feat: add Docling reader and node parser (#16406)

.github/workflows/coverage.yml

@@ -59,11 +59,11 @@ jobs:
             root="$file"
             # Keep going up the directory tree until we find a directory containing a marker file
             # (e.g., 'pyproject.toml' for python projects)
-            while [[ ! -f "$root/pyproject.toml" && "$root" != "." && "$root" != "/" ]]; do
+            while [[ ! -f "$root/pyproject.toml" && "$root" != "." && "$root" != "/" && "$root" != "./" ]]; do
               root=$(dirname "$root")
             done
 
-            if [[ ! "$FILTER_PATTERNS" =~ "$root" ]]; then
+            if [[ "$root" != "." && "$root" != "/" && "$root" != "./" && ! "$FILTER_PATTERNS" =~ "$root" ]]; then
               FILTER_PATTERNS="${FILTER_PATTERNS}'${root}',"
               CHANGED_ROOTS="${CHANGED_ROOTS} ${root}/::"
             fi

llama-index-integrations/node_parser/llama-index-node-parser-docling/.gitignore

@@ -0,0 +1,153 @@
+llama_index/_static
+.DS_Store
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+bin/
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+etc/
+include/
+lib/
+lib64/
+parts/
+sdist/
+share/
+var/
+wheels/
+pip-wheel-metadata/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+.ruff_cache
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+notebooks/
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+.python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+pyvenv.cfg
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# Jetbrains
+.idea
+modules/
+*.swp
+
+# VsCode
+.vscode
+
+# pipenv
+Pipfile
+Pipfile.lock
+
+# pyright
+pyrightconfig.json

llama-index-integrations/node_parser/llama-index-node-parser-docling/BUILD

@@ -0,0 +1,3 @@
+poetry_requirements(
+    name="poetry",
+)

llama-index-integrations/node_parser/llama-index-node-parser-docling/Makefile

@@ -0,0 +1,17 @@
+GIT_ROOT ?= $(shell git rev-parse --show-toplevel)
+
+help:	## Show all Makefile targets.
+	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | awk 'BEGIN {FS = ":.*?## "}; {printf "\033[33m%-30s\033[0m %s\n", $$1, $$2}'
+
+format:	## Run code autoformatters (black).
+	pre-commit install
+	git ls-files | xargs pre-commit run black --files
+
+lint:	## Run linters: pre-commit (black, ruff, codespell) and mypy
+	pre-commit install && git ls-files | xargs pre-commit run --show-diff-on-failure --files
+
+test:	## Run tests via pytest.
+	pytest tests
+
+watch-docs:	## Build and watch documentation.
+	sphinx-autobuild docs/ docs/_build/html --open-browser --watch $(GIT_ROOT)/llama_index/

llama-index-integrations/node_parser/llama-index-node-parser-docling/README.md

@@ -0,0 +1,35 @@
+# Docling Node Parser
+
+## Overview
+
+Docling Node Parser parses [Docling](https://github.com/DS4SD/docling) JSON output into LlamaIndex nodes with rich metadata for usage in downstream pipelines for RAG / QA etc.
+
+## Installation
+
+```console
+pip install llama-index-node-parser-docling
+```
+
+## Usage
+
+Docling Node Parser parses LlamaIndex documents containing JSON-serialized Docling format, as created by a Docling Reader.
+
+Basic usage looks like this:
+
+```python
+# docs = ...  # e.g. created using Docling Reader in JSON mode
+
+from llama_index.node_parser.docling import DoclingNodeParser
+
+node_parser = DoclingNodeParser()
+nodes = node_parser.get_nodes_from_documents(documents=docs)
+print(f"{nodes[6].text[:70]}...")
+# > Docling provides an easy code interface to convert PDF documents from ...
+
+print(nodes[6].metadata)
+# > {'dl_doc_hash': '556ad9e23b...',
+# >  'path': '#/main-text/22',
+# >  'heading': '2 Getting Started',
+# >  'page': 2,
+# >  'bbox': [107.40, 456.93, 504.20, 499.65]}
+```

llama-index-integrations/node_parser/llama-index-node-parser-docling/llama_index/node_parser/docling/BUILD

@@ -0,0 +1 @@
+python_sources()

llama-index-integrations/node_parser/llama-index-node-parser-docling/llama_index/node_parser/docling/__init__.py

@@ -0,0 +1,4 @@
+from llama_index.node_parser.docling.base import DoclingNodeParser
+
+
+__all__ = ["DoclingNodeParser"]

llama-index-integrations/node_parser/llama-index-node-parser-docling/llama_index/node_parser/docling/base.py

