[{"data":1,"prerenderedAt":4},["ShallowReactive",2],{"readme:docling-graph":3},"\u003Cp align=\"center\">\u003Cbr \u002F>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fdocling-project\u002Fdocling-graph\" rel=\"nofollow ugc noopener\">\n    \u003Cimg loading=\"lazy\" alt=\"Docling Graph\" src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fdocling-project\u002Fdocling-graph\u002FHEAD\u002Fdocs\u002Fassets\u002Flogo.png\" width=\"280\" \u002F>\n  \u003C\u002Fa>\n\u003C\u002Fp>\u003Ch1>Docling Graph\u003C\u002Fh1>\n\u003Cp>\u003Ca href=\"https:\u002F\u002Fdocling-project.github.io\u002Fdocling-graph\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fdocs-live-brightgreen\" alt=\"Docs\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpypi.org\u002Fproject\u002Fdocling-graph\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fpypi\u002Fv\u002Fdocling-graph?cacheSeconds=300\" alt=\"PyPI version\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fwww.python.org\u002Fdownloads\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPython-3.10%20%7C%203.11%20%7C%203.12-blue\" alt=\"Python 3.10 | 3.11 | 3.12\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fuv\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fendpoint?url=https:\u002F\u002Fraw.githubusercontent.com\u002Fastral-sh\u002Fuv\u002Fmain\u002Fassets\u002Fbadge\u002Fv0.json\" alt=\"uv\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fastral-sh\u002Fruff\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fendpoint?url=https:\u002F\u002Fraw.githubusercontent.com\u002Fastral-sh\u002Fruff\u002Fmain\u002Fassets\u002Fbadge\u002Fv2.json\" alt=\"Ruff\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fopensource.org\u002Flicenses\u002FMIT\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Fdocling-project\u002Fdocling-graph\" alt=\"License MIT\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fpydantic.dev\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fendpoint?url=https:\u002F\u002Fraw.githubusercontent.com\u002Fpydantic\u002Fpydantic\u002Fmain\u002Fdocs\u002Fbadge\u002Fv2.json\" alt=\"Pydantic v2\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fdocling-project\u002Fdocling\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDocling-VLM-red\" alt=\"Docling\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fnetworkx.org\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FNetworkX-3.0+-red\" alt=\"NetworkX\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Ftyper.tiangolo.com\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FTyper-CLI-purple\" alt=\"Typer\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002FTextualize\u002Frich\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FRich-terminal-purple\" alt=\"Rich\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fvllm.ai\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FvLLM-compatible-brightgreen\" alt=\"vLLM\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Follama.ai\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FOllama-compatible-brightgreen\" alt=\"Ollama\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Fwww.bestpractices.dev\u002Fprojects\u002F11598\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fwww.bestpractices.dev\u002Fprojects\u002F11598\u002Fbadge\" alt=\"OpenSSF Best Practices\" \u002F>\u003C\u002Fa>\n\u003Ca href=\"https:\u002F\u002Flfaidata.foundation\u002Fprojects\u002F\" rel=\"nofollow ugc noopener\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLF%20AI%20%26%20Data-003778?logo=linuxfoundation&logoColor=fff&color=0094ff&labelColor=003778\" alt=\"LF AI & Data\" \u002F>\u003C\u002Fa>\u003C\u002Fp>\n\u003Cp>Docling-Graph turns documents into validated \u003Cstrong>Pydantic\u003C\u002Fstrong> objects, then builds a \u003Cstrong>directed knowledge graph\u003C\u002Fstrong> with explicit semantic relationships.