Question

I have an XML document collection, an inverted file indexer, and a command-line tool for searching the index (or indices) produced by the indexer. Note that the latter returns a list of document IDs and various statistics about each document (rankings according to various functions, term hits, etc) rather than the actual document text. Both programs were written in straight C (by me).

• The collection is not huge (~1GB).
• The index is about 10-20% of the collection size.
• This is not intended (and never will be) for public use (using it will require logging in).
• It needs to run with client-side scripting totally disabled.

I’d like to whip up a simple web frontend that would allow me to query the index with a search term or terms and present the results appropriately, but it’s been a while since I touched any web stuff.

I want to see more or less the same info a query returns at the moment, but I’m not sure whether to write something (e.g. PHP, Ruby – alternative suggestions are welcome) that calls my command-line query program and processes the output, or whether re-implementing the query program would be more appropriate.

Are there any distinct advantages one has over the other? Security risks?
And can anyone recommend me a lightweight framework or library appropriate for any of this? (Like I said, haven’t touched web stuff in a while.)

Should I call the CLI query program? Why or why not?

(=/ I hope I’m not being too vague… do tell me if I should be asking this in a different manner.)

Answer 1

For something simple like this, I would use PHP and an Apache server. Why?

It doesn’t require a web framework to interface between Apache; less complexity = less time for you to spend configuring. You could just install Apache and the php module, then drop in this file in your web-root, and point a html form to http://127.0.0.1/indexer.php with the textareas "name" and "author":

<?php
$required_terms = array("name", "author");

foreach ($required_terms as $value) {
    if (!isset($_POST[$value])) {
        printf("The search term \"%s\" was missing", $value);
        exit;
    }
}

$terminal_command = sprintf("/usr/bin/indexer -n %s -a %s", $_POST["name"], $_POST["author"]);
print exec($terminal_command);

(Note this is just to show the simplicity, it needs validation of the post values received).

Then this would run your application with the 2 values as arguments, then print whatever was sent to stdout by your application. No more hassle or things to setup. It would take you a couple of minutes to get up and running.

So the main reason would be simple and fast to setup, for something internal and simple as this.