fix(knowledge): align document tag provenance selections with the serialized request (#6332)

* fix(knowledge): align document tag provenance selections with the serialized request

The create/upsert document tools counted one provenance selection pair per
parseDocumentTags entry, while the write route built targets from the
serialized documentTagsData and dropped entries whose value is the empty
string. A tag value that is truthy before coercion but stringifies to empty
(`[]`, `[null]`, `{ toString: () => '' }`) was therefore counted by the tool
and not by the route, and the bundle length check rejected the write with 400.

Both sides now read one shared parser over the exact bytes that go on the
wire, so their counts cannot diverge.

* test(knowledge): use as const for the empty-stringifying tag fixture

Author: waleedlatif1

apps/sim/app/api/knowledge/secret-provenance.ts

@@ -19,13 +19,13 @@ import {
   importKnowledgePersistedResponseSecretProvenance,
   type KnowledgeDocumentSourceValue,
   type KnowledgeDocumentWriteSecretProvenance,
-  parseKnowledgeDocumentTagProvenanceTargets,
 } from '@/lib/knowledge/secret-provenance'
 import {
   knowledgeDocumentContentSelectionKey,
   knowledgeDocumentFilenameSelectionKey,
   knowledgeDocumentTagNameSelectionKey,
   knowledgeDocumentTagValueSelectionKey,
+  parseKnowledgeDocumentTagProvenanceTargets,
 } from '@/lib/knowledge/secret-provenance-selection'
 import { ResolvedSecretTraceRegistry } from '@/executor/utils/resolved-secret-trace-registry'
 

apps/sim/lib/knowledge/secret-provenance-selection.ts

@@ -1,3 +1,34 @@
+export interface KnowledgeDocumentTagProvenanceTarget {
+  tagName: string
+  value: unknown
+}
+
+/**
+ * Parses only tag entries that can causally contribute a persisted tag value. Both the tool that
+ * builds the request selections and the route that builds the write targets read this one parser,
+ * so their counts can never diverge.
+ */
+export function parseKnowledgeDocumentTagProvenanceTargets(
+  documentTagsData: string | undefined
+): KnowledgeDocumentTagProvenanceTarget[] {
+  if (!documentTagsData) return []
+  try {
+    const parsed: unknown = JSON.parse(documentTagsData)
+    if (!Array.isArray(parsed)) return []
+    return parsed.flatMap((candidate) => {
+      if (!candidate || typeof candidate !== 'object' || Array.isArray(candidate)) return []
+      const record = candidate as Record<string, unknown>
+      const tagName = typeof record.tagName === 'string' ? record.tagName.trim() : ''
+      if (!tagName || record.value === undefined || record.value === null || record.value === '') {
+        return []
+      }
+      return [{ tagName, value: record.value }]
+    })
+  } catch {
+    return []
+  }
+}
+
 export function knowledgeDocumentFilenameSelectionKey(documentIndex: number): string {
   return `document-filename:${documentIndex}`
 }

apps/sim/lib/knowledge/secret-provenance.ts

@@ -56,33 +56,6 @@ export interface KnowledgeDocumentWriteSecretProvenance {
   }[]
 }
 
-interface KnowledgeDocumentTagProvenanceTarget {
-  tagName: string
-  value: unknown
-}
-
-/** Parses only tag entries that can causally contribute a persisted tag value. */
-export function parseKnowledgeDocumentTagProvenanceTargets(
-  documentTagsData: string | undefined
-): KnowledgeDocumentTagProvenanceTarget[] {
-  if (!documentTagsData) return []
-  try {
-    const parsed: unknown = JSON.parse(documentTagsData)
-    if (!Array.isArray(parsed)) return []
-    return parsed.flatMap((candidate) => {
-      if (!candidate || typeof candidate !== 'object' || Array.isArray(candidate)) return []
-      const record = candidate as Record<string, unknown>
-      const tagName = typeof record.tagName === 'string' ? record.tagName.trim() : ''
-      if (!tagName || record.value === undefined || record.value === null || record.value === '') {
-        return []
-      }
-      return [{ tagName, value: record.value }]
-    })
-  } catch {
-    return []
-  }
-}
-
 export type KnowledgeDocumentMetadataField = Exclude<
   keyof KnowledgeDocumentSourceValue,
   'fileUrl' | 'contentHash'

