docs: update full-text-search.md (#447)

This commit is contained in:
Ben Schlegel 2023-09-04 07:36:30 +02:00 committed by GitHub
parent b3bd6f7c01
commit 25541d2e35

View File

@ -6,9 +6,11 @@ tags:
Full-text search in Quartz is powered by [Flexsearch](https://github.com/nextapps-de/flexsearch). It's fast enough to return search results in under 10ms for Quartzs as large as half a million words.
It can be opened by either clicking on the search bar or pressing ⌘+K. The top 5 search results are shown on each query. Matching subterms are highlighted and the most relevant 30 words are excerpted. Clicking on a search result will navigate to that page.
It can be opened by either clicking on the search bar or pressing ``/`ctrl` + `K`. The top 5 search results are shown on each query. Matching subterms are highlighted and the most relevant 30 words are excerpted. Clicking on a search result will navigate to that page.
This component is also keyboard accessible: Tab and Shift+Tab will cycle forward and backward through search results and Enter will navigate to the highlighted result (first result by default).
To search content by tags, you can either press `⌘`/`ctrl` + `shift` + `K` or start your query with `#` (e.g. `#components`).
This component is also keyboard accessible: Tab and Shift+Tab will cycle forward and backward through search results and Enter will navigate to the highlighted result (first result by default). You are also able to navigate search results using `ArrowUp` and `ArrowDown`.
> [!info]
> Search requires the `ContentIndex` emitter plugin to be present in the [[configuration]].
@ -17,7 +19,7 @@ This component is also keyboard accessible: Tab and Shift+Tab will cycle forward
By default, it indexes every page on the site with **Markdown syntax removed**. This means link URLs for instance are not indexed.
It properly tokenizes Chinese, Korean, and Japenese characters and constructs separate indexes for the title and content, weighing title matches above content matches.
It properly tokenizes Chinese, Korean, and Japenese characters and constructs separate indexes for the title, content and tags, weighing title matches above content matches.
## Customization
@ -25,4 +27,4 @@ It properly tokenizes Chinese, Korean, and Japenese characters and constructs se
- Component: `quartz/components/Search.tsx`
- Style: `quartz/components/styles/search.scss`
- Script: `quartz/components/scripts/search.inline.ts`
- You can edit `contextWindowWords` or `numSearchResults` to suit your needs
- You can edit `contextWindowWords`, `numSearchResults` or `numTagResults` to suit your needs