@@ -0,0 +1,81 @@
+from typing import Any, Iterable, Sequence
+
+from llama_index.core.schema import Document as LIDocument
+from llama_index.core.node_parser import NodeParser
+
+from docling_core.transforms.chunker import BaseChunker, HierarchicalChunker
+from docling_core.types import Document as DLDocument
+from llama_index.core import Document as LIDocument
+from llama_index.core.node_parser import NodeParser
+from llama_index.core.node_parser.node_utils import IdFuncCallable, default_id_func
+from llama_index.core.schema import (
+    BaseNode,
+    NodeRelationship,
+    RelatedNodeType,
+    TextNode,
+)
+from llama_index.core.utils import get_tqdm_iterable
+
+_NODE_TEXT_KEY = "text"
+
+
+class DoclingNodeParser(NodeParser):
+    """Docling format node parser.
+
+    Splits the JSON format of `DoclingReader` into nodes corresponding
+    to respective document elements from Docling's data model
+    (paragraphs, headings, tables etc.).
+
+    Args:
+        chunker (BaseChunker, optional): The chunker to use. Defaults to `HierarchicalChunker(heading_as_metadata=True)`.
+        doc_meta_keys_allowed (set[str], optional): The Document metadata keys allowed to be included for embedding and LLM input. Defaults to `set()`.
+        node_meta_keys_allowed (set[str], optional): The Node metadata keys allowed to be included for embedding and LLM input. Defaults to `{"heading"}`.
+    """
+
+    chunker: BaseChunker = HierarchicalChunker(heading_as_metadata=True)
+    doc_meta_keys_allowed: set[str] = set()
+    node_meta_keys_allowed: set[str] = {"heading"}
+
+    def _parse_nodes(
+        self,
+        nodes: Sequence[BaseNode],
+        show_progress: bool = False,
+        **kwargs: Any,
+    ) -> list[BaseNode]:
+        id_func: IdFuncCallable = self.id_func or default_id_func
+        nodes_with_progress: Iterable[BaseNode] = get_tqdm_iterable(
+            items=nodes, show_progress=show_progress, desc="Parsing nodes"
+        )
+        all_nodes: list[BaseNode] = []
+        for input_node in nodes_with_progress:
+            li_doc = LIDocument.model_validate(input_node)
+            dl_doc: DLDocument = DLDocument.model_validate_json(li_doc.get_content())
+            chunk_iter = self.chunker.chunk(dl_doc=dl_doc)
+            for i, chunk in enumerate(chunk_iter):
+                rels: dict[NodeRelationship, RelatedNodeType] = {
+                    NodeRelationship.SOURCE: li_doc.as_related_node_info(),
+                }
+                metadata = chunk.model_dump(
+                    exclude=_NODE_TEXT_KEY,
+                    exclude_none=True,
+                )
+                # by default we exclude all meta keys from embedding/LLM — unless allowed
+                excl_meta_keys = [
+                    k for k in metadata if k not in self.node_meta_keys_allowed
+                ]
+                if self.include_metadata:
+                    excl_meta_keys = [
+                        k
+                        for k in li_doc.metadata
+                        if k not in self.doc_meta_keys_allowed
+                    ] + excl_meta_keys
+                node = TextNode(
+                    id_=id_func(i=i, doc=li_doc),
+                    text=chunk.text,
+                    excluded_embed_metadata_keys=excl_meta_keys,
+                    excluded_llm_metadata_keys=excl_meta_keys,
+                    relationships=rels,
+                )
+                node.metadata = metadata
+                all_nodes.append(node)
+        return all_nodes

llama-index-integrations/node_parser/llama-index-node-parser-docling/pyproject.toml

