feat: rm md build, autorag upload (#27049)
Some checks are pending
Publish / Production (push) Waiting to run

This commit is contained in:
Colby M. White 2025-12-10 14:29:17 -06:00 committed by GitHub
parent 6370fd99cd
commit acba5b6b81
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 0 additions and 96 deletions

View file

@ -37,19 +37,6 @@ jobs:
name: Deploy to Cloudflare Workers
env:
CLOUDFLARE_API_TOKEN: ${{ secrets.CLOUDFLARE_API_TOKEN }}
- name: Install rclone
run: sudo -v ; curl https://rclone.org/install.sh | sudo bash
- name: Build vendored Markdown
run: npx tsx bin/generate-index-md.ts
- name: Upload vendored Markdown files to AI Search DevDocs bucket
env:
AWS_ACCESS_KEY_ID: ${{ secrets.AUTORAG_DEVDOCS_ACCESS_KEY_ID }}
AWS_SECRET_ACCESS_KEY: ${{ secrets.AUTORAG_DEVDOCS_SECRET_ACCESS_KEY }}
run: |
rclone sync \
--config bin/rclone.conf \
distmd \
autorag:developer-docs-full
- uses: actions/cache/save@v4
if: always()
with:

View file

@ -1,76 +0,0 @@
import {
readFileSync,
writeFileSync,
mkdirSync,
appendFileSync,
} from "node:fs";
import glob from "fast-glob";
import { parse } from "node-html-parser";
import { htmlToMarkdown } from "~/util/markdown";
import YAML from "yaml";
const files = await glob("dist/**/*.html");
for (const file of files) {
if (file === "dist/index.html" || file === "dist/404.html") {
continue;
}
const html = readFileSync(file, "utf-8");
const dom = parse(html);
const url = dom
.querySelector("link[rel='alternate'][type='text/markdown']")
?.getAttribute("href");
if (!url) {
continue;
}
const markdown = await htmlToMarkdown(html, url);
if (!markdown) {
continue;
}
const product = file.split("/")[1];
const path = file.replace("dist/", "distmd/").replace(".html", ".md");
mkdirSync(path.split("/").slice(0, -1).join("/"), { recursive: true });
writeFileSync(path, markdown);
const llmsFullContent = ["<page>", markdown, "</page>\n\n"].join("\n");
mkdirSync(`distllms/${product}`, { recursive: true });
appendFileSync("distllms/llms-full.txt", llmsFullContent);
appendFileSync(`distllms/${product}/llms-full.txt`, llmsFullContent);
try {
const path = await glob(`src/content/products/${product}.*`).then((arr) =>
arr.at(0),
);
if (!path) {
continue;
}
const yaml = YAML.parse(readFileSync(path, "utf-8"));
const group = yaml.product?.group?.replaceAll(" ", "-").toLowerCase();
if (!group) {
continue;
}
mkdirSync(`distllms/${group}`, { recursive: true });
appendFileSync(`distllms/${group}/llms-full.txt`, llmsFullContent);
} catch (error) {
if (error instanceof Error) {
console.error(
`Failed to find a product group for ${product}:`,
error.message,
);
}
}
}

View file

@ -1,7 +0,0 @@
[autorag]
type = s3
provider = Cloudflare
env_auth = true
endpoint = https://cf9267a5bf4f6b3cceca48d554d81679.r2.cloudflarestorage.com
acl = private
no_check_bucket = true