TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Semantic Search-as-You-Type

16 点作者 generall超过 1 年前

2 条评论

akmodi超过 1 年前
This a interesting article!<p>How do you build the prefixes when multiple words share the same prefix?<p>If my understanding is correct, the method is:<p>1. Find common search terms<p>2. For each search term (eg test, compute it&#x27;s vector [1.23, 4.56...] and it&#x27;s prefixes [t, te, tes...])<p>3. Store these in Qdrant as t-&gt;[1.23, 4.56...] , te-&gt;[1.23, 4.56...] , tes-&gt;[1.23, 4.56...] and so on. Here, each of the prefixes are used as point_ids<p>4. When a search query comes in, call &#x2F;recommend and pass in the partial query as the point id
andre-z超过 1 年前
Made possible with Rust and a few optimization tricks.<p>- Qdrant as a vector search engine<p>- ONNX inference in Rust<p>- Embeddings cache &amp; lookup<p>- Parallel &amp; Batch requests<p>- Hybrid search with full-text filtering + vector re-scoring<p>Code repo <a href="https:&#x2F;&#x2F;github.com&#x2F;qdrant&#x2F;page-search">https:&#x2F;&#x2F;github.com&#x2F;qdrant&#x2F;page-search</a>