Function createDocumentRecords

apps/sim/lib/knowledge/documents/service.ts:854–987

(
  documents: Array<{
    filename: string
    fileUrl: string
    fileSize: number
    mimeType: string
    documentTagsData?: string
    tag1?: string
    tag2?: string
    tag3?: string
    tag4?: string
    tag5?: string
    tag6?: string
    tag7?: string
  }>,
  knowledgeBaseId: string,
  requestId: string,
  uploadedBy: string | null = null
)

}

export async function createDocumentRecords(
  documents: Array<{
    filename: string
    fileUrl: string
    fileSize: number
    mimeType: string
    documentTagsData?: string
    tag1?: string
    tag2?: string
    tag3?: string
    tag4?: string
    tag5?: string
    tag6?: string
    tag7?: string
  }>,
  knowledgeBaseId: string,
  requestId: string,
  uploadedBy: string | null = null
): Promise<DocumentData[]> {
  return await db.transaction(async (tx) => {
    await tx.execute(sql`SELECT 1 FROM knowledge_base WHERE id = ${knowledgeBaseId} FOR UPDATE`)

    const kb = await tx
      .select({
        id: knowledgeBase.id,
        workspaceId: knowledgeBase.workspaceId,
        userId: knowledgeBase.userId,
      })
      .from(knowledgeBase)
      .where(and(eq(knowledgeBase.id, knowledgeBaseId), isNull(knowledgeBase.deletedAt)))
      .limit(1)

    if (kb.length === 0) {
      throw new Error('Knowledge base not found')
    }

    const kbWorkspaceId = kb[0].workspaceId
    await assertKnowledgeBaseFileUrlsOwnership(
      documents.map((docData) => docData.fileUrl),
      kbWorkspaceId,
      kb[0].userId,
      requestId,
      tx
    )

    // One load per batch (was N+1); skip entirely if no doc carries tags.
    const hasTaggedDocs = documents.some((d) => d.documentTagsData)
    const tagDefinitions = hasTaggedDocs
      ? await loadTagDefinitions(knowledgeBaseId, tx)
      : (new Map() as TagDefinitionsByName)

    const now = new Date()
    const documentRecords = []
    const returnData: DocumentData[] = []

    for (const docData of documents) {
      const documentId = generateId()