AI
CodeGraph: Open-Taxonomy Knowledge Graph for Source Code with Wikidata Grounding
Publication: Hugging Face Daily Papers
Author: Not available in collected metadata
Source published: 23/09/2026, 20:00:00 UTC
Briefing created: 28/09/2026, 21:24:43 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 28/09/2026, 14:18:24 UTC.
Briefings
This article presents CodeGraph, an open-taxonomy knowledge graph for source code built using a code-specialized Large Language Model. The pipeline extracts entities grounded in Wikidata via a three-stage linking procedure, resulting in 158 million nodes and 1 billion typed edges across 14 programming languages. A calibrated quality-assurance protocol combines human gold data with an LLM-as-a-judge filter to quantify annotation precision. The research was applied to the 167 million files of the Stack-Edu corpus, creating the first large-scale open-taxonomy knowledge graph for source code.
Key points
- CodeGraph is the first large-scale open-taxonomy knowledge graph for source code, built using a code-specialized Large Language Model.
- The graph contains 158 million nodes and 1 billion typed edges across 14 programming languages, with 63,000 concept entities and 19,800 grounded Wikidata entities.
Context and limitations
The pipeline was applied to the Stack-Edu corpus, which contains 167 million files. The graph's coverage is limited to 14 programming languages, and the knowledge extraction relies on a code-specialized LLM rather than traditional syntactic analysis tools.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 28/09/2026, 21:24:43 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeigo5qenvbpyjfjm6t5pwmde4xv277dhgrdizhxkgkcsocyfn5r7cq