Given a browser, a search engine's transformation function, and a search term, perform the requested search in the browser.

Given a base URL, create the SearchEngine that escapes the query and appends it to the base. You can easily define a new engine locally using exported functions without needing to modify XMonad.Actions.Search:

myNewEngine = searchEngine "site" "https://site.com/search="

The important thing is that the site has a interface which accepts the escaped query string as part of the URL. Alas, the exact URL to feed searchEngine varies from site to site, often considerably, so there's no general way to cover this.

Generally, examining the resultant URL of a search will allow you to reverse-engineer it if you can't find the necessary URL already described in other projects such as Surfraw.

If your search engine is more complex than this (you may want to identify the kind of input and make the search URL dependent on the input or put the query inside of a URL instead of in the end) you can use the alternative searchEngineF function.

searchFunc :: String -> String
searchFunc s | "wiki:"    `isPrefixOf` s = "https://en.wikipedia.org/wiki/" ++ (escape $ tail $ snd $ break (==':') s)
             | "https://" `isPrefixOf` s = s
             | otherwise                 = (use google) s
myNewEngine = searchEngineF "mymulti" searchFunc

searchFunc here searches for a word in wikipedia if it has a prefix of "wiki:" (you can use the escape function to escape any forbidden characters), opens an address directly if it starts with "https://" and otherwise uses the provided google search engine. You can use other engines inside of your own through the use function as shown above to make complex searches.

The user input will be automatically escaped in search engines created with searchEngine, searchEngineF, however, completely depends on the transformation function passed to it.

Like search, but in this case, the string is not specified but grabbed from the user's response to a prompt. Example:

, ((modm, xK_g), promptSearch greenXPConfig google)

This specializes "promptSearchBrowser" by supplying the browser argument as supplied by getBrowser from XMonad.Prompt.Shell.

Like search, but for use with the output from a Prompt; it grabs the Prompt's result, passes it to a given searchEngine and opens it in a given browser.

Like promptSearchBrowser, but only suggest previous searches for the given SearchEngine in the prompt.

Like search, but for use with the X selection; it grabs the selection, passes it to a given searchEngine and opens it in the default browser . Example:

, ((modm .|. shiftMask, xK_g), selectSearch google)

This specializes "selectSearchBrowser" by supplying the browser argument as supplied by getBrowser from XMonad.Prompt.Shell.

Like search, but for use with the X selection; it grabs the selection, passes it to a given searchEngine and opens it in a given browser.

\(\mathcal{O}(\min(m,n))\). The isPrefixOf function takes two lists and returns True iff the first list is a prefix of the second.

>>> "Hello" `isPrefixOf` "Hello World!"
True

>>> "Hello" `isPrefixOf` "Wello Horld!"
False

Escape the search string so search engines understand it. Only digits and ASCII letters are not encoded. All non ASCII characters which are encoded as UTF8

Given an already defined search engine, extracts its transformation function, making it easy to create compound search engines. For an instance you can use use google to get a function which makes the same transformation as the google search engine would.

This function wraps up a search engine and creates a new one, which works like the argument, but goes directly to a URL if one is given rather than searching.

myIntelligentGoogleEngine = intelligent google

Now if you search for https://xmonad.org it will directly open in your browser

Connects a few search engines into one. If the search engines' names are "s1", "s2" and "s3", then the resulting engine will use s1 if the query is s1:word, s2 if you type s2:word and s3 in all other cases.

Example:

multiEngine = intelligent (wikipedia !> mathworld !> (prefixAware google))

Now if you type "wiki:Haskell" it will search for "Haskell" in Wikipedia, "mathworld:integral" will search mathworld, and everything else will fall back to google. The use of intelligent will make sure that URLs are opened directly.

Makes a search engine prefix-aware. Especially useful together with !>. It will automatically remove the prefix from a query so that you don't end up searching for google:xmonad if google is your fallback engine and you explicitly add the prefix.

Changes search engine's name

A customized prompt indicating we are searching, and the name of the site.