View source
<?php
function commerce_reports_patterns_load($min_support, &$context) {
if (empty($context['sandbox'])) {
$total_transactions = db_query("SELECT COUNT(order_id) FROM {commerce_order} WHERE status IN ('completed', 'pending')")
->fetchField();
$min_transactions = (int) floor($total_transactions * $min_support);
$context['results'] = array(
'total_transactions' => $total_transactions,
'minimum_transactions' => $min_transactions,
'frequent_sets' => array(
1 => array(
array(
'prefix' => array(),
'sets' => array(),
),
),
),
);
$context['sandbox']['products'] = db_query('SELECT sku, COUNT(li.line_item_id) AS occurences FROM {commerce_product} p LEFT JOIN {commerce_line_item} li ON p.sku = li.line_item_label GROUP BY sku HAVING occurences >= :min_occurences ORDER BY occurences DESC', array(
':min_occurences' => $min_transactions,
))
->fetchAll();
$context['sandbox']['progress'] = 0;
$context['sandbox']['max'] = count($context['sandbox']['products']);
if (!$context['sandbox']['max']) {
drupal_set_message(t('The value of your minimum support is too high to include any of the available transactions.'), 'error');
return;
}
}
$product = $context['sandbox']['products'][$context['sandbox']['progress']];
if ($product) {
$context['results']['frequent_sets'][1][0]['sets'][] = array(
'items' => array(
$product->sku,
),
'tidlist' => db_query("SELECT DISTINCT order_id FROM {commerce_line_item} WHERE line_item_label = :sku", array(
':sku' => $product->sku,
))
->fetchCol(),
);
}
$context['sandbox']['progress']++;
if ($context['sandbox']['progress'] < $context['sandbox']['max']) {
$context['finished'] = $context['sandbox']['progress'] / $context['sandbox']['max'];
}
else {
if (!count($context['results']['frequent_sets'][1][0]['sets'])) {
unset($context['results']['frequent_sets'][1]);
}
}
}
function commerce_reports_patterns_frequent_itemset(&$context) {
if (empty($context['sandbox'])) {
$context['sandbox']['set_length'] = 2;
$context['sandbox']['current_partition'] = 0;
$context['sandbox']['first_itemset'] = 0;
$context['sandbox']['second_itemset'] = 1;
}
$length = $context['sandbox']['set_length'];
if (empty($context['results']['frequent_sets'][$length - 1])) {
return;
}
if (empty($context['results']['frequent_sets'][$length])) {
$context['results']['frequent_sets'][$length] = array();
}
$prev_partitions =& $context['results']['frequent_sets'][$length - 1];
$prev_partitions_count = count($prev_partitions);
for ($x = 0; $x < 50; $x++) {
if (!empty($prev_partitions[$context['sandbox']['current_partition']])) {
$partition_itemset_count = count($prev_partitions[$context['sandbox']['current_partition']]['sets']);
}
else {
$partition_itemset_count = 0;
}
$a = $context['sandbox']['first_itemset'];
$b = $context['sandbox']['second_itemset'];
if ($context['sandbox']['second_itemset'] >= $partition_itemset_count) {
if ($context['sandbox']['first_itemset'] >= $partition_itemset_count) {
if ($context['sandbox']['current_partition'] >= $prev_partitions_count) {
$context['sandbox']['current_partition'] = 0;
$context['sandbox']['set_length']++;
break;
}
else {
$context['sandbox']['current_partition']++;
}
$context['sandbox']['first_itemset'] = 0;
}
else {
$context['sandbox']['first_itemset']++;
}
$context['sandbox']['second_itemset'] = $context['sandbox']['first_itemset'] + 1;
continue;
}
else {
$context['sandbox']['second_itemset']++;
}
$partition =& $prev_partitions[$context['sandbox']['current_partition']];
$partition_prefix =& $partition['prefix'];
$partition_itemset =& $partition['sets'];
$a_set =& $partition_itemset[$a];
$b_set =& $partition_itemset[$b];
$candidate = array(
'items' => array_values(array_unique(array_merge($a_set['items'], $b_set['items']))),
