Spent a bunch of time over the last two days working on integrating search into yarnd
, but got sidetracked instead with improving the the search engine called yarns
that I host at search.twtxt.net – Bluge and Bleve are such similar libraries I’m actually not sure which is better now 😅 – Fortunately I’ve built the search engine codebase in such a way that bleve can be swapped out by bluge by only implementing the Indexer
interface 👌
Going to do that tomorrow after I let yarns
recrawl and scrape the Yarn/Twtxt space from my iMac while I go get some sleep 😴 #Yarn.social #search
Made a bunch more improvements to the Twtxt Search Engine 👌 At least @lyse@lyse.isobeef.org likes the improvements so far, even if I completely suck at UI/UX 😅 #yarns #twtxt #search
📣 Updates to Search Engine:
- Remove the default fuzziness of 1 (but see below)
- Fixed several bugs in the pager (trying searching for “Hello World” and sort by “Oldest”)
- Try to auto-detect the type of the query string by parsing it as a query string, if that fails default to match query…
This last feature GuessQueryType()
now allows advanced users to type in a Query String query and letting the backend figure out that’s what you meant. So for example you can search for terms with fuzziness quite easily by just typing in the search box: fuck~1
(for example) or search for mentions by typing in mentions:prologic@twtxt.net
– Enjoy 🤗 #Twtxt #Yarn.social #Search #Engine
Added support for #tag clouds and #search to timeline. Based on code from @dfaria.eu@dfaria.eu🙏
Live at: http://darch.dk/timeline/?profile=https://darch.dk/twtxt.txt