2023-06-04 23:35:45 +07:00
|
|
|
import esbuild from 'esbuild'
|
2023-05-30 22:02:20 +07:00
|
|
|
import remarkParse from 'remark-parse'
|
|
|
|
import remarkRehype from 'remark-rehype'
|
|
|
|
import { Processor, unified } from "unified"
|
|
|
|
import { Root as MDRoot } from 'remark-parse/lib'
|
|
|
|
import { Root as HTMLRoot } from 'hast'
|
|
|
|
import { ProcessedContent } from '../plugins/vfile'
|
|
|
|
import { PerfTimer } from '../perf'
|
|
|
|
import { read } from 'to-vfile'
|
2023-07-13 14:19:35 +07:00
|
|
|
import { FilePath, ServerSlug, slugifyFilePath } from '../path'
|
2023-05-30 22:02:20 +07:00
|
|
|
import path from 'path'
|
2023-06-04 23:35:45 +07:00
|
|
|
import os from 'os'
|
|
|
|
import workerpool, { Promise as WorkerPromise } from 'workerpool'
|
2023-06-12 13:26:43 +07:00
|
|
|
import { QuartzTransformerPluginInstance } from '../plugins/types'
|
2023-06-05 00:37:43 +07:00
|
|
|
import { QuartzLogger } from '../log'
|
|
|
|
import chalk from 'chalk'
|
2023-05-30 22:02:20 +07:00
|
|
|
|
|
|
|
export type QuartzProcessor = Processor<MDRoot, HTMLRoot, void>
|
2023-06-12 13:26:43 +07:00
|
|
|
export function createProcessor(transformers: QuartzTransformerPluginInstance[]): QuartzProcessor {
|
2023-05-30 22:02:20 +07:00
|
|
|
// base Markdown -> MD AST
|
|
|
|
let processor = unified().use(remarkParse)
|
|
|
|
|
|
|
|
// MD AST -> MD AST transforms
|
2023-06-20 12:50:25 +07:00
|
|
|
for (const plugin of transformers.filter(p => p.markdownPlugins)) {
|
|
|
|
processor = processor.use(plugin.markdownPlugins!())
|
2023-05-30 22:02:20 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
// MD AST -> HTML AST
|
|
|
|
processor = processor.use(remarkRehype, { allowDangerousHtml: true })
|
|
|
|
|
|
|
|
|
|
|
|
// HTML AST -> HTML AST transforms
|
2023-06-20 12:50:25 +07:00
|
|
|
for (const plugin of transformers.filter(p => p.htmlPlugins)) {
|
|
|
|
processor = processor.use(plugin.htmlPlugins!())
|
2023-05-30 22:02:20 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
return processor
|
|
|
|
}
|
|
|
|
|
2023-06-04 23:35:45 +07:00
|
|
|
function* chunks<T>(arr: T[], n: number) {
|
|
|
|
for (let i = 0; i < arr.length; i += n) {
|
|
|
|
yield arr.slice(i, i + n)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-06-05 00:37:43 +07:00
|
|
|
async function transpileWorkerScript() {
|
2023-06-04 23:35:45 +07:00
|
|
|
// transpile worker script
|
|
|
|
const cacheFile = "./.quartz-cache/transpiled-worker.mjs"
|
|
|
|
const fp = "./quartz/worker.ts"
|
2023-06-05 00:37:43 +07:00
|
|
|
return esbuild.build({
|
2023-06-04 23:35:45 +07:00
|
|
|
entryPoints: [fp],
|
|
|
|
outfile: path.join("quartz", cacheFile),
|
|
|
|
bundle: true,
|
|
|
|
keepNames: true,
|
|
|
|
platform: "node",
|
|
|
|
format: "esm",
|
|
|
|
packages: "external",
|
|
|
|
plugins: [
|
|
|
|
{
|
|
|
|
name: 'css-and-scripts-as-text',
|
|
|
|
setup(build) {
|
|
|
|
build.onLoad({ filter: /\.scss$/ }, (_) => ({
|
|
|
|
contents: '',
|
|
|
|
loader: 'text'
|
|
|
|
}))
|
|
|
|
build.onLoad({ filter: /\.inline\.(ts|js)$/ }, (_) => ({
|
|
|
|
contents: '',
|
|
|
|
loader: 'text'
|
|
|
|
}))
|
|
|
|
}
|
|
|
|
}
|
|
|
|
]
|
|
|
|
})
|
|
|
|
}
|
|
|
|
|
2023-07-13 14:19:35 +07:00
|
|
|
export function createFileParser(transformers: QuartzTransformerPluginInstance[], baseDir: string, fps: FilePath[], allSlugs: ServerSlug[], verbose: boolean) {
|
2023-06-05 00:37:43 +07:00
|
|
|
return async (processor: QuartzProcessor) => {
|
|
|
|
const res: ProcessedContent[] = []
|
|
|
|
for (const fp of fps) {
|
|
|
|
try {
|
|
|
|
const file = await read(fp)
