From dc9afe2adf16e1375001eca9bc73c2211f73e27c Mon Sep 17 00:00:00 2001
From: Jimmy Berry <jberry@suse.com>
Date: Fri, 14 Sep 2018 15:50:37 -0500
Subject: [PATCH] metrics/access/ingest: detect product image paths and
 included in dump.

---
 metrics/access/ingest.php | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/metrics/access/ingest.php b/metrics/access/ingest.php
index 57c1a1a9..5582a537 100755
--- a/metrics/access/ingest.php
+++ b/metrics/access/ingest.php
@@ -3,11 +3,13 @@
 
 const REGEX_LINE = '/\S+ \S+ \S+ \[([^:]+:\d+:\d+:\d+ [^\]]+)\] "(\S+)(?: (\S+) \S+)?" (\S+) (\S+) "[^"]*" "[^"]*" .* size:(\S+) \S+(?: +"?(\S+-\S+-\S+-\S+-[^\s"]+|-)"? "?(dvd|ftp|-)"?)?/';
 const REGEX_PRODUCT = '#/(?:(tumbleweed)|distribution/(?:leap/)?(\d+\.\d+)|openSUSE(?:_|:/)(?:leap(?:_|:/))?(factory|tumbleweed|\d+\.\d+))#i';
+const REGEX_IMAGE = '#(?:/(?:iso|live)/[^/]+-(DVD|NET|GNOME-Live|KDE-Live|Rescue-CD|Kubic-DVD)-[^/]+\.iso(?:\.torrent)?|/jeos/[^/]+-(JeOS)\.[^/]+\.(?:qcow2|vhdx|vmdk|vmx)$)#';
 
 $total = 0;
 $total_invalid = 0;
 $total_product = [];
 $unique_product = [];
+$total_image_product = [];
 
 $file = $argc == 2 ? $argv[1] : 'php://stdin';
 $handle = fopen($file, 'r');
@@ -40,6 +42,15 @@ while (($line = fgets($handle)) !== false) {
     if (!isset($unique_product[$product][$uuid])) $unique_product[$product][$uuid] = 0;
     $unique_product[$product][$uuid] += 1;
   }
+
+  if (preg_match(REGEX_IMAGE, $match[3], $match_image)) {
+    // Remove empty match groups and select non-all match.
+    $values = array_filter($match_image);
+    $image = next($values);
+    if (!isset($total_image_product[$product])) $total_image_product[$product] = [];
+    if (!isset($total_image_product[$product][$image])) $total_image_product[$product][$image] = 0;
+    $total_image_product[$product][$image] += 1;
+  }
 }
 $position = ftell($handle);
 fclose($handle);
@@ -54,6 +65,7 @@ echo json_encode([
   'total' => $total,
   'total_product' => $total_product,
   'unique_product' => $unique_product,
+  'total_image_product' => $total_image_product,
   'total_invalid' => $total_invalid,
   'bytes' => $position,
 ]) . "\n"; // JSON_PRETTY_PRINT for debugging.