perf(build): replace specs glob with fs reads

[atharvadeosthale]

Summary

The 11 import.meta.glob calls in specs.ts (one per Appwrite version) expanded into ~58k lazy import() statements, producing 49,123 server chunks (500 MB) and pushing peak build RSS to 7 GB. The matched .md and .json files are static text — there's no reason to route them through Vite's module graph.

This PR resolves @appwrite.io/specs at runtime via createRequire and reads examples + OpenAPI specs with fs.readFile. The package moves from devDependencies to dependencies so it survives bun install --production and ships to the final image. vite-plugin-dynamic-import is removed (no longer needed).

Build impact

metric	before	after	delta
wall time	192s	113s	−41%
peak RSS	7.0 GB	5.6 GB	−21%
SSR modules	50,269	4,340	−91%
server chunks	49,123	3,194	−93%
build/server/ size	501 MB	80 MB	−84%
final image (build/server + specs in node_modules)	501 MB	376 MB	−125 MB

Test plan

• Build runs clean
• /docs/references/cloud/server-nodejs/databases renders code examples (Node.js)
• /docs/references/1.5.x/server-python/storage renders (Python, older version)
• /docs/references/1.7.x/client-android-kotlin/account renders (exercises the Android special path in loadExample)

[appwrite]

Appwrite Website

Project ID: 69d7efb00023389e8d27

Sites (1)

Site	Status	Logs	Preview	QR
website 69d7f2670014e24571ca	Building	View Logs	Preview URL

Website (appwrite/website)

Project ID: 684969cb000a2f6c0a02

Sites (1)

Site	Status	Logs	Preview	QR
website 68496a17000f03d62013	Queued	View Logs	Preview URL

---

Tip

Trigger functions via HTTP, SDKs, events, webhooks, or scheduled cron jobs

[greptile-apps]

Greptile Summary

• Replaces 11 import.meta.glob calls (which produced ~58k lazy imports, 49k server chunks, and 501 MB build output) with direct fs.readFile + createRequire-based resolution, cutting build time by 41% and server bundle size by 84%.
• Adds input validation (assertValidVersion, assertValidPlatform) to guard filesystem paths derived from URL route params, and an apiCache Map to avoid re-parsing large OpenAPI JSON files on every request.
• The apiCache key is ${version}|${platform} but the file actually read is keyed on mode (server/client/console), so multiple platform variants of the same mode/version get separate cache entries for identical file contents — the fix is to key by ${version}|${mode} instead.

Confidence Score: 4/5

Safe to merge after fixing the cache key; all other changes are well-structured and correct.

One P1 defect: the apiCache key uses the full platform string rather than the derived mode, causing the same large JSON to be parsed and stored once per platform variant instead of once per mode. The rest of the PR — glob removal, fs reads, path validation, parallel example loading, and build-time copy — is correct and a clear improvement.

src/routes/docs/references/[version]/[platform]/[service]/specs.ts — specifically the apiCache key on line 414.

Important Files Changed

Filename	Overview
src/routes/docs/references/[version]/[platform]/[service]/specs.ts	Replaces import.meta.glob with fs.readFile; adds path allowlist guards, an in-memory apiCache, and parallelised example loading — but apiCache keyed on full platform string causes redundant parses for same-mode platforms (P1).
scripts/build.js	New post-build step that copies specs data into build/_specs_data so the deployment artifact is self-contained; logic is straightforward and correct.
package.json	Moves @appwrite.io/specs from devDependencies to dependencies and removes vite-plugin-dynamic-import.
vite.config.ts	Removes the now-unneeded dynamicImport plugin; no other changes.

_{Reviews (4): Last reviewed commit: "move specs out of node_modules so doesnt..." | Re-trigger Greptile}

[greptile-apps]

The apiCache key includes the full platform string (e.g. server-nodejs, server-python), but the JSON file actually read is determined by mode — one of three fixed literals (server, client, console). Every server-side platform for the same version maps to the same open-api3-${version}-server.json file, so each distinct platform string creates a separate cache entry that parses and stores an identical multi-MB document. On a live SSR server where different users request server-nodejs, server-python, server-php, etc. concurrently, the same JSON is parsed and allocated once per platform variant — partially undoing the memory benefit of the cache. The key should be ${version}|${mode} so all server platforms share one entry.

Suggested change

	const cacheKey = `${version}|${platform}`;
	const cached = apiCache.get(cacheKey);
	if (cached) {
	return cached;
	}
	const isClient = platform.startsWith('client-');
	const mode = platform.startsWith('server-') ? 'server' : isClient ? 'client' : 'console';
	const filename = `open-api3-${version}-${mode}.json`;
	const isClient = platform.startsWith('client-');
	const mode = platform.startsWith('server-') ? 'server' : isClient ? 'client' : 'console';
	const cacheKey = `${version}|${mode}`;
	const cached = apiCache.get(cacheKey);
	if (cached) {
	return cached;
	}
	const filename = `open-api3-${version}-${mode}.json`;