1. If you press "f" on the first page of search result it returns to the terminal with "[ERROR] Already at the first page." - no need to interrupt session here.
2. Consider implementing history of previous 10 queries.
I like it, but it just reminds me of how much we missed by letting terminals stagnate through the 90's.
You want those URLs that turn up to be clickable. Modern terminals help you a bit here. But in a parrallel universe (Emacs perhaps?), the terminal itself would display the graphics and have markup for links etc.
Surprisingly, we live in an alternate reality. That's called a browser, but the keyboard input is a little different -- you have to hit Ctrl/Cmd + L every time you want to type in a new location.
I currently use http://surfraw.alioth.debian.org/ along with w3m and it works great but does not appear to be currently maintained. Would you have any interest in adding similar search options to your application? Query Wikipedia, HN or Stack Overflow.... directly?
We introduced googler @t in the latest release to handle that to some extent using Google Site Search. Yes, it would be fun to support direct site searches but it would also be quite difficult to maintain it.
I did give it a shot last week and found the search strings and result patterns of probably every other site different. Just imagine keeping up with changes in the URLs! For a while I thought "why not launch the URLs directly in the browser" to keep it simpler? But then googler wouldn't remain a fun terminal utility to develop anymore. ;)
Do you use a Google api or do you send http request and scrape the response html? As I understand the gogle search api got deprecated some years back. I am a few days from my laptop so can't properly read it for myself :-) thanks!
18 comments
[ 46.5 ms ] story [ 257 ms ] threadHomepage: https://github.com/jarun/googler
Features:
- Google Search, Google Site Search, Google News
- Fast and clean (no ads, stray URLs or clutter), custom color
- Navigate result pages from omniprompt, open URLs in browser
- Effortless keyword-based site search with googler @t add-on
- Fetch n results in a go, start at the nth result
- Disable automatic spelling correction and search exact keywords
- Specify duration, country/domain (default: worldwide/.com), language
- Google keywords (e.g. filetype:mime, site:somesite.com) support
- Open the first result directly in browser (as in I'm Feeling Lucky)
- Non-stop searches: fire new searches at omniprompt without exiting
- HTTPS proxy, User Agent, TLS 1.2 (default) support
- Man page with examples, completion scripts for Bash, Zsh and Fish
- Minimal dependencies
Couple small suggestions:
1. If you press "f" on the first page of search result it returns to the terminal with "[ERROR] Already at the first page." - no need to interrupt session here.
2. Consider implementing history of previous 10 queries.
Thanks
You want those URLs that turn up to be clickable. Modern terminals help you a bit here. But in a parrallel universe (Emacs perhaps?), the terminal itself would display the graphics and have markup for links etc.
I currently use http://surfraw.alioth.debian.org/ along with w3m and it works great but does not appear to be currently maintained. Would you have any interest in adding similar search options to your application? Query Wikipedia, HN or Stack Overflow.... directly?
I did give it a shot last week and found the search strings and result patterns of probably every other site different. Just imagine keeping up with changes in the URLs! For a while I thought "why not launch the URLs directly in the browser" to keep it simpler? But then googler wouldn't remain a fun terminal utility to develop anymore. ;)