'tidlist' => array_intersect($a_set['tidlist'], $b_set['tidlist']),
);
$occurences = max(count($candidate['tidlist']), 0);
if ($occurences >= $context['results']['minimum_transactions']) {
$new_items = array_diff($candidate['items'], $partition_prefix);
$prefix = array_merge($partition_prefix, array(
reset($new_items),
));
_commerce_reports_patterns_add_set($context['results']['frequent_sets'][$length], $prefix, $candidate);
}
}
if ($count = count($context['results']['frequent_sets'][$length])) {
$context['finished'] = 0;
$context['message'] = t('Generated %count frequent item sets with length %length', array(
'%count' => $count,
'%length' => $length,
));
}
else {
unset($context['results']['frequent_sets'][$length]);
}
}
function _commerce_reports_patterns_add_set(&$partitions, $prefix, $set) {
sort($prefix);
foreach ($partitions as &$partition) {
if ($partition['prefix'] == $prefix) {
$partition['sets'][] = $set;
return;
}
}
$partitions[] = array(
'prefix' => $prefix,
'sets' => array(
$set,
),
);
}
function commerce_reports_patterns_association_rules($min_confidence, &$context) {
if (empty($context['sandbox'])) {
$context['sandbox']['length'] = 2;
$context['sandbox']['partition'] = 0;
$context['sandbox']['set'] = 0;
$context['results']['count'] = 0;
$context['results']['rules'] = array();
db_query('TRUNCATE {commerce_reports_patterns}');
}
if ($context['sandbox']['set'] > count($context['results']['frequent_sets'][$context['sandbox']['length']][$context['sandbox']['partition']]['sets']) - 1) {
$context['sandbox']['partition']++;
$context['sandbox']['set'] = 0;
}
if ($context['sandbox']['partition'] > count($context['results']['frequent_sets'][$context['sandbox']['length']]) - 1) {
$context['sandbox']['length']++;
$context['sandbox']['partition'] = 0;
}
if ($context['sandbox']['length'] > count($context['results']['frequent_sets'])) {
$context['finished'] = 1;
return;
}
else {
$context['finished'] = 0;
}
$set =& $context['results']['frequent_sets'][$context['sandbox']['length']][$context['sandbox']['partition']]['sets'][$context['sandbox']['set']];
$items =& $set['items'];
$set_support_count = count($set['tidlist']);
$count = count($items);
$members = pow(2, $count);
for ($i = 0; $i < $members; $i++) {
$b = sprintf('%0' . $count . 'b', $i);
$rule = new stdClass();
$rule->size = $count;
$rule->if_clause = array();
$rule->then_clause = array();
for ($j = 0; $j < $count; $j++) {
if ($b[$j] == '1') {
$rule->if_clause[] = $items[$j];
}
else {
$rule->then_clause[] = $items[$j];
}
}
if (!empty($rule->then_clause) && ($if_count = count($rule->if_clause))) {
$if_support_count = db_query("SELECT COUNT(*) FROM (SELECT order_id FROM {commerce_line_item} WHERE line_item_label IN (:sku) GROUP BY order_id HAVING COUNT(line_item_id) = :count) cli", array(
':sku' => $rule->if_clause,
':count' => $if_count,
))
->fetchField();
if ($if_support_count && ($confidence = $set_support_count / $if_support_count) >= $min_confidence) {
$rule->confidence = $confidence;
drupal_write_record('commerce_reports_patterns', $rule);
$context['results']['count']++;
}
}
unset($set);
}
$context['sandbox']['set']++;
$context['message'] = t('Found %count association rules', array(
'%count' => $context['results']['count'],
));
}
function commerce_reports_patterns_finished($success, $results, $operations) {
if ($success) {
drupal_set_message(t('Successfully finished mining frequent itemsets, the largest set contained %length items.', array(
'%length' => count($results['frequent_sets']),
)));
drupal_set_message(t('Successfully finished generating %count association rules.', array(
'%count' => $results['count'],
)));
}
else {
drupal_set_message(t('There was a problem generating frequent itemsets.'), 'error');
}
}
