PHP search engine

My search engine is written in PHP.
It has a simple Linux grep underneath it.
e.g. Search for grep.

Plan

1. We will search the Works of Shakespeare.
2. Take in an argument from a form.
3. The argument is in the URL (HTTP GET).
4. Make a command-line call to something like:

    cd (somedir) ; grep -i (searcharg) */*html 

   (";" allows two commands on one line)
5. Collect the grep output in a PHP variable.
6. Note we did case-insensitive search.

What to link the hit to

• What to link the hit to is a problem. My copy of the Works of Shakespeare is not on the web. And the University of Adelaide site has closed down.
• Basic solution: Link the hit to the Internet Archive copy of the page.
  For example, link Macbeth Act 3 to:

  https://web.archive.org/web/2019/https://ebooks.adelaide.edu.au/s/shakespeare/william/macbeth/act3.html

Make the hits clickable

• Go through the grep output line by line. Split the output into an array of lines by using explode on PHP_EOL.
• Go through the array with foreach.
• Each line looks like:

   filename: hit 

• Q. Doing an "explode" on ":" is no good. Why?

• Instead: For each line, find position of first ":" on line.
• filename is the substring before that position, hit is substring after
• Output something like:

  	
   $url = "https://web.archive.org/web/2019/https://ebooks.adelaide.edu.au/s/shakespeare/william/$filename";
   print "<a href='$url'>$filename</a>: $hit"; 
  

How to stop "infinite" output

What if grep returns tens of thousands of lines?
What if someone searches for a single character?

Here is a simple solution: Pipe grep to head, which collects the first n lines of what is piped into it.

1. Collect the first 500 lines and no more:

    grep string files | head -500 

2. The previous will actually terminate grep once 500 is reached.

ancientbrain.com      w2mind.org      humphrysfamilytree.com

On the Internet since 1987.      New 250 G VPS server.

Note: Links on this site to user-generated content like Wikipedia are highlighted in red as possibly unreliable. My view is that such links are highly useful but flawed.