fix(search): use search-response spaces for verification and fall back for legacy blog-post references

Author: waleedlatif1

apps/sim/lib/sim-search/live/atlassian.test.ts

@@ -1,5 +1,6 @@
 import { describe, expect, it, vi } from 'vitest'
 import { readAtlassian, searchAtlassian } from '@/lib/sim-search/live/atlassian'
+import { NativeSearchError } from '@/lib/sim-search/live/http'
 import type { NativeClient } from '@/lib/sim-search/live/types'
 
 const SITE = { id: 'cloud', url: 'https://acme.atlassian.net' }
@@ -50,7 +51,22 @@ describe('Confluence live documents', () => {
         content: expect.stringContaining('Shipped search.'),
       }
     )
-    expect(api.json).toHaveBeenCalledWith(`${v2}/pages/123`, { query: { 'body-format': 'view' } })
+  })
+
+  it('reads a legacy reference without a kind as a blog post when no page has that id', async () => {
+    const api: NativeClient = {
+      json: vi.fn(async (path: string) => {
+        if (path === '/oauth/token/accessible-resources') return [SITE]
+        if (path === `${v2}/blogposts/9`)
+          return { id: '9', title: 'Release notes', body: { view: { value: '<p>Shipped.</p>' } } }
+        throw new NativeSearchError('unavailable', 'Provider request failed (404).')
+      }),
+      text: vi.fn(),
+    }
+    await expect(readAtlassian(api, 'confluence', '9', 'cloud')).resolves.toMatchObject({
+      kind: 'blogpost',
+      content: expect.stringContaining('Shipped.'),
+    })
   })
 
   it('reads a space result as its homepage, keeping the space as the document', async () => {
@@ -71,7 +87,6 @@ describe('Confluence live documents', () => {
       title: 'Engineering',
       content: expect.stringContaining('Team charter.'),
     })
-    expect(api.json).toHaveBeenCalledWith(`${v2}/spaces`, { query: { keys: 'ENG' } })
   })
 
   it('records whether a search result is a page, blog post, or space so its read picks the endpoint', async () => {
@@ -82,6 +97,7 @@ describe('Confluence live documents', () => {
             content: {
               id: '123',
               type: 'page',
+              space: { key: 'ENG' },
               title: 'Runbook',
               _links: { webui: '/spaces/ENG/pages/123' },
             },
@@ -115,5 +131,7 @@ describe('Confluence live documents', () => {
       { id: 'ENG', kind: 'space' },
     ])
     expect(page.documents[2]?.url).toBe('https://acme.atlassian.net/wiki/spaces/ENG')
+    expect(page.documents[0]?.accessMetadata).toEqual({ spaceKey: 'ENG' })
+    expect(page.documents[2]?.accessMetadata).toEqual({ spaceKey: 'ENG' })
   })
 })

apps/sim/lib/sim-search/live/atlassian.ts

@@ -71,6 +71,7 @@ function page(row: Record<string, unknown>, cloudId: string, site: string): Nati
     return {
       id: string(space.key),
       kind: 'space',
+      accessMetadata: { spaceKey: string(space.key) },
       container: cloudId,
       title: string(row.title) || string(space.name),
       url: `${site}/wiki${string(row.url) || `/spaces/${segment(string(space.key))}`}`,
@@ -81,9 +82,11 @@ function page(row: Record<string, unknown>, cloudId: string, site: string): Nati
   const content = Object.keys(object(row.content)).length ? object(row.content) : row
   const links = object(content._links)
   const version = object(content.version)
+  const spaceKey = string(object(content.space).key)
   return {
     id: string(content.id),
     kind: string(content.type) === 'blogpost' ? 'blogpost' : 'page',
+    ...(spaceKey ? { accessMetadata: { spaceKey } } : {}),
     container: cloudId,
     title: string(content.title) || string(row.title),
     url: `${site}/wiki${string(links.webui) || `/pages/${segment(string(content.id))}`}`,
@@ -174,7 +177,7 @@ export async function searchAtlassian(
               order
             ),
             limit: String(input.limit),
-            expand: 'content.version',
+            expand: 'content.version,content.space',
             ...(input.native?.cursor && single ? { cursor: input.native.cursor } : {}),
           },
         })
@@ -248,11 +251,19 @@ async function readConfluence(
     contentId = string(space.homepageId)
     contentKind = 'page'
   }
+  const content = (type: string) =>
+    client.json(`${api}/${type}/${segment(contentId)}`, { query: { 'body-format': 'view' } })
+  /** A reference issued before kinds were recorded may name a blog post; its page read is a 404. */
   const row = object(
-    await client.json(
-      `${api}/${contentKind === 'blogpost' ? 'blogposts' : 'pages'}/${segment(contentId)}`,
-      { query: { 'body-format': 'view' } }
-    )
+    kind === undefined
+      ? await content('pages').catch((error: unknown) => {
+          if (error instanceof NativeSearchError && error.status === 'unavailable') {
+            contentKind = 'blogpost'
+            return content('blogposts')
+          }
+          throw error
+        })
+      : await content(contentKind === 'blogpost' ? 'blogposts' : 'pages')
   )
   const pageTitle = string(row.title)
   return {

apps/sim/lib/sim-search/live/policy.test.ts

@@ -204,6 +204,15 @@ describe('organization search scope enforcement', () => {
       ).toBe(false)
     }
   )
+  it('checks a Confluence search hit by the space its search response named, without requests', async () => {
+    const verify = createPolicyVerifier('confluence', selected(['ENG']), client({}), '')
+    expect(await verify({ id: '123', container: 'site', kind: 'page' }, { spaceKey: 'ENG' })).toBe(
+      true
+    )
+    expect(await verify({ id: '124', container: 'site', kind: 'page' }, { spaceKey: 'HR' })).toBe(
+      false
+    )
+  })
   it('checks Confluence spaces by key and blog posts through their own endpoint', async () => {
     const api = client({
       '/ex/confluence/site/wiki/api/v2/blogposts/9': { id: '9', spaceId: '7' },

apps/sim/lib/sim-search/live/policy.ts

@@ -202,7 +202,8 @@ export function createPolicyVerifier(
       }
       /** v2 reads, like document reads, so the check needs only the granular read scopes. */
       const api = `/ex/confluence/${segment(document.container)}/wiki/api/v2`
-      let spaceKey = document.kind === 'space' ? document.id : ''
+      /** The search response names each hit's space; only reads without that evidence look it up. */
+      let spaceKey = document.kind === 'space' ? document.id : string(providerMetadata?.spaceKey)
       if (!spaceKey) {
         const row = object(
           await json(