Searching for a text in markdown

[NenadP]

Hi, I am not seeing this in the docs, or maybe its not apparent to me.

Is it possible to search for a text anywhere if I had markdown files?

Say I have a bunch of markdown .md files in my content dir, and I would like to present a separate page to search anywhere in these md files?

In my case:

• user would search by search term
• that searches anywhere in .md files in content directory
• I preset user the list of page titles and a little excerpt from the page where match was found

Questions:

• if this is presently achievable, and if yes, how?
• inf not, if its possible to put it on a feature request list?

[NenadP]

I am looking at the queries, I guess there is body with all of its children but that is likely hard to search.

A workaround would be to provide a custom field in .md files

---
bodyPlainText: ... // copy of all the text without markdown in the md document
---

And then use query content:

const articles = await queryContent()
  .where({ bodyPlainText: { $contains: ['my search text'] } })
  .find()

but bodyPlainText and its value would need to be built programmatically.

(this is obviously naive approach, lots of things to consider like ordering best results etc.)

[farnabaz]

Supporting full-text search is not planned yet. To generate bodyPlainText for contents you can use transformers.

[NenadP]

@farnabaz yeah, I figured so. Depending on the requirement it could be searched also by traversing trees like so:

 // getting all the articles - I guess this approach would be very computational intensive if there is a large amount
 // of content to be searched on the client side and not using something like Algolia
 // In the below code example, all results are aggregated per `_path` (so per page). 
 // But could be easily modified to  a different strategy
 // The search match is wrapped inside the <strong> tag so it could be highlighted (needs to be interpolated with `v-html`)

 // Handy lib for traversing trees
 import { visit } from 'unist-util-visit'

 const articles = await queryContent().find()
 const searchResults = ref([])

 const getSearchResults = (searchTerm: string) => {
    searchResults.value = []
    const searchTermLowerCase = searchTerm.toLocaleLowerCase()
    articles.forEach((article: ParsedContent) => {
      visit(article.body, 'text', (node) => {
        if (node.value.toLocaleLowerCase().includes(searchTermLowerCase)) {
          const existingPageResult = searchResults.value.find(
            (result) => result.path === article._path
          )
          const searchResult = node.value.replaceAll(
            new RegExp(searchTermLowerCase, 'gi'),
            (match) => `<strong>${match}</strong>`
          )
          if (existingPageResult) {
            existingPageResult.searchResult += `<br />${searchResult }`
          } else {
            searchResults.value.push({
              title: article.title,
              searchResult 
            })
          }
        }
      })
    })
  }