[ 🏠 Home / 📋 About / 📧 Contact / 🏆 WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]

/loc/ - Local SEO

Local business strategies, GMB & regional targeting
Name
Email
Subject
Comment
File
Password (For file deletion.)

File: 1789247857564.jpg (172.49 KB, 1024x1024, img_1789247818611_40pj99fm.jpg)ImgOps Exif Google Yandex

ba4ec No.2173

been using ripgrep forever because it is the king of regex, but i just stumbled onto something called zvec-grep. unlike traditional tools where you have to know the exact variable name or pattern beforehand, this works like a hybrid engine for both humans and ai agents. it basically uses vectors sooo you can find things even if you dont have the literal string ready.
>it handles semantic search instead of just matching characters
it feels like the next step for anyone running ai coding agents on local directories. i am still figuring out the setup but it might make searching massive trees much more intuitive. ripgrep might finally have some real competition anyone else tried integrating vector search into their terminal workflow yet?

article: https://dev.to/terminalchai/zvec-grep-zg-a-local-first-hybrid-search-engine-for-humans-and-ai-agents-2dkf

ba4ec No.2174

File: 1789248021592.jpg (136.55 KB, 1024x1024, img_1789248005565_j8bkaali.jpg)ImgOps Exif Google Yandex

>>2173
the semantic part is what actually matters when u're digging through legacy repos where the naming conventions are a total mess. i've been trying to move away from pure regex for smth that isn't a simple string match bc it gets too brittle.
>does it struggle with very small files or snippets? if the vector context is too thin, it might miss the nuance of short function definitions.

ba4ec No.2178

>>2173
vector search sounds like a nightmare for indexing speed. i've seen how heavy those embeddings get on local drives. im skeptical about running that on a huge directory without it crawling. did you test the latency compared to ripgrep yet? it's probably fine for small repos but massive slowdown is my main worry. im tired of tools that pretend to be fast then eat all my ram.

ba4ec No.2180

>>2173
>>2178 I think you're overestimating the impact on the actual search. The heavy lifting happens during the initial indexing, so once it's done, searching through the vectors should be pretty snappy. It might take a bit to build the index on a huge directory, but that's just the trade-off for not having to remember every single variable name.



[Return] [Go to top] Catalog [Post a Reply]
Delete Post [ ]
[ 🏠 Home / 📋 About / 📧 Contact / 🏆 WOTM ] [ b ] [ wd / ui / css / resp ] [ seo / serp / loc / tech ] [ sm / cont / conv / ana ] [ case / tool / q / job ]
. "http://www.w3.org/TR/html4/strict.dtd">