fix(search): keep Confluence searchable for colon queries and say when nothing was searched

Confluence rejects a CQL text query holding a colon or backslash with HTTP 400,
even escaped, which dropped it from those searches; both now go as word breaks.
A plain live search that reaches no connected account is reported as partial
with guidance naming where to connect one, and the results view offers the
connection instead of claiming no results.

Author: waleedlatif1

apps/sim/app/workspace/[workspaceId]/home/components/knowledge-search-results/knowledge-search-results.tsx

@@ -1,12 +1,13 @@
 'use client'
 
 import { useEffect, useMemo, useState } from 'react'
-import { Chip } from '@sim/emcn'
+import { Chip, ChipLink } from '@sim/emcn'
 import { useQueryStates } from 'nuqs'
 import { ActivityStatus } from '@/components/ui/activity-status'
 import type { WorkspaceSearchFilters } from '@/lib/api/contracts/knowledge'
 import { useSession } from '@/lib/auth/client'
 import { type ResourceScope, resourceScopeKey } from '@/lib/core/resource-scope'
+import { organizationRoutes } from '@/lib/navigation/paths'
 import { SearchFilters } from '@/app/workspace/[workspaceId]/home/components/knowledge-search-results/search-filters'
 import {
   groupResultsByDocument,
@@ -150,6 +151,25 @@ function LiveSearchResults({
             Search found no results.
           </p>
         )}
+      {!isFetching && data?.live && accounts.length === 0 && (
+        <div className='flex items-center gap-2 px-2 py-2'>
+          <p role='status' className='text-[var(--text-muted)] text-caption'>
+            {filters.source
+              ? 'You haven’t connected this app.'
+              : 'Connect your apps to search them.'}
+          </p>
+          <ChipLink
+            variant='border'
+            href={
+              scope.kind === 'organization'
+                ? organizationRoutes(scope.organizationId).integrations
+                : `/workspace/${scope.workspaceId}/integrations`
+            }
+          >
+            Connect apps
+          </ChipLink>
+        </div>
+      )}
       {accounts.some((account) =>
         ['unavailable', 'timeout', 'rate_limited', 'reconnect'].includes(account.status)
       ) && (

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

@@ -666,6 +666,28 @@ describe('authorized live retrieval', () => {
     expect(result.live?.accounts.map((status) => status.queryIndex)).toEqual([0, 1])
     expect(result.live?.guidance).toBe('Live coverage: slack')
   })
+  it('reports a search that reached no account as nothing searched, not as no results', async () => {
+    mocks.accounts.mockResolvedValue([])
+    const result = await searchLiveKnowledge.execute({ principal, input })
+    expect(mocks.search).not.toHaveBeenCalled()
+    expect(result.results).toEqual([])
+    expect(result.retrieval.status).toBe('partial')
+    expect(result.live?.accounts).toEqual([])
+    expect(result.live?.guidance).toMatch(
+      /^Nothing was searched: you have no connected sources here\. Connect your accounts at \S*\/workspace\/workspace\/integrations,/
+    )
+  })
+  it('names the filtered source when only other sources are connected', async () => {
+    const result = await searchLiveKnowledge.execute({
+      principal,
+      input: { ...input, filters: { source: 'slack' } },
+    })
+    expect(mocks.search).not.toHaveBeenCalled()
+    expect(result.retrieval.status).toBe('partial')
+    expect(result.live?.guidance).toMatch(
+      /^Nothing was searched: you have no connected slack account here\. .* or search without the source filter\./
+    )
+  })
   describe('result previews', () => {
     const excerpts = ['a', 'b', 'c', 'd'].map((id) => ({
       ...document,

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

@@ -32,6 +32,7 @@ import {
   mapWithConcurrency,
   waitWithAbort,
 } from '@/lib/core/utils/concurrency'
+import { getBaseUrl } from '@/lib/core/utils/urls'
 import { MANAGED_MCP_CONNECTORS } from '@/lib/credential-groups/managed-mcp-connectors'
 import { requireOrganizationSearchAvailable } from '@/lib/knowledge/access/availability'
 import { defineAuthorizedKnowledgeUseCase } from '@/lib/knowledge/application/authorized-knowledge-use-case'
@@ -49,6 +50,7 @@ import {
 import { RRF_K } from '@/lib/knowledge/search/recency'
 import { matchPassage } from '@/lib/knowledge/search/snippet'
 import { isKnowledgeSourceUrl } from '@/lib/knowledge/search/source-url'
+import { organizationRoutes } from '@/lib/navigation/paths'
 import { connectorDisplayName } from '@/lib/sim-search/connectors'
 import {
   type LiveAccountSession,
@@ -440,6 +442,24 @@ function compareAccounts(left: LiveAccount, right: LiveAccount): number {
   )
 }
 
