MediaWiki Code2Code Search：面向开源软件实体语义发现的神经检索系统

View PDF HTML (experimental)

Abstract:Code search in large-scale ecosystems is often hindered by the lexical gap between user queries and implementation details, alongside the trade-off between the low latency of traditional Information Retrieval (IR) and the precision of Deep Learning (DL). We present MediaWiki Code2Code Search, a neural retrieval system for semantic code-to-code discovery. By indexing 1.29 million structural entities (functions, types, and templates) across 2,500+ MediaWiki repositories, our system enables retrieval based on computational intent rather than surface tokens. We employ a split-build architecture, decoupling GPU-intensive offline indexing from a CPU-only serving layer; our FAISS IVF-PQ index occupies 168.6 MiB: a 96.6\% reduction compared to a flat float32 baseline, and achieves a median query latency of 1.85 seconds on commodity hardware, satisfying the 6 GiB RAM constraint of Wikimedia Toolforge. In a blinded evaluation graded by a three-model LLM jury (inter-judge Krippendorff's $\alpha = 0.74$), our system improves ranking quality over a BM25 baseline across a 27-query benchmark, raising the jury-mean nDCG@10 (to 0.81 from 0.63), a positive and consistent gain across all judges. Gains are most pronounced in name-obfuscated and cross-language tasks where lexical methods fail. The system is available at this https URL under the Apache 2.0 licence and provides an open RESTful API.

Comments:	Published in SoftwareX. 24 pages, 3 figures, 5 tables
Subjects:	Information Retrieval (cs.IR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Software Engineering (cs.SE)
MSC classes:	68P20, 68N30, 68T50
ACM classes:	H.3.3; D.2.7; I.2.7; D.2.8
Cite as:	arXiv:2607.26766 [cs.IR]
	(or arXiv:2607.26766v2 [cs.IR] for this version)
	https://doi.org/10.48550/arXiv.2607.26766 arXiv-issued DOI via DataCite
Journal reference:	SoftwareX 36 (2026) 103127
Related DOI:	https://doi.org/10.1016/j.softx.2026.103127 DOI(s) linking to related resources

Submission history

From: Francesco Tosoni [view email]
[v1] Wed, 29 Jul 2026 11:07:07 UTC (200 KB)
[v2] Sat, 31 Oct 2026 11:58:28 UTC (171 KB)