I recently wrote a python script to surf the web (see search results) directly into the terminal.
Here is a demo
The script is available on my github
Itโs not that big of a deal since all credit goes to searx a privacy respecting FOSS metasearch engine
Why Searx
Why not?
- It respects your privacy. No user-tracking
- Open-source
- You can even set up your own (private/public) searx instance on a VPS. List of public instances
- Aggregates results from more than 70 search services (DuckDuckGo, Wikipedia, etc)
Installation
The script doesnโt require any dependencies (except Python 3), just download the script and run it.
Note that surf was written for/as a Linux user. It should work on Mac without any problems. If you are on Windows consider sending patches if the script breaks. Thanks!
wget -q https://raw.githubusercontent.com/Bhupesh-V/.Varshney/master/scripts/surf
chmod +x surf && mv surf $HOME/.local/bin/
If you are more of a bad boy/girl then add the location of surf script to your PATH
export PATH="DIR_PATH_WHICH_CONTAINS_SURF:$PATH"
# e.g
export PATH="$HOME/Documents/.Varshney/scripts:$PATH"
You can view help & usage on surf by doing a simple surf --help
. Letโs go through all the options available.
Usage
- Just run the script, without any arguments. surf will prompt you for a search query:
$ surf
- Show link description.
$ surf -d
- Provide search query as an argument (also
--query
):
$ surf -dq "how to change the world"
- You can also pipe data to surf
ยฏ\_(ใ)_/ยฏ
:
$ echo "search this" | surf
- Use
-c
to specify search category (default: general):
$ surf -c "videos" -dq "how to make a pizza"
Other categories include: โnewsโ, โfilesโ, โimagesโ, โmapโ
- Update searx instance cache list before fetching results:
$ surf -udq "when is doomsday"
Updating cache is not usually required, I recommend running it once/twice a week to get a list of active instances.
The cache is located at $HOME/.local/.searx_instances
on linux and $HOME/Library/Caches/.searx_instances
on Mac
- Show only top N results (also
--top
):
$ surf -t 5 -dq "check if key exists in map"
searx supports those sneaky search tricks as well,
surf -dq "site:youtube.com how to make pizza"
surf -dq "site:stackoverflowcom <my-error-message>"
surf -dq "intitle:best vim plugins"
surf -dq "inurl:docs.djangoproject.com templates"
Well now you would be asking โwhatโs the use of this bhupesh, If I have to open the link in the browser at the endโ
But what I realized is that it may accidentally fit right into my/your workflow if you use Vim, let me show how
Browsing Inside Vim, say whaat
We need a way to open links right from our vim terminal. Add this to your vimrc
or init.vim
" Open hyper link in current line
function! OpenLink()
let links = []
try
call substitute(getline('.'), 'http[s]\?:\/\/[^) \"]*', '\=add(links, submatch(0))', 'g')
echo "Opening " . links[0]
exe "silent! !xdg-open " . links[0]
catch E684
echo "No link found :("
return
endtry
endfunction
The line
exe "silent! !xdg-open " . links[0]
will directly open your default browser with the link.
If you are not using X11, then you can just swap it directly with the browser of your choice or your default file manager
- Firefox
exe "silent! !firefox --new-tab " . links[0]
- Chromium
exe "silent! !chromium-browser " . links[0]
Also, I recently found a less bulky approach on vimtricks.com
function! OpenURLUnderCursor()
let s:uri = expand('<cWORD>')
let s:uri = substitute(s:uri, '?', '\\?', '')
let s:uri = shellescape(s:uri, 1)
if s:uri != ''
silent exec "!open '".s:uri."'"
:redraw!
endif
endfunction
nnoremap gx :call OpenURLUnderCursor()<CR>
Searching Error messages inside Vim
Say I am debugging something and end up with a never before seen Error message. I can quickly send the error string to surf
and get links to StackOverflow or GitHub!
Here is a demo of me using flake on surf & searching about rules:
How this works is pretty straightforward. We just need to get the current visual selection.
function! GetVisualSelection()
" Thanks: https://stackoverflow.com/a/6271254/8209510
let [line_start, column_start] = getpos("'<")[1:2]
let [line_end, column_end] = getpos("'>")[1:2]
let lines = getline(line_start, line_end)
let lines[-1] = lines[-1][: column_end - (&selection == 'inclusive' ? 1 : 2)]
let lines[0] = lines[0][column_start - 1:]
return join(lines, "\n")
endfunction
xnoremap <leader>t <esc>:split <bar> call SendQueryToTerm()<CR>
function! SendQueryToTerm()
let selection = GetVisualSelection()
exe ":term surf -dq \'" . selection . "\'"
endfunction
This will open a split above with a new terminal buffer.
Convinced yet?
Maybe not
If you liked surf make sure to subscribe to the Atom Feed or watch my dotfiles repo for the latest development or better fork & hack your own features!!
Top comments (4)
I got through about half this post thinking, "why not just use lynx|links?" and then saw your Vim integration and switched to, "hey that's an idea".
I think I might take this idea play with it - instead of using
xdg-open
, make it figure out what browser is available and uselinks
if you're on a remote system throughssh
- or maybe to take results and put them in the quickfix window?With more and more ugly big banners and ads, this is the way how to find relevant information on the internet
Nice article. This is good information.
Thanks :)