feat(theme-algolia): add option.replaceSearchResultPathname to process/replaceAll search result urls

[mellson]

Implemented solution

replaceSearchResultPathname added - doc: https://deploy-preview-8428--docusaurus-2.netlify.app/docs/search/#connecting-algolia

{
  // Optional: Replace parts of the item URLs from Algolia. Useful when using the same search index for multiple deployments using a different baseUrl. You can use regexp or string in the `from` param. For example: localhost:3000 vs myCompany.com/docs
  replaceSearchResultPathname: {
    from: '/docs/', // or as RegExp: /\/docs\//
    to: '/',
    },
}

Pre-flight checklist

• I have read the Contributing Guidelines on pull requests.
• If this is a code change: I have written unit tests and/or added dogfooding pages to fully verify the new behavior.
• If this is a new API or substantial change: the PR has an accompanying issue (closes #0000) and the maintainers have approved on my working plan.

Motivation

We have an upcoming documentation site we're building at https://stately.ai/docs/. So we've set our baseUrl in the config to /docs/. However when we do our preview deploys, these are served from root /.

This setup makes it hard to use the search on the preview deploys because it indexes based on the site with a different baseUrl.

This PR adds the option to replace the item's URL that Algolia returns. So in our case we would like to replace /docs/ with /.

Test Plan

Test links

Deploy preview: https://deploy-preview-8428--docusaurus-2.netlify.app/

Related issues/PRs

[netlify]

✅ [V2]

Name	Link
🔨 Latest commit	7847734
🔍 Latest deploy log	https://app.netlify.com/sites/docusaurus-2/deploys/63a48a09e7b3280008e3ecaf
😎 Deploy Preview	https://deploy-preview-8428--docusaurus-2.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site settings.

[github-actions]

⚡️ Lighthouse report for the deploy preview of this PR

URL	Performance	Accessibility	Best Practices	SEO	PWA	Report
/	🟢 94	🟢 97	🟢 100	🟢 100	🟢 90	Report
/docs/installation	🟠 76	🟢 100	🟢 100	🟢 100	🟢 90	Report

[slorber]

👍 Agree this feature could be useful, and I would probably use it myself

Not 100% fan of the API as it's not very flexible and the naming doesn't feel right

If we can make it better how let's do this. Otherwise I won't block this PR too much because it's a useful but quite niche feature and we can do breaking changes later if needed.

[mellson]

👍 Agree this feature could be useful, and I would probably use it myself

Not 100% fan of the API as it's not very flexible and the naming doesn't feel right

If we can make it better how let's do this. Otherwise I won't block this PR too much because it's a useful but quite niche feature and we can do breaking changes later if needed.

Thank you for your feedback, I agree with all your suggestions and have implemented them in the latest commit.

[slorber]

LGTM 👍

Can you add this to our localhost + deploypreview? That would make sure that we use our own new feature and we could be sure it works end 2 end with the preview.

Example search result:

• https://deploy-preview-8428--docusaurus-2.netlify.app/docs/next/docusaurus-core/
  =>
• https://deploy-preview-8428--docusaurus-2.netlify.app/docs/docusaurus-core/

I guess this should work: {from: "/docs/next", to: "/docs"}

[mellson]

LGTM 👍

Can you add this to our localhost + deploypreview? That would make sure that we use our own new feature and we could be sure it works end 2 end with the preview.

Example search result:

• https://deploy-preview-8428--docusaurus-2.netlify.app/docs/next/docusaurus-core/
  =>
• https://deploy-preview-8428--docusaurus-2.netlify.app/docs/docusaurus-core/

I guess this should work: {from: "/docs/next", to: "/docs"}

That makes sense 👍🏻 I've updated the config to use this in dev- and preview mode. And made it possible to provide a regexp or a string in the from param.

[slorber]

Not super fan of the getRegexpOrString.

I'd prefer either a better solution or to have it removed and just support regexp strings.

I'm a bit afraid this might lead to confusing behaviors when using special chars in from including [({&+... I'd rather let the user decide explicitly if they want to use a RexExp or a simple string replacement

[slorber]

🤔 I don't really like this implementation and would prefer to just support regexp. Adding unit tests would likely reveal the problematic cases.

My idea was more:

• config accepts both string + RegExp
• config validation normalizes this as a serializable RegExp string
• client-side always use new RexExp()

The config normalization can be something like:

if (from instanceof string) {
  return escapeRegexp(from);
} else if (from instanceof RegExp) {
 return from.source;
} else {
  throw new Error('unexpected")
}

=> always return a valid Regexp string source

Does it make sense?

[mellson]

That makes sense; I wasn't terribly happy with getRegexpOrString either. I've tried to address your suggestions in the latest commit, hope I got it right this time 😅

[slorber]

I'm going to push some little refactors in the PR before merging so please don't commit 🤪

[slorber]

Thanks 👍

Looks like it works fine on both the search bar and search page, on localhost + deploy preview