M3 - New ActivityPub Client

[maisutton]

Launch date: May 10th

Develop a new ActivityPub client which will support natively loading published data over p2p protocols, skipping the need for always-online HTTP servers for accessing content.

• New project reader.distributed.press
• Render ActivityPub posts in UI by loading them on the fly (1 week) @akhileshthite
  • Support loading from p2p URLs (native or via gateways)
  • UI to configure gateway
• Read from users outboxes into a feed (0.5 week)
  • Read staticpub inbox to start, have interface that takes a "num posts" and "page" param and has an async iterator of Activities.
  • Handle create/update/
• Ingest posts/outbox into local Database (leveljs over indexeddb)
  • RDF traversal library that works over fetch and resolves links as we traverse (1 week)
  • Basic API for handling AP traversal
  • Index Notes from activities in addition to activities themselves (1 week)
  • Track update times
• Track followed accounts and have a button to download updates (1 week)
  • Have a page to track them
  • Save their IDs in local storage
  • Code to traverse their outboxes and ingest them into the db
• Render "timeline" of posts (maybe we can experiment here) (2 weeks)

[RangerMauve]

• New project reader.distributed.press
• Low javascript, no framework, multi page application using web components, no server component (little or no) build system.
• Render ActivityPub posts in UI by loading them on the fly
• Support loading from p2p URLs (native or via gateways)
• UI to configure gateway
• Read from users outboxes into a feed
• Ingest posts/outbox into local Database (leveljs over indexeddb)
• Render "timeline" of posts (maybe we can experiment here)

Nice to have: Different filter options for timeline (using querystrings and being able to save views)
Nice to have: Basic color scheme customization to make it feel like "their own"

[RangerMauve]

Load this as a post: hyper://staticpub.mauve.moe/helloworld.html

• Fetch from IPFS
• Detect that it's html and get the link[rel=alternate] tag to get the json-ld URL
• Resolve the json-ld object and render it's content

[RangerMauve]

Use https://developer.mozilla.org/en-US/docs/Web/API/DOMParser for parsing the html, then querySelector to get the link

[RangerMauve]

Assume that the IPFS URLs of all the objects will be the same domain but with ipfs://

Test it in Agregore or Peersky or Brave to try direct IPFS loading