Extract text from the Canadian business directory

extract.php

#!/usr/bin/env php
<?php
/**
 * Note: This requires the html5-php library.
 * https://github.com/html5lib/html5lib-php
 */

/*
 * Exit Codes
 */
define ('ERROR_NONE',       0);
define ('ERROR_USAGE',      1);
define ('ERROR_NOSUCHFILE', 2);
define ('ERROR_PARSING',    3);

// Ensure usage is correct
if ($_SERVER['argc'] != 2) {
	echo sprintf ("Usage: %s [input file]\n", $_SERVER['argv'][0]);
	exit (ERROR_USAGE);
}

// Ensure that a proper filename was passed
$input_file = $_SERVER['argv'][1];
if (!file_exists ($input_file) || !is_readable ($input_file)) {
	echo sprintf ("Input file must exist and be readable.\n");
	exit (ERROR_NOSUCHFILE);
}

// Support library.  This uses proper separators to load html5lib-php/library/HTML5/Parser.php
require_once  __DIR__ . join(DIRECTORY_SEPARATOR, array ('', 'html5lib-php', 'library', 'HTML5', 'Parser.php'));

// Load the document
$doc = HTML5_Parser::parse (file_get_contents ($input_file));
$xpath = new DOMXpath($doc);

// Look for divs with the class of "col-md-4 mrgn-sm-sm"
$elements = $xpath->query('//div[contains(@class, "col-md-4") and contains(@class, "mrgn-sm-sm")]');
if (count ($elements) != 1) {
	echo "Unable to parse document.  There should be exactly one tag with the class col-md5 mrgn-sm-sm.\n";
	exit (ERROR_PARSING);
}
$element = $elements->item(0);
unset ($elements);

/**
 * Temporary variable used to store the current attribute
 */
$attr_name = NULL;

/**
 * Stores the current business
 */
$business = array();

// Now that we have the section we want, extract the data
foreach ($element->childNodes as $child_node) {
	// h2 tags are labels
	if ($child_node instanceof DOMElement) {
		if ($child_node->tagName == 'h2') {
			// Clean any whitespace or colons
			$attr_name = trim ($attr_name, " \t\n\r\0\x0B:");

			// Convert to key for storage by converting "Mailing Address:" to "mailing_address"
			$attr_name = str_replace (' ', '_', strtolower ($child_node->nodeValue));
		} else if ($child_node->tagName == 'address') {
			// The key will come from the H2 tag, and the value will come from the address tag
			$address = $child_node->nodeValue;

			/*
			 * Clean up white space a little bit.
			 */
			// Eliminate any DOS-style line endings, and replace all multiple newlines and tabs with a single newline
			$address = preg_replace ("/[\t\r\n]+/", "\n", $address);

			// Eliminate any spaces after a newline
			$address = preg_replace ("/\n[\s]+/", "\n", $address);

			// Remove leading and trailing spaces
			$address = trim ($address);

			// Save the data
			$business[$attr_name] = $address;
		} else if ($child_node->tagName == 'p') {
			/*
			 * Paragraphs are a little different.  They have a <abbr> tag, then the value.
			 */
			foreach ($child_node->childNodes as $sub_node) {
				if (($sub_node instanceof DOMElement) && $sub_node->tagName == 'abbr') {
					// The title element contains the full value
					$attr_name = strtolower ($sub_node->getAttribute ('title'));
				} else if ($sub_node instanceof DOMText) {
					// Clean any leading or trailing whitespace or colons
					$value = trim ($sub_node->nodeValue, " \t\n\r\0\x0B:");

					// Save this to the business
					$business[$attr_name] = $value;

					// Clean up (not necessary, but still good practice)
					unset ($value);
				}
			}
		}
	}
}

var_dump ($business);