Proper data cache tagging for computed content (#2915)

.changeset/real-elephants-lie.md

@@ -0,0 +1,6 @@
+---
+'gitbook-v2': patch
+'gitbook': patch
+---
+
+Improving data cache management for computed content

.changeset/two-fans-joke.md

@@ -0,0 +1,5 @@
+---
+'@gitbook/cache-tags': minor
+---
+
+Initial version of the package

bun.lock

@@ -22,6 +22,17 @@
         "wrangler": "^3.109.2",
       },
     },
+    "packages/cache-tags": {
+      "name": "@gitbook/cache-tags",
+      "version": "0.0.0",
+      "dependencies": {
+        "@gitbook/api": "0.96.1",
+        "assert-never": "^1.2.1",
+      },
+      "devDependencies": {
+        "typescript": "^5.5.3",
+      },
+    },
     "packages/colors": {
       "name": "@gitbook/colors",
       "version": "0.2.0",
@@ -42,6 +53,7 @@
       "dependencies": {
         "@gitbook/api": "0.96.1",
         "@gitbook/cache-do": "workspace:*",
+        "@gitbook/cache-tags": "workspace:*",
         "@gitbook/colors": "workspace:*",
         "@gitbook/emoji-codepoints": "workspace:*",
         "@gitbook/icons": "workspace:*",
@@ -122,6 +134,7 @@
       "version": "0.1.1",
       "dependencies": {
         "@gitbook/api": "0.96.1",
+        "@gitbook/cache-tags": "workspace:*",
         "@sindresorhus/fnv1a": "^3.1.0",
         "next": "^15.2.0",
         "react": "^19.0.0",
@@ -618,6 +631,8 @@
 
     "@gitbook/cache-do": ["@gitbook/cache-do@workspace:packages/cache-do"],
 
+    "@gitbook/cache-tags": ["@gitbook/cache-tags@workspace:packages/cache-tags"],
+
     "@gitbook/colors": ["@gitbook/colors@workspace:packages/colors"],
 
     "@gitbook/emoji-codepoints": ["@gitbook/emoji-codepoints@workspace:packages/emoji-codepoints"],

packages/cache-tags/.gitignore

@@ -0,0 +1 @@
+dist/

packages/cache-tags/README.md

@@ -0,0 +1,3 @@
+# `@gitbook/cache-tags`
+
+Utility to generate cache tags for GitBook Open.

packages/cache-tags/package.json

@@ -0,0 +1,25 @@
+{
+    "name": "@gitbook/cache-tags",
+    "type": "module",
+    "exports": {
+        ".": {
+            "types": "./dist/index.d.ts",
+            "development": "./src/index.ts",
+            "default": "./dist/index.js"
+        }
+    },
+    "version": "0.0.0",
+    "dependencies": {
+        "@gitbook/api": "0.96.1",
+        "assert-never": "^1.2.1"
+    },
+    "devDependencies": {
+        "typescript": "^5.5.3"
+    },
+    "scripts": {
+        "build": "tsc",
+        "typecheck": "tsc --noEmit",
+        "dev": "tsc -w"
+    },
+    "files": ["dist", "src", "README.md", "CHANGELOG.md"]
+}

packages/cache-tags/src/index.ts

@@ -0,0 +1,175 @@
+import type { ComputedContentSource } from '@gitbook/api';
+import assertNever from 'assert-never';
+
+/**
+ * Get a stringified cache tag for a given object.
+ */
+export function getCacheTag(
+    spec: /**
+     * All data related to a user
+     * @deprecated - in v2, no tag as this is an immutable data
+     */
+        | {
+              tag: 'user';
+              user: string;
+          }
+        /**
+         * All data related to a space
+         */
+        | {
+              tag: 'space';
+              space: string;
+          }
+        /**
+         * All data related to an integration.
+         */
+        | {
+              tag: 'integration';
+              integration: string;
+          }
+        /**
+         * All data related to a change request
+         */
+        | {
+              tag: 'change-request';
+              space: string;
+              changeRequest: string;
+          }
+        /**
+         * Immutable data related to a revision
+         * @deprecated - in v2, no tag as this is an immutable data
+         */
+        | {
+              tag: 'revision';
+              space: string;
+              revision: string;
+          }
+        /**
+         * Immutable data related to a document
+         * @deprecated - in v2, no tag as this is an immutable data
+         */
+        | {
+              tag: 'document';
+              space: string;
+              document: string;
+          }
+        /**
+         * Immutable data related to a computed document
+         * @deprecated - in v2, no tag as this is an immutable data
+         */
+        | {
+              tag: 'computed-document';
+              space: string;
+              integration: string;
+          }
+        /**
+         * All data related to the URL of a content
+         */
+        | {
+              tag: 'url';
+              hostname: string;
+          }
+        /**
+         * All data related to a site
+         */
+        | {
+              tag: 'site';
+              site: string;
+          }
+        /**
+         * All data related to an OpenAPI spec
+         */
+        | {
+              tag: 'openapi';
+              organization: string;
+              openAPISpec: string;
+          }
+): string {
+    switch (spec.tag) {
+        case 'user':
+            return `user:${spec.user}`;
+        case 'url':
+            return `url:${spec.hostname}`;
+        case 'space':
+            return `space:${spec.space}`;
+        case 'change-request':
+            return `space:${spec.space}:change-request:${spec.changeRequest}`;
+        case 'revision':
+            return `space:${spec.space}:revision:${spec.revision}`;
+        case 'document':
+            return `space:${spec.space}:document:${spec.document}`;
+        case 'computed-document':
+            return `space:${spec.space}:computed-document:${spec.integration}`;
+        case 'site':
+            return `site:${spec.site}`;
+        case 'integration':
+            return `integration:${spec.integration}`;
+        case 'openapi':
+            return `organization:${spec.organization}:openapi:${spec.openAPISpec}`;
+        default:
+            assertNever(spec);
+    }
+}
+
+/**
+ * Get the tags for a computed content source.
+ */
+export function getComputedContentSourceCacheTags(
+    inContext: {
+        spaceId: string;
+        organizationId: string;
+    },
+    source: ComputedContentSource
+) {
+    const tags: string[] = [];
+
+    // We add the dependencies as tags, to ensure that the computed content is invalidated
+    // when the dependencies are updated.
+    const dependencies = Object.values(source.dependencies ?? {});
+    if (dependencies.length > 0) {
+        dependencies.forEach((dependency) => {
+            switch (dependency.ref.kind) {
+                case 'space':
+                    tags.push(
+                        getCacheTag({
+                            tag: 'space',
+                            space: dependency.ref.space,
+                        })
+                    );
+                    break;
+                case 'openapi':
+                    tags.push(
+                        getCacheTag({
+                            tag: 'openapi',
+                            organization: inContext.organizationId,
+                            openAPISpec: dependency.ref.spec,
+                        })
+                    );
+                    break;
+                default:
+                    // Do not throw for unknown dependency types
+                    // as it might mean we are lacking behind the API version
+                    break;
+            }
+        });
+    } else {
+        // Push a dummy tag, as the v1 is only using the first tag
+        tags.push(
+            getCacheTag({
+                tag: 'computed-document',
+                space: inContext.spaceId,
+                integration: source.integration,
+            })
+        );
+    }
+
+    // We invalidate the computed content when a new version of the integration is deployed.
+    tags.push(
+        getCacheTag({
+            tag: 'integration',
+            integration: source.integration,
+        })
+    );
+
+    return tags;
+}

