fix(files): address search review findings

Author: icecrasher321

apps/docs/content/docs/integrations/file.mdx

@@ -89,7 +89,7 @@ Search indexed text across active workspace files using literal smart-case subst
 |   ↳ `text` | string | Matching line or bounded match-centered preview. |
 | `count` | number | Number of returned matching lines. |
 | `truncated` | boolean | Whether more matching lines exist beyond the configured hard cap. |
-| `complete` | boolean | Whether every current file revision is indexed without failures. |
+| `complete` | boolean | Whether indexing has no pending or failed current revisions; skipped and partial coverage is reported separately. |
 | `indexStatus` | object | Current workspace search-index coverage by file status. |
 |   ↳ `readyFiles` | number | Files whose current revision is searchable. |
 |   ↳ `pendingFiles` | number | Files still waiting to be indexed. |

apps/sim/blocks/blocks.test.ts

@@ -170,6 +170,7 @@ describe.concurrent('Blocks Module', () => {
       expect(block?.subBlocks[0].options?.map((option) => option.id)).toEqual([
         'file_read',
         'file_get_content',
+        'file_search',
         'file_fetch',
         'file_write',
         'file_append',

apps/sim/blocks/blocks/file.test.ts

@@ -83,6 +83,19 @@ describe('FileV5Block', () => {
     expect(maxResults?.value?.()).toBe('50')
   })
 
+  it.each(['10.5', '10results', '0', '201'])(
+    'rejects invalid builder-configured search cap %s',
+    (maxResults) => {
+      expect(() =>
+        buildParams({
+          operation: 'file_search',
+          query: 'needle',
+          maxResults,
+        })
+      ).toThrow('Maximum Results must be an integer between 1 and 200')
+    }
+  )
+
   it('read returns only the files output (no redundant file)', () => {
     expect(FileV5Block.outputs.files).toBeDefined()
     expect(FileV5Block.outputs.contents).toBeDefined()

apps/sim/blocks/blocks/file.ts

@@ -1254,10 +1254,13 @@ export const FileV5Block: BlockConfig<FileParserV3Output> = {
         const operation = params.operation || 'file_read'
 
         if (operation === 'file_search') {
-          const maxResults = Number.parseInt(String(params.maxResults ?? '50'), 10)
+          const maxResults = Number(params.maxResults ?? '50')
+          if (!Number.isInteger(maxResults) || maxResults < 1 || maxResults > 200) {
+            throw new Error('Maximum Results must be an integer between 1 and 200')
+          }
           return {
             query: params.query,
-            maxResults: Number.isNaN(maxResults) ? 50 : maxResults,
+            maxResults,
           }
         }
 

apps/sim/lib/internal/file/execute-tool.ts

@@ -1,6 +1,6 @@
 import { resolvePrincipalAttribution, resolvePrincipalSubject } from '@sim/auth/principal'
 import { createLogger } from '@sim/logger'
-import { getErrorMessage } from '@sim/utils/errors'
+import { getErrorMessage, toError } from '@sim/utils/errors'
 import { z } from 'zod'
 import { fileParseContract } from '@/lib/api/contracts/storage-transfer'
 import { fileManageContract } from '@/lib/api/contracts/tools/file'
@@ -184,6 +184,7 @@ export const executeFileTool: InternalToolOperationHandler = async (request) =>
     const message = getErrorMessage(error, 'Unknown error')
     logger.error('File operation dispatch failed', {
       error: isSearchFailure ? 'Workspace file search failed' : message,
+      errorType: isSearchFailure ? toError(error).name : undefined,
       requestId: request.requestId,
       toolId: request.toolId,
     })

apps/sim/lib/workspace-files/application/operations.test.ts

@@ -44,6 +44,7 @@ describe('file operation registry', () => {
     expect(executorOperationIds).toEqual([
       'files.read_metadata',
       'files.read_content',
+      'files.search_content',
       'files.download',
       'files.create',
       'files.update_content',

apps/sim/lib/workspace-files/search/indexing.ts

@@ -70,17 +70,49 @@ async function extractIndexText(
   }
 }
 
-async function clearRevision(fileId: string, sourceContentUpdatedAt: Date): Promise<void> {
+async function clearRevision(
+  workspaceId: string,
+  fileId: string,
+  sourceContentUpdatedAt: Date
+): Promise<void> {
   await db
     .delete(workspaceFileSearchSegment)
     .where(
       and(
+        eq(workspaceFileSearchSegment.workspaceId, workspaceId),
         eq(workspaceFileSearchSegment.fileId, fileId),
         eq(workspaceFileSearchSegment.sourceContentUpdatedAt, sourceContentUpdatedAt)
       )
     )
 }
 
