"]
 license = "MIT"
@@ -21,9 +21,23 @@ keywords= ["docling", "convert", "document", "pdf", "layout model", "segmentatio
 packages = [{include = "docling"}]
 
 [tool.poetry.dependencies]
+##############
+# constraints:
+##############
+torch = [
+  {version = "^2.2.2", optional = true, markers = "sys_platform != 'darwin' or platform_machine != 'x86_64'"},
+  {version = "~2.2.2", optional = true, markers = "sys_platform == 'darwin' and platform_machine == 'x86_64'"}
+]
+torchvision = [
+  {version = "^0", optional = true, markers = "sys_platform != 'darwin' or platform_machine != 'x86_64'"},
+  {version = "~0.17.2", optional = true, markers = "sys_platform == 'darwin' and platform_machine == 'x86_64'"}
+]
+######################
+# actual dependencies:
+######################
 python = "^3.10"
 pydantic = "^2.0.0"
-docling-core = "^1.5.0"
+docling-core = "^1.6.2"
 docling-ibm-models = "^1.2.0"
 deepsearch-glm = "^0.21.1"
 filetype = "^1.2.0"
@@ -37,29 +51,6 @@ certifi = ">=2024.7.4"
 rtree = "^1.3.0"
 scipy = "^1.14.1"
 pyarrow = "^16.1.0"
-
-#########
-# extras:
-#########
-python-dotenv = { version = "^1.0.1", optional = true }
-llama-index-embeddings-huggingface = { version = "^0.3.1", optional = true }
-llama-index-llms-huggingface-api = { version = "^0.2.0", optional = true }
-llama-index-vector-stores-milvus = { version = "^0.2.1", optional = true }
-langchain-huggingface = { version = "^0.0.3", optional = true}
-langchain-milvus = { version = "^0.1.4", optional = true }
-langchain-text-splitters = { version = "^0.2.4", optional = true }
-
-##############
-# constraints:
-##############
-torch = [
-  {version = "^2.2.2", optional = true, markers = "sys_platform != 'darwin' or platform_machine != 'x86_64'"},
-  {version = "~2.2.2", optional = true, markers = "sys_platform == 'darwin' and platform_machine == 'x86_64'"}
-]
-torchvision = [
-  {version = "^0", optional = true, markers = "sys_platform != 'darwin' or platform_machine != 'x86_64'"},
-  {version = "~0.17.2", optional = true, markers = "sys_platform == 'darwin' and platform_machine == 'x86_64'"}
-]
 typer = "^0.12.5"
 
 [tool.poetry.group.dev.dependencies]
@@ -82,20 +73,13 @@ nbqa = "^1.9.0"
 
 [tool.poetry.group.examples.dependencies]
 datasets = "^2.21.0"
-
-[tool.poetry.extras]
-examples = [
-    "python-dotenv",
-    # LlamaIndex examples:
-    "llama-index-embeddings-huggingface",
-    "llama-index-llms-huggingface-api",
-    "llama-index-vector-stores-milvus",
-    # LangChain examples:
-    "langchain-huggingface",
-    "langchain-milvus",
-    "langchain-text-splitters",
-]
-
+python-dotenv = "^1.0.1"
+llama-index-embeddings-huggingface = "^0.3.1"
+llama-index-llms-huggingface-api = "^0.2.0"
+llama-index-vector-stores-milvus = "^0.2.1"
+langchain-huggingface = "^0.0.3"
+langchain-milvus = "^0.1.4"
+langchain-text-splitters = "^0.2.4"
 
 [tool.poetry.scripts]
 docling = "docling.cli.main:app"