{
  "@context":[
    "https://www.w3.org/ns/activitystreams",
    {"Hashtag":"as:Hashtag","sensitive":"as:sensitive"},
    {"toot":"http://joinmastodon.org/ns#","focalPoint":"toot:focalPoint"}
  ],
  "published":"2025-11-06T16:14:30.492Z",
  "attributedTo":"https://epiktistes.com/actors/toddsundsted",
  "inReplyTo":"https://epiktistes.com/objects/0TxHnMksF5I",
  "replies":"https://epiktistes.com/objects/8Om8QWJvTeE/replies",
  "to":["https://www.w3.org/ns/activitystreams#Public","https://epiktistes.com/actors/toddsundsted"],
  "cc":["https://epiktistes.com/actors/toddsundsted/followers"],
  "content":"<p>I ended up with a solution that exaggerates the focal point offset from center to try to get more of the image around the focal point inside the container:</p><pre>def normalized_focal_point\n  return nil unless has_focal_point?\n  x, y = focal_point.not_nil!\n  norm_x = x / 2 + 0.5      # normalized x = x / 2 + 0.5\n  norm_y = -y / 2 + 0.5     # normalized y = -y / 2 + 0.5 (y inverted)\n  # push the focal point toward the edges so that more of the focused image is in view\n  {\n    exaggerate(norm_x),\n    exaggerate(norm_y)\n  }\nend\n\nprivate def exaggerate(value, strength = 0.75)\n  centered = value - 0.5\n  exaggerated = centered.sign * (centered.abs ** strength)\n  (exaggerated + 0.5).clamp(0.0, 1.0)\nend</pre><p>Reading up on <code>object-position</code> and its use for focal point support in Mastodon and looking at examples in practice it doesn’t seem like the existing implemented approach works well (<a href=\"https://github.com/mastodon/mastodon/issues/6879\">link</a>). I'm open to better ways to do this, and I welcome course correction if I'm heading in the wrong direction.</p><p><a href=\"https://epiktistes.com/tags/ktistec\" class=\"hashtag\" rel=\"tag\">#ktistec</a> <a href=\"https://epiktistes.com/tags/activitypub\" class=\"hashtag\" rel=\"tag\">#activitypub</a></p>",
  "contentMap":{
    "en-US":"<p>I ended up with a solution that exaggerates the focal point offset from center to try to get more of the image around the focal point inside the container:</p><pre>def normalized_focal_point\n  return nil unless has_focal_point?\n  x, y = focal_point.not_nil!\n  norm_x = x / 2 + 0.5      # normalized x = x / 2 + 0.5\n  norm_y = -y / 2 + 0.5     # normalized y = -y / 2 + 0.5 (y inverted)\n  # push the focal point toward the edges so that more of the focused image is in view\n  {\n    exaggerate(norm_x),\n    exaggerate(norm_y)\n  }\nend\n\nprivate def exaggerate(value, strength = 0.75)\n  centered = value - 0.5\n  exaggerated = centered.sign * (centered.abs ** strength)\n  (exaggerated + 0.5).clamp(0.0, 1.0)\nend</pre><p>Reading up on <code>object-position</code> and its use for focal point support in Mastodon and looking at examples in practice it doesn’t seem like the existing implemented approach works well (<a href=\"https://github.com/mastodon/mastodon/issues/6879\">link</a>). I'm open to better ways to do this, and I welcome course correction if I'm heading in the wrong direction.</p><p><a href=\"https://epiktistes.com/tags/ktistec\" class=\"hashtag\" rel=\"tag\">#ktistec</a> <a href=\"https://epiktistes.com/tags/activitypub\" class=\"hashtag\" rel=\"tag\">#activitypub</a></p>"
  },
  "mediaType":"text/html",
  "attachment":[],
  "tag":[
      {"type":"Hashtag","name":"#ktistec","href":"https://epiktistes.com/tags/ktistec"},
      {"type":"Hashtag","name":"#activitypub","href":"https://epiktistes.com/tags/activitypub"}
  ],
  "type":"Note",
  "id":"https://epiktistes.com/objects/yfrZIZfDszU"
}