xerj-search
reference corpus · status live · category DATA (Reference implementations — data engines) · lane A · refresh pinned
What an agent uses it for
an agent working on full-text search internals queries this for how lucene/tantivy/quickwit/meilisearch/sonic/elasticsearch do BM25, merge policy and the ES wire
Use it
xerj corpus add --from https://raw.githubusercontent.com/xerj-org/xerj/corpus-hub/tools/xerj-code/hub/xerj-search.jsonxerj corpus index xerj-search → xerj code xerj-search "your question"Sources & pins
| source | pin | licence | size | review note |
|---|---|---|---|---|
| tantivy | 3a55bc986d10… | MIT | 34.8 MB / 527 files | |
| quickwit | fb99535bccc4… | Apache-2.0 | 30.4 MB / 1,555 files | LICENSE is plain Apache-2.0; LICENSE-3rdparty.csv is a dependency inventory, not this project's licence. |
| meilisearch | be181e115792… | MIT AND BUSL-1.1 | 51.5 MB / 1,734 files | MIT core, but Enterprise Edition parts are BUSL-1.1 (LICENSE-EE). Check the file's header and path before copying anything. |
| sonic | 484339e891a5… | MPL-2.0 | 1.6 MB / 244 files | File-level copyleft. XERJ is Apache-2.0, so read the design and write your own code; do not lift MPL files in. |
| elasticsearch | ac375eb50ac5… | AGPL-3.0-only OR SSPL-1.0 OR Elastic-2.0 | 559.5 MB / 46,387 files | Triple-licensed; the detector reports only the first (most restrictive) match. Never copy ES code into XERJ — beyond the licence, XERJ's public position is that it shares no code and no architecture with Elasticsearch. |
| lucene | 9cda46563a4b… | Apache-2.0 | 188.6 MB / 7,557 files | Plain Apache-2.0 — the same licence as XERJ — so unlike elasticsearch in this corpus it is adapt-with-attribution, not approach-only. Lucene is the library elasticsearch is built on, so it is the primary source for postings, codecs, segment merge, BM25 scoring and HNSW; prefer it over the ES entry. |
866.5 MB across 58,004 files · added 2026-08-07
Rights
AGPL-3.0-only OR SSPL-1.0 OR Elastic-2.0 + Apache-2.0 + MIT + MIT AND BUSL-1.1 + MPL-2.0 — redistribution permitted with attribution per the licence review . The review block records a human opening each licence file at the pin; detector output is a hint, never the verdict.
Retrieval spot-check (G7)
not yet spot-checked; queries are pre-registered in the registry before grading.