All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
- Added link to personal website on About page.
- Allow
data:images in Netlify CSP. - Fix inability to DM login links in releases (#104).
- Fixed crawler not finding API in some cases
- Fix CSP issues for Plausible analytics
- Allow Plausible privacy-preserving analytics in CSP
- Update dependencies
- Update to Elixir 1.10
- Fixed CSP headers for data: images
- Remove staging server
- Added AppSignal logo to "Special thanks" section
- Bring back
developstaging backup (now managed in DNS) - Increase default number of concurrent crawlers to 100
- Accessibility improvements (according to axe review)
- Update dependencies
- Add rate limiting of auth endpoints
- Added security headers to netlify frontend
- Sanitize crawled HTML in the backend
- Update default number of concurrent crawlers
- Crawl every 30 minutes instead of every hour
- Fixed link to Mastodon account
- Update links to @fediversespace Mastodon account
- Remove staging backend server
- Fixed frontend crash when instance node missing
- Update dependencies
- Fixed some unsuccessful crawls being saved without their errors
- Fix insularity score only working for > 0
- Fixed error when viewing some uncrawlable instances
- Fix navbar z-index
- Optimize query used for generating status rate
- Add support for logging in via an ActivityPub direct message to the instance admin.
- Added option to hide edges between instances if there are only mentions in one direction (off by default).
- Added note to neighbors tab to make it explicit that blocked instances may appear.
- Added federation tab that shows federation restrictions (only available for some Pleroma instances).
- Add tabular view of instances.
- Edges are no longer shown between instances where one blocks the other (based on the federation list in nodeinfo).
- Add caching to graph + instance endpoints to better handle traffic spikes.
- Added ON DELETE to
most_recent_crawltable, such that it can handle previously-crawled but now-dead instances. - You can now login to the admin view by clicking, not just by pressing enter.
- Add handling for weirdly-formatted Friendica peers.
- If the details of an instance fail to load, it's now easy to dismiss the error.
- Add Friendica crawler (only supports peers; there's no timeline API endpoint.)
- Color more server types on the map -- Hubzilla, Plume, Pixelfed, and Wordpress.
- Cleaned up ElasticSearch configuration in backend.
- Remove color-coding by activity per user. The vast majority of instances had the exact same color so this wasn't very useful.
- Added missing indices on
crawlsandcrawl_interactionstables. - Added table to store most recent crawl. This speeds up the instance view by a lot!
- Add nodeinfo and GNU Social crawler.
- Thanks to nodeinfo, Peertube and Writefreely are now also displayed on the map.
- Note that the information about connections comes from other instances.
- You can now zoom slightly further out on the map to see more of the fediverse at once.
- Database deletions are now properly handled with
ON DELETE CASCADEwhere necessary.
- Added Misskey crawler.
- Crawl instances that are down or unrecognized less often.
- Fixed broken instance view on mobile devices.
- Increased database connection timeout - required as the database grows!
- Fixed a wonky search UI when there are no results.
- You can now click a button in the search bar to search (you can also still just press enter, of course).
- You can now filter searches by instance type.
- Added toggle to show/hide edges on graph.
- Full-text search across instance descriptions now supports the following languages: arabic, armenian, basque, bengali, brazilian, bulgarian, catalan, cjk (i.e. chinese, japanese, korean), czech, danish, dutch, english, finnish, french, galician, german, greek, hindi, hungarian, indonesian, irish, italian, latvian, lithuanian, norwegian, persian, romanian, russian, sorani, spanish, swedish, turkish, thai.
- Added a warning on mobile devices suggesting to view the site on a larger computer.
- Performance improvements when opening the app on something that isn't the graph.
- There are now fewer irrelevant search results.
- Clarify that the admin page only works for Mastodon and Pleroma instances.
- Fixed some instances being duplicated (due to un-normalized data).
- Fixed mobile instance view erroring for uncrawled instances.
- Improved error handling in admin login page.
- Instances that opt-out will no longer show up in search results ever, nor are they accessible through the API.
- Instance administrators can now log in to opt in or out of crawling.
- Added ElasticSearch full-text search over instance domains and descriptions.
- Search results are now highlighted on the graph.
- When you hover a search result, it is now highlighted on the graph.
- Instance details now show activity rate (average number of statuses posted per day).
- It's now possible to color code by activity rate.
- Instances are now crawled hourly instead of every 30 minutes.
- There are now 100 concurrent crawl workers by default (increased from 50).
- The colors for color coding have been made brighter (more visible against the dark background).
- Fixed a process leak that could cause the server to OOM.
- Domains are no longer added to the queue twice.
- The server administrator can now be notified if there's a new potential spam domain.
- It is now possible to color code the graph by instance type (e.g. Mastodon, Pleroma, etc.)
- It's now shown in the front-end if an instance wasn't crawled because of its robots.txt.
- You can now link directly to instances at e.g. /instance/mastodon.social.
- Instance details now have a link to the corresponding fediverse.network page.
- The main graph is no longer displayed on mobile. Instead, a smaller neighborhood graph is shown.
- You no longer have to zoom completely in to see labels.
- Label size is now dependent on the instance size.
- The instance lookup field is now front-and-center. Is also uses the backend for faster lookups. This is to improve performance, and it lays the groundwork for full-text search over instance names and descriptions.
- The reset-graph-view button now explains what it's for when you hover over it.
- Previously, direct links to /about would return a 404 on Netlify's infrastructure. No longer.
- The backend has been completely rewritten in Elixir for improved stability and performance.
- An "insularity score" was added to show the percentage of mentions to users on the same instance.
- The crawler now respects robots.txt.
- Migrated the frontend graph from Sigma.js to Cytoscape.js.
- To improve performance, instances with no neighbors are no longer shown on the graph.
- The /api/v1 endpoint no longer exists; now there's a new /api.
- Spam domains can be blacklisted in the backend crawler's config.
- Add basic automated security scanning (using Sobelow and Gitlab's dependency scanning).
- Initial release. The date above is inaccurate; this first version was released sometime in the fall of 2018.
- This release had a Django backend and a Sigma.js graph.