packages/cache-tags/tsconfig.json

@@ -0,0 +1,24 @@
+{
+    "compilerOptions": {
+        "target": "esnext",
+        "lib": ["dom", "dom.iterable", "esnext"],
+        "allowJs": true,
+        "skipLibCheck": true,
+        "strict": true,
+        "noEmit": false,
+        "declaration": true,
+        "outDir": "dist",
+        "esModuleInterop": true,
+        "module": "esnext",
+        "moduleResolution": "bundler",
+        "resolveJsonModule": true,
+        "isolatedModules": true,
+        "jsx": "react-jsx",
+        "incremental": true,
+        "types": [
+            "bun-types" // add Bun global
+        ]
+    },
+    "include": ["src/**/*.ts", "src/**/*.tsx"],
+    "exclude": ["node_modules"]
+}

packages/gitbook-v2/package.json

@@ -7,6 +7,7 @@
         "react": "^19.0.0",
         "react-dom": "^19.0.0",
         "@gitbook/api": "0.96.1",
+        "@gitbook/cache-tags": "workspace:*",
         "@sindresorhus/fnv1a": "^3.1.0",
         "server-only": "^0.0.1"
     },

packages/gitbook-v2/src/app/sites/static/[mode]/[siteURL]/[pagePath]/page.tsx

@@ -3,8 +3,8 @@ import {
     generateSitePageMetadata,
     generateSitePageViewport,
 } from '@/components/SitePage';
+import { getCacheTag } from '@gitbook/cache-tags';
 import { type RouteParams, getPagePathFromParams, getStaticSiteContext } from '@v2/app/utils';
-import { getSiteCacheTag } from '@v2/lib/cache';
 import type { Metadata, Viewport } from 'next';
 import { unstable_cacheTag as cacheTag } from 'next/cache';
 
@@ -21,7 +21,12 @@ export default async function Page(props: PageProps) {
     const context = await getStaticSiteContext(params);
     const pathname = getPagePathFromParams(params);
 
-    cacheTag(getSiteCacheTag(context.site.id));
+    cacheTag(
+        getCacheTag({
+            tag: 'site',
+            site: context.site.id,
+        })
+    );
 
     return <SitePage context={context} pageParams={{ pathname }} redirectOnFallback={true} />;
 }

packages/gitbook-v2/src/app/sites/static/[mode]/[siteURL]/layout.tsx

@@ -4,8 +4,8 @@ import {
     generateSiteLayoutMetadata,
     generateSiteLayoutViewport,
 } from '@/components/SiteLayout';
+import { getCacheTag } from '@gitbook/cache-tags';
 import { type RouteLayoutParams, getStaticSiteContext } from '@v2/app/utils';
-import { getSiteCacheTag } from '@v2/lib/cache';
 import { GITBOOK_DISABLE_TRACKING } from '@v2/lib/env';
 import { unstable_cacheTag as cacheTag } from 'next/cache';
 
@@ -21,7 +21,12 @@ export default async function SiteStaticLayout({
 
     const context = await getStaticSiteContext(await params);
 
-    cacheTag(getSiteCacheTag(context.site.id));
+    cacheTag(
+        getCacheTag({
+            tag: 'site',
+            site: context.site.id,
+        })
+    );
 
     return (
         <CustomizationRootLayout customization={context.customization}>