@kurajs/cli 0.0.2 → 0.0.4

package/dist/cli.js

@@ -6,7 +6,6 @@
 //     and never reads the filesystem (Workers-safe; mirrors June's app/_content.ts freeze).
 //     A content hash short-circuits re-embedding when nothing changed (cheap to run pre-dev).
 import { buildIndex } from "@kurajs/docs/search";
-import { transformers } from "@kurajs/transformers";
 import fs from "node:fs";
 import path from "node:path";
 import crypto from "node:crypto";
@@ -53,28 +52,45 @@ async function cmdIndex() {
             if (e.locale === locale)
                 variants.push(e);
     const allEntries = [...DOCS, ...variants];
-    // Content hash — skip the (slow) embed when nothing changed, so `kura index` is cheap to run
-    // before every dev/build. Covers locale + slug + body of every entry, plus the model id.
-    const hashInput = JSON.stringify([model, allEntries.map((e) => [e.locale ?? "", e.slug, e.body])]);
+    // --no-embed: skip the search index entirely (build MDX only). Use it when the site runs
+    // WITHOUT a runtime embedder — search degrades to a lexical scan, so no model/index is needed
+    // (e.g. a Cloudflare Workers deploy without Workers AI).
+    const noEmbed = process.argv.includes("--no-embed");
+    // Content hash — skip rebuilds when nothing changed, so `kura index` is cheap to run before
+    // every dev/build. Covers the mode + model + locale/slug/body of every entry.
+    const hashInput = JSON.stringify([model, noEmbed, allEntries.map((e) => [e.locale ?? "", e.slug, e.body])]);
     const contentHash = crypto.createHash("sha256").update(hashInput).digest("hex").slice(0, 16);
-    if (fs.existsSync(mdxTs) && fs.existsSync(indexTs)) {
-        const prev = fs.readFileSync(indexTs, "utf8").match(/CONTENT_HASH = "([^"]+)"/)?.[1];
-        if (prev === contentHash) {
-            console.log(`kura index: up to date (${allEntries.length} docs, hash ${contentHash}) — skipped`);
-            return;
-        }
+    const stamp = `// content-hash: ${contentHash}\n`;
+    const hashOf = (f) => (fs.existsSync(f) ? fs.readFileSync(f, "utf8").match(/content-hash: (\S+)/)?.[1] : undefined);
+    const upToDate = noEmbed ? hashOf(mdxTs) === contentHash : hashOf(indexTs) === contentHash && fs.existsSync(mdxTs);
+    if (upToDate) {
+        console.log(`kura index: up to date (${allEntries.length} docs, hash ${contentHash}) — skipped`);
+        return;
+    }
+    fs.mkdirSync(path.dirname(mdxTs), { recursive: true });
+    if (!noEmbed) {
+        const localeTag = locales.size ? ` (+${variants.length} variants across ${locales.size} locales)` : "";
+        console.log(`kura index: embedding ${DOCS.length} docs${localeTag} (model ${model})…`);
+        const t0 = Date.now();
+        // Lazy import: only the embed path needs the ML stack, so --no-embed sites never load it
+        // (and never install @kurajs/transformers — it's an optional peer).
