Function fileToDocument

apps/sim/connectors/gitlab/gitlab.ts:322–384

* Builds a repository-file document from a fetched (non-raw) file response. Returns * null for binary, oversized, or empty files so they are not indexed.

(
  apiBase: string,
  encodedProject: string,
  host: string,
  projectPath: string,
  ref: string,
  path: string,
  file: GitLabFile
)

 * null for binary, oversized, or empty files so they are not indexed.
 */
function fileToDocument(
  apiBase: string,
  encodedProject: string,
  host: string,
  projectPath: string,
  ref: string,
  path: string,
  file: GitLabFile
): ExternalDocument | null {
  const blobSha = file.blob_id?.trim()
  if (!blobSha) return null

  const title = path.split('/').pop() || path
  const skippedForSize = (size: number): ExternalDocument => {
    logger.info('Skipping oversized GitLab file', { path, size })
    return markSkipped(
      {
        externalId: `${FILE_PREFIX}${path}`,
        title,
        content: '',
        mimeType: 'text/plain',
        sourceUrl: buildFileSourceUrl(apiBase, encodedProject, host, projectPath, ref, path),
        contentHash: buildFileContentHash(encodedProject, path, blobSha),
        metadata: { contentType: 'file', title, path, size },
      },
      sizeLimitSkipReason(MAX_FILE_SIZE)
    )
  }

  if (typeof file.size === 'number' && file.size > MAX_FILE_SIZE) {
    return skippedForSize(file.size)
  }

  const raw = typeof file.content === 'string' ? file.content : ''
  const buffer = file.encoding === 'base64' ? Buffer.from(raw, 'base64') : Buffer.from(raw, 'utf8')
  if (isBinaryBuffer(buffer)) {
    logger.info('Skipping binary GitLab file', { path })
    return null
  }
  if (buffer.byteLength > MAX_FILE_SIZE) {
    return skippedForSize(buffer.byteLength)
  }

  const content = buffer.toString('utf8')
  const body = composeBody(title, content)
  if (!body.trim()) return null

  return {
    externalId: `${FILE_PREFIX}${path}`,
    title,
    content: body,
    contentDeferred: false,
    mimeType: 'text/plain',
    sourceUrl: buildFileSourceUrl(apiBase, encodedProject, host, projectPath, ref, path),
    contentHash: buildFileContentHash(encodedProject, path, blobSha),
    metadata: {
      contentType: 'file',
      title,