Get base64 encoding of an attachment

Ask for Help
2

True, this is hard to use in the renderhtml (HTML Viewer) widget. It is possible already with some effort, and I put an example here:

https://public.getgrist.com/mSxCywdTTSJa/Attachments-in-HTML-Viewer/p/1#a1.s1.r1.c2

  1. Firstly, since you need to run JS, there is a hidden way to allow that in the renderhtml widget. Use the widget by choosing Custom URL option, and use the URL of https://gristlabs.github.io/grist-widget/renderhtml?tags=script. (The tags=script option allows <script> tags in the HTML to be rendered.)

  2. Second is you need a fairly sophisticated formula that combines Python and Javascript. The one in the example is this:

    setImageSrcScript = """
<script>
  grist.docApi.getAccessToken({readOnly: true})
  .then(tokenInfo => {
    for (let img of document.querySelectorAll('img')) {
      if (img.dataset.att) {
        img.src = `${tokenInfo.baseUrl}/attachments/${img.dataset.att}/download?auth=${tokenInfo.token}`;
      }
    }
  })
</script>
"""

return """
<div style="padding: 24px; background-color: #F4F0FF">
{setImageSrcScript}
<h1>{Company}</h1>
<p>Hello world!
{att}
</div>
""".format(
  setImageSrcScript=setImageSrcScript,
  Company=$Company,
  att='\n'.join('<p><img data-att="{}" width="100%">'.format(att.id) for att in $Attachments)
)

The setImageSrcScript part is meant to be reusable with no changes in other such image-rendering formulas; it goes through all <img> tags with data-att attribute equal to the id of an attachment, and adds a suitable src tag for those.

As for a better approach, the problem with using base64 encoding in a formula to compute HTML that includes the full body of attachments is that it can produce very large formula results. Grist doesn’t handle those very efficiently. For a large table, it could quickly become a problem. I think a more promising approach would be an improvement to the HTML Viewer widget that automatically does something similar to what’s done by the setImageSrcScript above.

1 Like