+async function discardObsoleteRevision(options: {
+  workspaceId: string
+  fileId: string
+  sourceContentUpdatedAt: Date
+}): Promise<void> {
+  await db.transaction(async (tx) => {
+    await tx
+      .delete(workspaceFileSearchSegment)
+      .where(
+        and(
+          eq(workspaceFileSearchSegment.workspaceId, options.workspaceId),
+          eq(workspaceFileSearchSegment.fileId, options.fileId),
+          eq(workspaceFileSearchSegment.sourceContentUpdatedAt, options.sourceContentUpdatedAt)
+        )
+      )
+    await tx
+      .delete(workspaceFileSearchIndex)
+      .where(
+        and(
+          eq(workspaceFileSearchIndex.workspaceId, options.workspaceId),
+          eq(workspaceFileSearchIndex.fileId, options.fileId),
+          eq(workspaceFileSearchIndex.sourceContentUpdatedAt, options.sourceContentUpdatedAt)
+        )
+      )
+  })
+}
+
 async function markTerminal(options: {
   workspaceId: string
   fileId: string
@@ -114,6 +146,7 @@ async function markTerminal(options: {
         .delete(workspaceFileSearchSegment)
         .where(
           and(
+            eq(workspaceFileSearchSegment.workspaceId, options.workspaceId),
             eq(workspaceFileSearchSegment.fileId, options.fileId),
             eq(workspaceFileSearchSegment.sourceContentUpdatedAt, options.sourceContentUpdatedAt)
           )
@@ -122,6 +155,7 @@ async function markTerminal(options: {
         .delete(workspaceFileSearchIndex)
         .where(
           and(
+            eq(workspaceFileSearchIndex.workspaceId, options.workspaceId),
             eq(workspaceFileSearchIndex.fileId, options.fileId),
             eq(workspaceFileSearchIndex.sourceContentUpdatedAt, options.sourceContentUpdatedAt)
           )
@@ -234,12 +268,15 @@ export async function indexWorkspaceFileForSearch(
     throwOnError: true,
   })
   if (!file || !sameRevision(file.contentUpdatedAt, sourceContentUpdatedAt)) {
-    await clearRevision(payload.fileId, sourceContentUpdatedAt)
+    await discardObsoleteRevision({ ...payload, sourceContentUpdatedAt })
     return
   }
 
   const [state] = await db
-    .select({ status: workspaceFileSearchIndex.status })
+    .select({
+      status: workspaceFileSearchIndex.status,
+      workspaceId: workspaceFileSearchIndex.workspaceId,
+    })
     .from(workspaceFileSearchIndex)
     .where(
       and(
@@ -248,6 +285,10 @@ export async function indexWorkspaceFileForSearch(
       )
     )
     .limit(1)
+  if (state && state.workspaceId !== payload.workspaceId) {
+    await discardObsoleteRevision({ ...payload, sourceContentUpdatedAt })
+    return
+  }
   if (state?.status === 'ready' || state?.status === 'skipped') return
 
   await db
@@ -270,7 +311,7 @@ export async function indexWorkspaceFileForSearch(
         updatedAt: new Date(),
       },
     })
-  await clearRevision(payload.fileId, sourceContentUpdatedAt)
+  await clearRevision(payload.workspaceId, payload.fileId, sourceContentUpdatedAt)
 
   if (file.size > FILE_SEARCH_MAX_SOURCE_BYTES) {
     await markTerminal({
@@ -315,7 +356,7 @@ export async function indexWorkspaceFileForSearch(
   } catch (error) {
     if (signal.aborted) throw error
     if (isPayloadSizeLimitError(error)) {
-      await clearRevision(payload.fileId, sourceContentUpdatedAt)
+      await clearRevision(payload.workspaceId, payload.fileId, sourceContentUpdatedAt)
       await markTerminal({
         ...payload,
         sourceContentUpdatedAt,
@@ -324,7 +365,7 @@ export async function indexWorkspaceFileForSearch(
       })
       return
     }
-    await clearRevision(payload.fileId, sourceContentUpdatedAt)
+    await clearRevision(payload.workspaceId, payload.fileId, sourceContentUpdatedAt)
     logger.error('Workspace file search indexing failed', {
       workspaceId: payload.workspaceId,
       fileId: payload.fileId,
@@ -340,7 +381,7 @@ export async function markWorkspaceFileSearchIndexFailed(
 ): Promise<void> {
   const sourceContentUpdatedAt = new Date(payload.sourceContentUpdatedAt)
   if (Number.isNaN(sourceContentUpdatedAt.getTime())) return
-  await clearRevision(payload.fileId, sourceContentUpdatedAt)
+  await clearRevision(payload.workspaceId, payload.fileId, sourceContentUpdatedAt)
   await markTerminal({
     ...payload,
     sourceContentUpdatedAt,

