From c90dbacab05d2ceab84b53c66c7be77b4dda47ef Mon Sep 17 00:00:00 2001
From: Anton Bulakh <him@necauq.ua>
Date: Tue, 07 Jan 2025 20:33:34 +0000
Subject: [PATCH] chore(build): separate markdown and html handling into two separate stages (#1675)
---
quartz/processors/parse.ts | 90 ++++++++++++++++++++++++++++++++------------
1 files changed, 65 insertions(+), 25 deletions(-)
diff --git a/quartz/processors/parse.ts b/quartz/processors/parse.ts
index 2bd530c..479313f 100644
--- a/quartz/processors/parse.ts
+++ b/quartz/processors/parse.ts
@@ -4,18 +4,20 @@
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 { MarkdownContent, ProcessedContent } from "../plugins/vfile"
import { PerfTimer } from "../util/perf"
import { read } from "to-vfile"
-import { FilePath, QUARTZ, slugifyFilePath } from "../util/path"
+import { FilePath, FullSlug, QUARTZ, slugifyFilePath } from "../util/path"
import path from "path"
import workerpool, { Promise as WorkerPromise } from "workerpool"
import { QuartzLogger } from "../util/log"
import { trace } from "../util/trace"
import { BuildCtx } from "../util/ctx"
-export type QuartzProcessor = Processor<MDRoot, MDRoot, HTMLRoot>
-export function createProcessor(ctx: BuildCtx): QuartzProcessor {
+export type QuartzMdProcessor = Processor<MDRoot, MDRoot, MDRoot>
+export type QuartzHtmlProcessor = Processor<undefined, MDRoot, HTMLRoot>
+
+export function createMdProcessor(ctx: BuildCtx): QuartzMdProcessor {
const transformers = ctx.cfg.plugins.transformers
return (
@@ -24,14 +26,20 @@
.use(remarkParse)
// MD AST -> MD AST transforms
.use(
- transformers
- .filter((p) => p.markdownPlugins)
- .flatMap((plugin) => plugin.markdownPlugins!(ctx)),
- )
+ transformers.flatMap((plugin) => plugin.markdownPlugins?.(ctx) ?? []),
+ ) as unknown as QuartzMdProcessor
+ // ^ sadly the typing of `use` is not smart enough to infer the correct type from our plugin list
+ )
+}
+
+export function createHtmlProcessor(ctx: BuildCtx): QuartzHtmlProcessor {
+ const transformers = ctx.cfg.plugins.transformers
+ return (
+ unified()
// MD AST -> HTML AST
.use(remarkRehype, { allowDangerousHtml: true })
// HTML AST -> HTML AST transforms
- .use(transformers.filter((p) => p.htmlPlugins).flatMap((plugin) => plugin.htmlPlugins!(ctx)))
+ .use(transformers.flatMap((plugin) => plugin.htmlPlugins?.(ctx) ?? []))
)
}
@@ -75,8 +83,8 @@
export function createFileParser(ctx: BuildCtx, fps: FilePath[]) {
const { argv, cfg } = ctx
- return async (processor: QuartzProcessor) => {
- const res: ProcessedContent[] = []
+ return async (processor: QuartzMdProcessor) => {
+ const res: MarkdownContent[] = []
for (const fp of fps) {
try {
const perf = new PerfTimer()
@@ -100,10 +108,32 @@
res.push([newAst, file])
if (argv.verbose) {
- console.log(`[process] ${fp} -> ${file.data.slug} (${perf.timeSince()})`)
+ console.log(`[markdown] ${fp} -> ${file.data.slug} (${perf.timeSince()})`)
}
} catch (err) {
- trace(`\nFailed to process \`${fp}\``, err as Error)
+ trace(`\nFailed to process markdown \`${fp}\``, err as Error)
+ }
+ }
+
+ return res
+ }
+}
+
+export function createMarkdownParser(ctx: BuildCtx, mdContent: MarkdownContent[]) {
+ return async (processor: QuartzHtmlProcessor) => {
+ const res: ProcessedContent[] = []
+ for (const [ast, file] of mdContent) {
+ try {
+ const perf = new PerfTimer()
+
+ const newAst = await processor.run(ast as MDRoot, file)
+ res.push([newAst, file])
+
+ if (ctx.argv.verbose) {
+ console.log(`[html] ${file.data.slug} (${perf.timeSince()})`)
+ }
+ } catch (err) {
+ trace(`\nFailed to process html \`${file.data.filePath}\``, err as Error)
}
}
@@ -113,6 +143,7 @@
const clamp = (num: number, min: number, max: number) =>
Math.min(Math.max(Math.round(num), min), max)
+
export async function parseMarkdown(ctx: BuildCtx, fps: FilePath[]): Promise<ProcessedContent[]> {
const { argv } = ctx
const perf = new PerfTimer()
@@ -126,9 +157,8 @@
log.start(`Parsing input files using ${concurrency} threads`)
if (concurrency === 1) {
try {
- const processor = createProcessor(ctx)
- const parse = createFileParser(ctx, fps)
- res = await parse(processor)
+ const mdRes = await createFileParser(ctx, fps)(createMdProcessor(ctx))
+ res = await createMarkdownParser(ctx, mdRes)(createHtmlProcessor(ctx))
} catch (error) {
log.end()
throw error
@@ -140,17 +170,27 @@
maxWorkers: concurrency,
workerType: "thread",
})
-
- const childPromises: WorkerPromise<ProcessedContent[]>[] = []
- for (const chunk of chunks(fps, CHUNK_SIZE)) {
- childPromises.push(pool.exec("parseFiles", [ctx.buildId, argv, chunk, ctx.allSlugs]))
+ const errorHandler = (err: any) => {
+ console.error(`${err}`.replace(/^error:\s*/i, ""))
+ process.exit(1)
}
- const results: ProcessedContent[][] = await WorkerPromise.all(childPromises).catch((err) => {
- const errString = err.toString().slice("Error:".length)
- console.error(errString)
- process.exit(1)
- })
+ const mdPromises: WorkerPromise<[MarkdownContent[], FullSlug[]]>[] = []
+ for (const chunk of chunks(fps, CHUNK_SIZE)) {
+ mdPromises.push(pool.exec("parseMarkdown", [ctx.buildId, argv, chunk]))
+ }
+ const mdResults: [MarkdownContent[], FullSlug[]][] =
+ await WorkerPromise.all(mdPromises).catch(errorHandler)
+
+ const childPromises: WorkerPromise<ProcessedContent[]>[] = []
+ for (const [_, extraSlugs] of mdResults) {
+ ctx.allSlugs.push(...extraSlugs)
+ }
+ for (const [mdChunk, _] of mdResults) {
+ childPromises.push(pool.exec("processHtml", [ctx.buildId, argv, mdChunk, ctx.allSlugs]))
+ }
+ const results: ProcessedContent[][] = await WorkerPromise.all(childPromises).catch(errorHandler)
+
res = results.flat()
await pool.terminate()
}
--
Gitblit v1.10.0