Download File

import * as wmill from "windmill-client"

import { S3Object } from "windmill-client"

const graph = "https://graph.microsoft.com/v1.0"

export type DynSelect_site_id = string

export async function site_id(auth: RT.MsSharepoint) {

  const response = await fetch(

    `${graph}/sites?search=*&$select=id,displayName,webUrl&$top=200`,

    {

      headers: {

        Authorization: `Bearer ${auth.token}`,

        Accept: "application/json",

      },

    }

  )

  if (!response.ok) {

    throw new Error(`${response.status} ${await response.text()}`)

  }

  const { value } = (await response.json()) as {

    value: { id: string; displayName: string; webUrl: string }[]

  }

  return value.map((s) => ({

    value: s.id,

    label: `${s.displayName} (${s.webUrl})`,

  }))

}

export type DynSelect_drive_id = string

export async function drive_id(auth: RT.MsSharepoint, site_id: string) {

  const response = await fetch(

    `${graph}/sites/${site_id}/drives?$select=id,name`,

    {

      headers: {

        Authorization: `Bearer ${auth.token}`,

        Accept: "application/json",

      },

    }

  )

  if (!response.ok) {

    throw new Error(`${response.status} ${await response.text()}`)

  }

  const { value } = (await response.json()) as {

    value: { id: string; name: string }[]

  }

  return value.map((d) => ({ value: d.id, label: d.name }))

}

// The site's default document library, or the library given by drive_id

function driveUrl(site_id: string, drive_id: string | undefined) {

  return drive_id != null && drive_id !== ""

    ? `${graph}/drives/${encodeURIComponent(drive_id)}`

    : `${graph}/sites/${site_id}/drive`

}

// An item is addressed by its id, or by a path from the drive root starting with "/"

function ref(item: string | undefined) {

  if (item == null || item === "" || item === "/") return "root"

  if (!item.startsWith("/")) return `items/${encodeURIComponent(item)}`

  const path = item

    .replace(/\/+$/, "")

    .split("/")

    .map(encodeURIComponent)

    .join("/")

  return `root:${path}:`

}

/**

 * Download File

 * Download a file in a site's default document library, or in another of its libraries when `drive_id` is set to Windmill object storage and return its S3 reference. `item` is an id, or a path from the drive root starting with "/" (e.g. `/Reports/2026/summary.xlsx`). `convert_to_pdf` converts Office documents (docx, xlsx, pptx and others) to PDF on the way out. Requires Sites.Read.All (covered by the default Sites.ReadWrite.All).

 */

export async function main(

  auth: RT.MsSharepoint,

  site_id: DynSelect_site_id,

  item: string,

  destination: S3Object,

  convert_to_pdf: boolean = false,

  drive_id: DynSelect_drive_id | undefined

) {

  const drive = driveUrl(site_id, drive_id)

  const url = new URL(`${drive}/${ref(item)}/content`)

  if (convert_to_pdf) url.searchParams.append("format", "pdf")

  // Graph answers 302 to a pre-authenticated download URL, which fetch follows

  const response = await fetch(url, {

    headers: { Authorization: `Bearer ${auth.token}` },

  })

  if (!response.ok) {

    throw new Error(`${response.status} ${await response.text()}`)

  }

  await wmill.writeS3File(

    destination,

    response.body as ReadableStream<Uint8Array>

  )

  return destination

}