StackMap
Subscribe

docling-graph vs knowledge_graph

Documents to validated knowledge graphs: Docling parses, an LLM or VLM fills Pydantic schemas, and you get a directed NetworkX graph with provenance, Cypher/CSV export and HTML views. — versus — Notebook recipe that turns any text corpus into a concept graph with a local Mistral 7B via Ollama — chunk, extract concepts and relations, add proximity edges — for Graph RAG and KG QA.

The curated verdict

Both go documents → knowledge graph. docling-graph parses with Docling and fills Pydantic schemas so the graph is validated; knowledge_graph is schema-less concept co-occurrence — quick to see, hard to trust.

docling-graphknowledge_graph
Stars8524.0k
Forks88627
LanguagePythonJupyter Notebook
LicenseMITMIT
Last activity3 days ago21 days ago
Topicsknowledge-graphs, ocrknowledge-graphs, rag, local
Curated connections63

docling-graph — the curator's take

Built for the domains where an approximate embedding is worthless — chemistry, finance, legal — because what you need is the exact edge between a compound and a reaction or an instrument and its dependency. Pydantic schemas make extraction validated rather than hopeful, template generation bootstraps those schemas from example documents or an OWL/RDFS ontology, the provenance ledger carries bounding-box geometry with no extra LLM calls, and graph fusion merges multiple graphs deterministically. Under the Linux Foundation AI & Data umbrella with the Docling project's engineering behind it. Caveats: you own the schema design, which is the actual work; VLM extraction needs the heavy `vlm` extra plus local GPU or a served model; and the output is a NetworkX graph you export, not a queryable graph database.

knowledge_graph — the curator's take

Read this to *learn* how a text-to-knowledge-graph pipeline works: it is a single notebook that chunks a PDF, asks a local model for concepts per chunk (concepts, not NER entities — the author's good point), treats co-occurrence as a weighted edge, and visualises with NetworkX/pyvis. Free to run, GPU-free, honest about being simple. NOT a library: no schema, no incremental updates, no store, no retrieval layer — 4k stars are for the tutorial, not for shipping. When you've understood the recipe, move to a tool that validates against a schema and persists.