@@ -0,0 +1,59 @@
+[build-system]
+build-backend = "poetry.core.masonry.api"
+requires = ["poetry-core"]
+
+[tool.codespell]
+check-filenames = true
+check-hidden = true
+# Feel free to un-skip examples, and experimental, you will just need to
+# work through many typos (--write-changes and --interactive will help)
+skip = "*.csv,*.html,*.json,*.jsonl,*.pdf,*.txt,*.ipynb"
+
+[tool.llamahub]
+contains_example = false
+import_path = "llama_index.node_parser.docling"
+
+[tool.llamahub.class_authors]
+DoclingNodeParser = "vagenas"
+
+[tool.mypy]
+disallow_untyped_defs = true
+# Remove venv skip when integrated with pre-commit
+exclude = ["_static", "build", "examples", "notebooks", "venv"]
+ignore_missing_imports = true
+python_version = "3.10"
+
+[tool.poetry]
+authors = ["Panos Vagenas <[email protected]>"]
+description = "llama-index node_parser docling integration"
+license = "MIT"
+name = "llama-index-node-parser-docling"
+packages = [{include = "llama_index/"}]
+readme = "README.md"
+version = "0.1.0"
+
+[tool.poetry.dependencies]
+python = "^3.10"
+llama-index-core = "^0.11.0"
+docling-core = "^1.7.1"
+
+[tool.poetry.group.dev]
+
+[tool.poetry.group.dev.dependencies]
+black = {extras = ["jupyter"], version = "<=23.9.1,>=23.7.0"}
+codespell = {extras = ["toml"], version = ">=v2.2.6"}
+ipython = "8.10.0"
+jupyter = "^1.0.0"
+mypy = "0.991"
+pre-commit = "3.2.0"
+pylint = "2.15.10"
+pytest = "7.2.1"
+pytest-mock = "3.11.1"
+ruff = "0.0.292"
+tree-sitter-languages = "^1.8.0"
+types-Deprecated = ">=0.1.0"
+types-PyYAML = "^6.0.12.12"
+types-protobuf = "^4.24.0.4"
+types-redis = "4.5.5.0"
+types-requests = "2.28.11.8"  # TODO: unpin when mypy>0.991
+types-setuptools = "67.1.0.0"

llama-index-integrations/node_parser/llama-index-node-parser-docling/tests/BUILD

@@ -0,0 +1,3 @@
+python_tests(
+    interpreter_constraints=["==3.10.*", "==3.11.*"]
+)

llama-index-integrations/node_parser/llama-index-node-parser-docling/tests/data/inp_li_doc.json

@@ -0,0 +1,18 @@
+{
+  "id_": "129210df929c78e70d74e6f141a46d8326905ce58562f2081819c80c3921d5a3",
+  "embedding": null,
+  "metadata": {
+    "dl_doc_hash": "129210df929c78e70d74e6f141a46d8326905ce58562f2081819c80c3921d5a3"
+  },
+  "excluded_embed_metadata_keys": ["dl_doc_hash"],
+  "excluded_llm_metadata_keys": ["dl_doc_hash"],
+  "relationships": {},
+  "text": "{\"_name\":\"\",\"type\":\"pdf-document\",\"description\":{\"title\":null,\"abstract\":null,\"authors\":null,\"affiliations\":null,\"subjects\":null,\"keywords\":null,\"publication_date\":null,\"languages\":null,\"license\":null,\"publishers\":null,\"url_refs\":null,\"references\":null,\"publication\":null,\"reference_count\":null,\"citation_count\":null,\"citation_date\":null,\"advanced\":null,\"analytics\":null,\"logs\":[],\"collection\":null,\"acquisition\":null},\"file-info\":{\"filename\":\"\",\"filename-prov\":null,\"document-hash\":\"129210df929c78e70d74e6f141a46d8326905ce58562f2081819c80c3921d5a3\",\"#-pages\":null,\"collection-name\":null,\"description\":null,\"page-hashes\":null},\"main-text\":[{\"text\":\"A duckling is a young duck in downy plumage[1] or baby duck,[2] but in the food trade a young domestic duck which has just reached adult size and bulk and its meat is still fully tender, is sometimes labelled as a duckling.\",\"type\":\"paragraph\",\"name\":\"text\",\"font\":null,\"prov\":[{\"bbox\":[1.0,2.0,3.0,4.0],\"page\":1,\"span\":[0,1],\"__ref_s3_data\":null}]},{\"text\":\"A male is called a drake and the female is called a duck, or in ornithology a hen.\",\"type\":\"paragraph\",\"name\":\"text\",\"font\":null,\"prov\":[{\"bbox\":[1.0,2.0,3.0,4.0],\"page\":1,\"span\":[0,2],\"__ref_s3_data\":null}]}],\"figures\":null,\"tables\":null,\"bitmaps\":null,\"equations\":null,\"footnotes\":null,\"page-dimensions\":null,\"page-footers\":null,\"page-headers\":null,\"_s3_data\":null,\"identifiers\":null}",
+  "mimetype": "text/plain",
+  "start_char_idx": null,
+  "end_char_idx": null,
+  "text_template": "{metadata_str}\n\n{content}",
+  "metadata_template": "{key}: {value}",
+  "metadata_seperator": "\n",
+  "class_name": "Document"
+}