apps/sim/lib/workspace-files/search/text.test.ts

@@ -43,6 +43,14 @@ describe('workspace file search text utilities', () => {
     expect(preview).not.toContain('�')
   })
 
+  it('maps case-folded offsets back to the original line', () => {
+    const line = `${'İ'.repeat(1200)}needle${'x'.repeat(1200)}`
+    const preview = createFileSearchPreview(line, 'needle', false)
+
+    expect(preview).toContain('needle')
+    expect(Buffer.byteLength(preview, 'utf8')).toBeLessThanOrEqual(2048)
+  })
+
   it('shows omitted logical-line content beyond the selected segment', () => {
     expect(
       createFileSearchPreview('needle and nearby text', 'needle', false, 2048, {

apps/sim/lib/workspace-files/search/text.ts

@@ -110,6 +110,42 @@ export function truncateUtf8ToBytes(text: string, maxBytes: number): string {
   return encoded.subarray(0, end).toString('utf8')
 }
 
+function findFileSearchMatchRange(
+  line: string,
+  query: string,
+  caseSensitive: boolean
+): { start: number; end: number } {
+  if (caseSensitive) {
+    const start = Math.max(0, line.indexOf(query))
+    return { start, end: Math.min(line.length, start + query.length) }
+  }
+
+  const searchableLine = line.toLocaleLowerCase()
+  const searchableQuery = query.toLocaleLowerCase()
+  const foldedStart = searchableLine.indexOf(searchableQuery)
+  if (foldedStart < 0) return { start: 0, end: Math.min(line.length, query.length) }
+
+  const originalStarts: number[] = []
+  const originalEnds: number[] = []
+  for (let offset = 0; offset < line.length; ) {
+    const codePoint = line.codePointAt(offset)
+    if (codePoint === undefined) break
+    const character = String.fromCodePoint(codePoint)
+    const foldedCharacter = character.toLocaleLowerCase()
+    const end = offset + character.length
+    for (let foldedOffset = 0; foldedOffset < foldedCharacter.length; foldedOffset += 1) {
+      originalStarts.push(offset)
+      originalEnds.push(end)
+    }
+    offset = end
+  }
+
+  const foldedEnd = foldedStart + searchableQuery.length
+  const start = originalStarts[foldedStart] ?? 0
+  const end = originalEnds[foldedEnd - 1] ?? Math.min(line.length, start + query.length)
+  return { start, end }
+}
+
 export function createFileSearchPreview(
   line: string,
   query: string,
@@ -124,10 +160,7 @@ export function createFileSearchPreview(
     return `${boundaries.prefixOmitted ? '…' : ''}${line}${boundaries.suffixOmitted ? '…' : ''}`
   }
 
-  const searchableLine = caseSensitive ? line : line.toLocaleLowerCase()
-  const searchableQuery = caseSensitive ? query : query.toLocaleLowerCase()
-  const matchStart = Math.max(0, searchableLine.indexOf(searchableQuery))
-  const matchEnd = Math.min(line.length, matchStart + query.length)
+  const { start: matchStart, end: matchEnd } = findFileSearchMatchRange(line, query, caseSensitive)
   const leadingEllipsis = boundaries.prefixOmitted || matchStart > 0 ? '…' : ''
   const trailingEllipsis = boundaries.suffixOmitted || matchEnd < line.length ? '…' : ''
   const ellipsisBytes = Buffer.byteLength(leadingEllipsis + trailingEllipsis, 'utf8')

apps/sim/tools/file/search.ts

@@ -82,7 +82,8 @@ export const fileSearchTool: InternalToolConfig<FileSearchParams, FileSearchResp
     },
     complete: {
       type: 'boolean',
-      description: 'Whether every current file revision is indexed without failures.',
+      description:
+        'Whether indexing has no pending or failed current revisions; skipped and partial coverage is reported separately.',
     },
     indexStatus: {
       type: 'object',