Get Document as Text

//native

/**

 * Get Document as Text

 * Get a Google Doc's title and its body as plain text (paragraphs and table cells, in reading order). Reads the first tab unless `tab_id` (from List Tabs) names another one.

 */

export async function main(

  auth: RT.Gdocs,

  document_id: string,

  tab_id: string | undefined

) {

  const url = new URL(`https://docs.googleapis.com/v1/documents/${document_id}`)

  if (tab_id != null && tab_id !== "")

    url.searchParams.append("includeTabsContent", "true")

  const response = await fetch(url, {

    headers: {

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

      Accept: "application/json",

    },

  })

  if (!response.ok) {

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

  }

  const doc = await response.json()

  let content = doc.body?.content

  if (tab_id) {

    const tab = findTab(doc.tabs ?? [], tab_id)

    if (!tab) throw new Error(`No tab ${tab_id} in document ${document_id}`)

    content = tab.documentTab?.body?.content

  }

  return {

    document_id: doc.documentId,

    title: doc.title,

    text: toText(content ?? []),

  }

}

function findTab(tabs: any[], id: string): any {

  for (const t of tabs) {

    if (t.tabProperties?.tabId === id) return t

    const child = findTab(t.childTabs ?? [], id)

    if (child) return child

  }

}

function toText(content: any[]): string {

  let text = ""

  for (const el of content) {

    for (const e of el.paragraph?.elements ?? [])

      text += e.textRun?.content ?? ""

    for (const row of el.table?.tableRows ?? [])

      for (const cell of row.tableCells ?? [])

        text += toText(cell.content ?? [])

    if (el.tableOfContents) text += toText(el.tableOfContents.content ?? [])

  }

  return text

}