+/**
+ * Why a plain search that reached no account returned nothing, and where the caller connects one;
+ * an empty result would otherwise read as a search that found nothing.
+ */
+function nothingSearchedGuidance(
+  owner: ResourceOwner,
+  source: string | undefined,
+  hasOtherAccounts: boolean
+): string {
+  const path = owner.organizationId
+    ? organizationRoutes(owner.organizationId).integrations
+    : `/workspace/${owner.workspaceId}/integrations`
+  const connect = new URL(path, getBaseUrl()).href
+  return source && hasOtherAccounts
+    ? `Nothing was searched: you have no connected ${source} account here. Connect one at ${connect}, or search without the source filter.`
+    : `Nothing was searched: you have no connected sources here. Connect your accounts at ${connect}, then search again.`
+}
+
 const liveKnowledgeSearch = defineAuthorizedKnowledgeUseCase({
   operation: knowledgeOperations.search,
   resolveContext: ({ input }: { input: LiveSearchInput }) => resolveKnowledgeOwnerContext(input),
@@ -503,6 +523,7 @@ const liveKnowledgeSearch = defineAuthorizedKnowledgeUseCase({
           nativesFor(account).length > 0
       )
       .sort(compareAccounts)
+    const nothingSearched = !queries && eligible.length === 0
     const selected = eligible.slice(0, MAX_ACCOUNTS)
     const openingAccount = createConcurrencyLimit(ACCOUNT_OPEN_CONCURRENCY)
     annotateSearchDiagnostics({
@@ -909,9 +930,14 @@ const liveKnowledgeSearch = defineAuthorizedKnowledgeUseCase({
       }
     })
     /** Live coverage guides the providers this call searched or tried to search. */
-    const guidance = liveSearchGuidance(
-      accounts.filter((account) => account.status !== 'reconnect').map(({ provider }) => provider)
-    )
+    const guidance = [
+      ...(nothingSearched
+        ? [nothingSearchedGuidance(input, filters?.source, allAccounts.length > 0)]
+        : []),
+      liveSearchGuidance(
+        accounts.filter((account) => account.status !== 'reconnect').map(({ provider }) => provider)
+      ),
+    ].join('\n')
     for (const account of eligible.slice(MAX_ACCOUNTS)) {
       for (const { queryIndex } of nativesFor(account))
         accounts.push({
@@ -934,9 +960,10 @@ const liveKnowledgeSearch = defineAuthorizedKnowledgeUseCase({
           message: 'No connection with this provider is configured and approved in this scope.',
         })
     }
-    const retrievalStatus = accounts.some((account) => account.status !== 'ok')
-      ? 'partial'
-      : 'complete'
+    const retrievalStatus =
+      nothingSearched || accounts.some((account) => account.status !== 'ok')
+        ? 'partial'
+        : 'complete'
     annotateSearchDiagnostics({
       resultCount: Math.min(ranked.length, input.topK),
       retrievalStatus,

apps/sim/lib/sim-search/live/text-match.test.ts

@@ -59,6 +59,16 @@ describe('text match clauses', () => {
     )
   })
 
+  it('sends a colon or backslash to Confluence as a word break, which it otherwise rejects', () => {
+    expect(confluenceTextClause('Security: Weekly Report', 'all')).toBe(
+      'text ~ "Security  Weekly Report"'
+    )
+    expect(confluenceTextClause(String.raw`C:\builds\out`, 'any')).toBe(
+      'siteSearch ~ "C  builds out"'
+    )
+    expect(confluenceTextClause(String.raw` :\ `, 'all')).toBe('')
+  })
+
   it('falls back to the all-terms clause when fewer than two terms are distinctive', () => {
     expect(jiraTextClause('how does it work', 'any')).toBe('text ~ "how does it work"')
   })

apps/sim/lib/sim-search/live/text-match.ts

@@ -150,9 +150,14 @@ export function jiraTextClause(text: string, match: TextMatch = 'all'): string {
  * The text clause for a plain query in Confluence CQL. `any` uses `siteSearch`, Confluence's own
  * relevance search, which ranks a page with the query's rarer words above the many pages sharing
  * one common word, where a `text ~` OR query ranks them the other way round.
+ *
+ * Confluence answers HTTP 400 to a text query holding a colon or backslash, even escaped, and its
+ * analyzer splits words on both, so they go as spaces. Nothing left to match means no clause.
  */
 export function confluenceTextClause(text: string, match: TextMatch = 'all'): string {
-  return `${match === 'any' ? 'siteSearch' : 'text'} ~ ${escapeSearchPhrase(text)}`
+  const searchable = text.replace(/[:\\]/g, ' ').trim()
+  if (!searchable) return ''
+  return `${match === 'any' ? 'siteSearch' : 'text'} ~ ${escapeSearchPhrase(searchable)}`
 }
 
 /** How many distinct terms occur as whole words in the text. */