apps/sim/tools/knowledge/secret-provenance.test.ts

@@ -0,0 +1,66 @@
+/**
+ * @vitest-environment node
+ */
+import { describe, expect, it } from 'vitest'
+import {
+  knowledgeDocumentContentSelectionKey,
+  knowledgeDocumentFilenameSelectionKey,
+  knowledgeDocumentTagNameSelectionKey,
+  knowledgeDocumentTagValueSelectionKey,
+  parseKnowledgeDocumentTagProvenanceTargets,
+} from '@/lib/knowledge/secret-provenance-selection'
+import { selectKnowledgeDocumentWriteSecretProvenance } from '@/tools/knowledge/secret-provenance'
+import { formatDocumentTagsForAPI, parseDocumentTags } from '@/tools/shared/tags'
+
+/** Mirrors the selection keys the knowledge write route derives from the serialized request body. */
+function serverSelectionKeys(documentTags: unknown): string[] {
+  const { documentTagsData } = formatDocumentTagsForAPI(parseDocumentTags(documentTags))
+  return [
+    knowledgeDocumentFilenameSelectionKey(0),
+    knowledgeDocumentContentSelectionKey(0),
+    ...parseKnowledgeDocumentTagProvenanceTargets(documentTagsData).flatMap((_tag, tagIndex) => [
+      knowledgeDocumentTagNameSelectionKey(0, tagIndex),
+      knowledgeDocumentTagValueSelectionKey(0, tagIndex),
+    ]),
+  ]
+}
+
+const EMPTY_STRINGIFYING_TAG_VALUES = [
+  ['empty array', []],
+  ['array of null', [null]],
+  ['array of undefined', [undefined]],
+  ['object stringifying to empty', { toString: () => '' }],
+] as const
+
+describe('selectKnowledgeDocumentWriteSecretProvenance', () => {
+  it.each(EMPTY_STRINGIFYING_TAG_VALUES)(
+    'agrees with the server target count for a tag value that is a %s',
+    (_label, tagValue) => {
+      const documentTags = [
+        { tagName: 'kept', value: 'value' },
+        { tagName: 'dropped', value: tagValue },
+      ]
+      const selections = selectKnowledgeDocumentWriteSecretProvenance({
+        name: 'doc.md',
+        content: 'content',
+        documentTags,
+      })
+
+      expect(selections.map((selection) => selection.key)).toEqual(
+        serverSelectionKeys(documentTags)
+      )
+    }
+  )
+
+  it('keeps tags whose serialized value is non-empty', () => {
+    const documentTags = { alpha: 'one', beta: 2, gamma: false }
+    const selections = selectKnowledgeDocumentWriteSecretProvenance({
+      name: 'doc.md',
+      content: 'content',
+      documentTags,
+    })
+
+    expect(selections.map((selection) => selection.key)).toEqual(serverSelectionKeys(documentTags))
+    expect(selections).toHaveLength(8)
+  })
+})

apps/sim/tools/knowledge/secret-provenance.ts

@@ -4,9 +4,10 @@ import {
   knowledgeDocumentFilenameSelectionKey,
   knowledgeDocumentTagNameSelectionKey,
   knowledgeDocumentTagValueSelectionKey,
+  parseKnowledgeDocumentTagProvenanceTargets,
 } from '@/lib/knowledge/secret-provenance-selection'
 import { inferDocumentFileInfo } from '@/tools/knowledge/types'
-import { parseDocumentTags } from '@/tools/shared/tags'
+import { formatDocumentTagsForAPI, parseDocumentTags } from '@/tools/shared/tags'
 
 /** Selects each causally independent persisted document field before request serialization. */
 export function selectKnowledgeDocumentWriteSecretProvenance(params: {
@@ -17,7 +18,9 @@ export function selectKnowledgeDocumentWriteSecretProvenance(params: {
   const name = typeof params.name === 'string' ? params.name.trim() : ''
   const content = typeof params.content === 'string' ? params.content.trim() : params.content
   const filename = inferDocumentFileInfo(name).filename
-  const tags = parseDocumentTags(params.documentTags)
+  const tags = parseKnowledgeDocumentTagProvenanceTargets(
+    formatDocumentTagsForAPI(parseDocumentTags(params.documentTags)).documentTagsData
+  )
 
   return [
     { key: knowledgeDocumentFilenameSelectionKey(0), value: filename },