\u003C\u002Fp>\n\u003Cp>This transformation enables high-precision use cases in \u003Cstrong>chemistry, finance, and legal\u003C\u002Fstrong> domains, where AI must capture exact entity connections (compounds and reactions, instruments and dependencies, properties and measurements) \u003Cstrong>rather than rely on approximate text embeddings\u003C\u002Fstrong>.\u003C\u002Fp>\n\u003Cp>This toolkit supports two extraction paths: \u003Cstrong>local VLM extraction\u003C\u002Fstrong> via Docling, and \u003Cstrong>LLM-based extraction\u003C\u002Fstrong> routed through \u003Cstrong>LiteLLM\u003C\u002Fstrong> for local runtimes (vLLM, Ollama) and API providers (OpenAI, Gemini, IBM watsonx, Mistral and more), all orchestrated through a flexible, config-driven pipeline.\u003C\u002Fp>\n\u003Ch2>Key Capabilities\u003C\u002Fh2>\n\u003Cul>\n\u003Cli>\u003Cp>\u003Cstrong>✍🏻 Input formats:\u003C\u002Fstrong> \u003Ca href=\"https:\u002F\u002Fdocling-project.github.io\u002Fdocling\u002Fusage\u002Fsupported_formats\u002F\" rel=\"nofollow ugc noopener\">Docling\u003C\u002Fa>’s supported inputs: PDF, images, DocLang, markdown, Office and more.\u003C\u002Fp>\n\u003C\u002Fli>\n\u003Cli>\u003Cp>\u003Cstrong>🧠 Extraction:\u003C\u002Fstrong> \u003Ca href=\"https:\u002F\u002Fdocling-project.github.io\u002Fdocling-graph\u002Ffundamentals\u002Fpipeline-configuration\u002Fbackend-selection\u002F\" rel=\"nofollow ugc noopener\">LLM\u003C\u002Fa> or \u003Ca href=\"https:\u002F\u002Fdocling-project.github.io\u002Fdocling-graph\u002Ffundamentals\u002Fpipeline-configuration\u002Fbackend-selection\u002F\" rel=\"nofollow ugc noopener\">VLM\u003C\u002Fa> backends, with \u003Ca href=\"https:\u002F\u002Fdocling-project.github.io\u002Fdocling-graph\u002Ffundamentals\u002Fextraction-process\u002Fchunking-strategies\u002F\" rel=\"nofollow ugc noopener\">chunking\u003C\u002Fa> and \u003Ca href=\"https:\u002F\u002Fdocling-project.github.io\u002Fdocling-graph\u002Ffundamentals\u002Fpipeline-configuration\u002Fprocessing-modes\u002F\" rel=\"nofollow ugc noopener\">processing modes\u003C\u002Fa>.\u003C\u002Fp>\n\u003C\u002Fli>\n\u003Cli>\u003Cp>\u003Cstrong>💎 Graphs:\u003C\u002Fstrong> Pydantic to \u003Ca href=\"https:\u002F\u002Fdocling-project.github.io\u002Fdocling-graph\u002Ffundamentals\u002Fgraph-management\u002Fgraph-conversion\u002F\" rel=\"nofollow ugc noopener\">NetworkX\u003C\u002Fa> directed graphs with stable IDs, edge and \u003Ca href=\"https:\u002F\u002Fdocling-project.github.io\u002Fdocling-graph\u002Ffundamentals\u002Fgraph-management\u002Fprovenance\u002F\" rel=\"nofollow ugc noopener\">provenance\u003C\u002Fa> metadata.\u003C\u002Fp>\n\u003C\u002Fli>\n\u003Cli>\u003Cp>\u003Cstrong>📦 Export:\u003C\u002Fstrong> \u003Ca href=\"https:\u002F\u002Fdocling-project.github.io\u002Fdocling-graph\u002Ffundamentals\u002Fgraph-management\u002Fexport-formats\u002F#csv-export\" rel=\"nofollow ugc noopener\">CSV\u003C\u002Fa>, \u003Ca href=\"https:\u002F\u002Fdocling-project.github.io\u002Fdocling-graph\u002Ffundamentals\u002Fgraph-management\u002Fexport-formats\u002F#cypher-export\" rel=\"nofollow ugc noopener\">Cypher\u003C\u002Fa>, and other K\u003C\u002Fp>\n\u003C\u002Fli>\n\u003C\u002Ful>\n",1787530083373]