|
|
|
|
|
2023-07-07 09:18:18 +07:00
|
|
|
// strip leading and trailing whitespace
|
|
|
|
file.value = file.value.toString().trim()
|
|
|
|
|
2023-06-20 12:50:25 +07:00
|
|
|
// Text -> Text transforms
|
|
|
|
for (const plugin of transformers.filter(p => p.textTransform)) {
|
|
|
|
file.value = plugin.textTransform!(file.value)
|
|
|
|
}
|
|
|
|
|
2023-06-05 00:37:43 +07:00
|
|
|
// base data properties that plugins may use
|
2023-07-13 14:19:35 +07:00
|
|
|
file.data.slug = slugifyFilePath(path.relative(baseDir, file.path) as FilePath)
|
2023-07-10 09:32:24 +07:00
|
|
|
file.data.allSlugs = allSlugs
|
2023-06-05 00:37:43 +07:00
|
|
|
file.data.filePath = fp
|
|
|
|
|
|
|
|
const ast = processor.parse(file)
|
|
|
|
const newAst = await processor.run(ast, file)
|
|
|
|
res.push([newAst, file])
|
|
|
|
|
|
|
|
if (verbose) {
|
|
|
|
console.log(`[process] ${fp} -> ${file.data.slug}`)
|
|
|
|
}
|
|
|
|
} catch (err) {
|
2023-06-06 14:00:38 +07:00
|
|
|
console.log(chalk.red(`\nFailed to process \`${fp}\`: `) + err)
|
2023-06-05 00:37:43 +07:00
|
|
|
process.exit(1)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return res
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-07-13 14:19:35 +07:00
|
|
|
export async function parseMarkdown(transformers: QuartzTransformerPluginInstance[], baseDir: string, fps: FilePath[], verbose: boolean): Promise<ProcessedContent[]> {
|
2023-05-30 22:02:20 +07:00
|
|
|
const perf = new PerfTimer()
|
2023-06-05 00:37:43 +07:00
|
|
|
const log = new QuartzLogger(verbose)
|
2023-06-04 23:35:45 +07:00
|
|
|
|
|
|
|
const CHUNK_SIZE = 128
|
|
|
|
let concurrency = fps.length < CHUNK_SIZE ? 1 : os.availableParallelism()
|
2023-06-05 00:37:43 +07:00
|
|
|
|
2023-07-10 09:32:24 +07:00
|
|
|
// get all slugs ahead of time as each thread needs a copy
|
2023-07-13 14:19:35 +07:00
|
|
|
const allSlugs = fps.map(fp => slugifyFilePath(path.relative(baseDir, path.resolve(fp)) as FilePath))
|
2023-07-10 09:32:24 +07:00
|
|
|
|
|
|
|
let res: ProcessedContent[] = []
|
2023-06-05 00:37:43 +07:00
|
|
|
log.start(`Parsing input files using ${concurrency} threads`)
|
2023-06-04 23:35:45 +07:00
|
|
|
if (concurrency === 1) {
|
|
|
|
const processor = createProcessor(transformers)
|
2023-07-10 09:32:24 +07:00
|
|
|
const parse = createFileParser(transformers, baseDir, fps, allSlugs, verbose)
|
2023-06-05 00:37:43 +07:00
|
|
|
res = await parse(processor)
|
2023-06-04 23:35:45 +07:00
|
|
|
} else {
|
2023-06-05 00:37:43 +07:00
|
|
|
await transpileWorkerScript()
|
2023-06-04 23:35:45 +07:00
|
|
|
const pool = workerpool.pool(
|
|
|
|
'./quartz/bootstrap-worker.mjs',
|
|
|
|
{
|
|
|
|
minWorkers: 'max',
|
|
|
|
maxWorkers: concurrency,
|
|
|
|
workerType: 'thread'
|
|
|
|
}
|
|
|
|
)
|
2023-05-30 22:02:20 +07:00
|
|
|
|
2023-06-04 23:35:45 +07:00
|
|
|
const childPromises: WorkerPromise<ProcessedContent[]>[] = []
|
|
|
|
for (const chunk of chunks(fps, CHUNK_SIZE)) {
|
2023-07-10 09:32:24 +07:00
|
|
|
childPromises.push(pool.exec('parseFiles', [baseDir, chunk, allSlugs, verbose]))
|
2023-05-30 22:02:20 +07:00
|
|
|
}
|
2023-06-05 00:37:43 +07:00
|
|
|
|
2023-06-04 23:35:45 +07:00
|
|
|
const results: ProcessedContent[][] = await WorkerPromise.all(childPromises)
|
2023-06-05 00:37:43 +07:00
|
|
|
res = results.flat()
|
2023-06-04 23:35:45 +07:00
|
|
|
await pool.terminate()
|
2023-05-30 22:02:20 +07:00
|
|
|
}
|
|
|
|
|
2023-06-05 00:37:43 +07:00
|
|
|
log.success(`Parsed ${res.length} Markdown files in ${perf.timeSince()}`)
|
2023-05-30 22:02:20 +07:00
|
|
|
return res
|
|
|
|
}
|