@delali/narsil 0.4.3 → 0.4.4-next.304
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +1 -1
- package/dist/server.mjs +1 -1
- package/package.json +5 -7
package/README.md
CHANGED
|
@@ -15,7 +15,7 @@ Narsil is an open source search engine with full-text, vector, hybrid, and geose
|
|
|
15
15
|
The engine partitions large indexes across workers and merges the results of the partitions into a single ranked list. This TypeScript package is the reference implementation of the `.nrsl` format, and [a Rust crate and a Go module](#other-languages) hold the same engine for programs in those languages.
|
|
16
16
|
|
|
17
17
|
<!-- BENCH:headline START -->
|
|
18
|
-
On BEIR SciFact, Narsil's
|
|
18
|
+
On BEIR SciFact, the BM25 ranking of Narsil's TypeScript server scores 0.681 nDCG@10, which is within 0.002 of the Anserini reference of 0.679. Elasticsearch scores 0.679 and OpenSearch scores 0.679 on the same queries. A paired significance test over those queries finds no significant difference between Narsil, Elasticsearch, and OpenSearch. On SciFact, Narsil's TypeScript server answers 5,152 keyword queries a second at its peak, and OpenSearch follows with 3,842. Narsil's Rust server answers 5,766 keyword queries a second at its peak. On DBpedia entities 1M with every engine held at 0.99 recall, Narsil's TypeScript server answers 1,129 vector queries a second at its peak, while OpenSearch leads with 1,465. Narsil's Rust server answers 1,299 vector queries a second at its peak. With each engine's recommended production settings on the same set, Narsil's TypeScript server answers 1,649 vector queries a second at its peak, and OpenSearch follows with 1,592. Narsil's Rust server answers 1,642 vector queries a second at its peak. Narsil has the highest peak throughput in 16 of the 20 keyword, vector, and hybrid comparisons in the newest recorded run: 10 of 12 with every engine on equal settings, and 6 of 8 with each engine's recommended production settings. The harness recorded these figures on 2026-11-17 in run `20261117T070716Z`. GCP c3-standard-8 hardware in us-central1-a, us-east1-b, us-east4-a, us-south1-b, us-west1-a, us-west2-a, us-west3-b, and us-west4-a hosted the engines. You'll find every dataset, every engine's settings, and the full method in [`BENCHMARKS.md`](https://github.com/assetcorp/narsil/blob/main/BENCHMARKS.md).
|
|
19
19
|
<!-- BENCH:headline END -->
|
|
20
20
|
|
|
21
21
|
Try it in your browser at [narsil.sondelali.com/demo](https://narsil.sondelali.com/demo). Read the full documentation at [narsil.sondelali.com/docs](https://narsil.sondelali.com/docs).
|
package/dist/server.mjs
CHANGED
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@delali/narsil",
|
|
3
|
-
"version": "0.4.
|
|
3
|
+
"version": "0.4.4-next.304",
|
|
4
4
|
"type": "module",
|
|
5
5
|
"description": "Distributed search engine with full-text, vector, hybrid, and geosearch. One engine, embedded or on a server, with a cross-language .nrsl format.",
|
|
6
6
|
"author": "Delali (https://sondelali.com)",
|
|
@@ -656,11 +656,9 @@
|
|
|
656
656
|
}
|
|
657
657
|
},
|
|
658
658
|
"optionalDependencies": {
|
|
659
|
-
"@delali/narsil-native-
|
|
660
|
-
"@delali/narsil-native-
|
|
661
|
-
"@delali/narsil-native-
|
|
662
|
-
"@delali/narsil-native-
|
|
663
|
-
"@delali/narsil-native-win32-arm64": "0.4.3",
|
|
664
|
-
"@delali/narsil-native-win32-x64": "0.4.3"
|
|
659
|
+
"@delali/narsil-native-linux-arm64": "0.4.4-next.304",
|
|
660
|
+
"@delali/narsil-native-linux-x64": "0.4.4-next.304",
|
|
661
|
+
"@delali/narsil-native-win32-arm64": "0.4.4-next.304",
|
|
662
|
+
"@delali/narsil-native-win32-x64": "0.4.4-next.304"
|
|
665
663
|
}
|
|
666
664
|
}
|