feat(gatsby): page engine

[pieh]

I suggest to use "Hide whitespace changes" in diff viewer - https://github.com/gatsbyjs/gatsby/pull/33067/files?diff=unified&w=1

Building on top of GraphQL Engine introduced in #33030

Description

This feature is disabled by default. Only available when building gatsby itself with COMPILER_OPTIONS="GATSBY_MAJOR=4"

This adds generation of "page engine" - importable standalone module that allows generation of page-data and html content. It uses "query engine" (introduced in #33030 ) under the hood.

Example usage (after initial gatsby build):

const { GraphQLEngine } = require(`./.cache/query-engine`)
const { getData, renderHTML, renderPageData } = require(`./.cache/page-ssr`)

const graphqlEngine = new GraphQLEngine({
  dbPath: process.cwd() + `/.cache/data/datastore`,
})

async function run() {
  const data = await getData({ graphqlEngine, pathName })
  const pageData = await renderPageData({ data })
  const html = await renderHTML({ data, pageData })

  console.log({ pageData, html })
}

run()

[ch37670]

Fixes #32892

[pieh]

Extracted to separate module from packages/gatsby/src/utils/worker/child/render-html.ts

[vladar]

Just to clarify - we need to have it as a separate module because we re-use it later in bundle-webpack, right?

[pieh]

Hmm, possibly we could revert this split because we don't use it in engine itself. But it's still beneficial because we don't have to load more things from workers code than needed (child/render-html.ts is not loaded by main process, just by html/pqr worker processes)

[pieh]

Extracted to separate module from packages/gatsby/src/utils/page-data.ts + addition of getPagePathFromPageDataPath to reverse /page-data/<page-path>/page-data.json to <page-path>

[pieh]

Extracted to separate module from packages/gatsby/src/utils/worker/child/render-html.ts

[vladar]

LGTM 👍