+        const { transformers } = await import("@kurajs/transformers").catch(() => {
+            console.error("kura index: embedding needs @kurajs/transformers — install it, or use --no-embed.");
+            return process.exit(1);
+        });
+        const bytes = await buildIndex({ entries: allEntries, embedder: transformers({ model }) });
+        const b64 = Buffer.from(bytes).toString("base64");
+        fs.writeFileSync(indexTs, stamp +
+            "// AUTO-GENERATED by `kura index` — do not edit. Frozen search index (base64) so the worker\n" +
+            "// bundle imports it instead of reading the filesystem (Workers-safe).\n" +
+            `export const INDEX_B64 = ${JSON.stringify(b64)};\n`);
+        console.log(`kura index: wrote app/_index.ts (${(bytes.length / 1024).toFixed(0)}KB index) in ${((Date.now() - t0) / 1000).toFixed(1)}s`);
+    }
+    else {
+        if (fs.existsSync(indexTs))
+            fs.rmSync(indexTs);
+        console.log("kura index: --no-embed — MDX only, no search index (runtime search is lexical)");
     }
-    const localeTag = locales.size ? ` (+${variants.length} variants across ${locales.size} locales)` : "";
-    console.log(`kura index: embedding ${DOCS.length} docs${localeTag} (model ${model})…`);
-    const t0 = Date.now();
-    const bytes = await buildIndex({ entries: allEntries, embedder: transformers({ model }) });
-    const b64 = Buffer.from(bytes).toString("base64");
-    fs.mkdirSync(path.dirname(indexTs), { recursive: true });
-    fs.writeFileSync(indexTs, "// AUTO-GENERATED by `kura index` — do not edit. Frozen search index (base64) so the worker\n" +
-        "// bundle imports it instead of reading the filesystem (Workers-safe).\n" +
-        `export const CONTENT_HASH = ${JSON.stringify(contentHash)};\n` +
-        `export const INDEX_B64 = ${JSON.stringify(b64)};\n`);
-    console.log(`kura index: wrote app/_index.ts (${(bytes.length / 1024).toFixed(0)}KB index) in ${((Date.now() - t0) / 1000).toFixed(1)}s`);
     // Precompile MDX -> static HTML with curated components (build-time; Workers-safe at runtime).
     // Bucketed by locale: "default" for the flat (default-locale) files, plus one per variant locale.
     const { mdxToHtml } = await import("@kurajs/docs/mdx");
@@ -93,7 +109,8 @@ async function cmdIndex() {
         await render("default", d);
     for (const e of variants)
         await render(e.locale, e);
-    fs.writeFileSync(mdxTs, "// AUTO-GENERATED by `kura index` — do not edit. Frozen precompiled MDX (locale → slug → html).\n" +
+    fs.writeFileSync(mdxTs, stamp +
+        "// AUTO-GENERATED by `kura index` — do not edit. Frozen precompiled MDX (locale → slug → html).\n" +
         `export const MDX: Record<string, Record<string, string>> = ${JSON.stringify(map)};\n`);
     const total = Object.values(map).reduce((n, b) => n + Object.keys(b).length, 0);
     const tag = locales.size ? ` across ${locales.size + 1} locales` : "";
@@ -104,6 +121,8 @@ if (cmd === "index") {
     await cmdIndex();
 }
 else {
-    console.log("Kura CLI\n  kura index [--model <hf-model>]   build & freeze the docs search index + MDX");
+    console.log("Kura CLI\n" +
+        "  kura index [--model <hf-model>]   build & freeze the docs search index + MDX\n" +
+        "  kura index --no-embed             MDX only, no search index (runtime search is lexical)");
     process.exit(cmd ? 1 : 0);
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@kurajs/cli",
-  "version": "0.0.2",
+  "version": "0.0.4",
   "type": "module",
   "description": "Kura CLI — build the docs search index (kura index).",
   "bin": {
@@ -15,14 +15,18 @@
   "license": "MIT",
   "dependencies": {
     "@kurajs/core": "^0.0.1",
-    "@kurajs/docs": "^0.0.2",
-    "@kurajs/transformers": "^0.0.1"
+    "@kurajs/docs": "^0.0.3"
   },
   "peerDependencies": {
-    "@huggingface/transformers": ">=3"
+    "@kurajs/transformers": "^0.0.1"
+  },
+  "peerDependenciesMeta": {
+    "@kurajs/transformers": {
+      "optional": true
+    }
   },
   "devDependencies": {
-    "@huggingface/transformers": "3.8.1",
+    "@kurajs/transformers": "*",
     "typescript": "^5.7.0"
   },
   "author": "Lawrence Lin",