From f49e5870a2a2ad8cb34a1b20785fd4f402f8133e Mon Sep 17 00:00:00 2001
From: Gabriel Reimers <gabriel@wokabulary.com>
Date: Fri, 23 Sep 2016 17:56:40 +0200
Subject: [PATCH] - adds RISM demo crawler

---
 command_line_scripts/indexRISM.php | 40 ++++++++++++++++++++++++++++++
 src/RISMIncipitCrawler.php         |  2 +-
 2 files changed, 41 insertions(+), 1 deletion(-)
 create mode 100644 command_line_scripts/indexRISM.php

diff --git a/command_line_scripts/indexRISM.php b/command_line_scripts/indexRISM.php
new file mode 100644
index 0000000..3e1a8d1
--- /dev/null
+++ b/command_line_scripts/indexRISM.php
@@ -0,0 +1,40 @@
+<?php
+namespace ADWLM\IncipitSearch;
+
+require 'vendor/autoload.php';
+
+
+/**
+ * This is to be called from command line to index the Gluck Gesamtausgabe catalog.
+ * ATTENTION: this must be called from the same directory as index.php
+ * or paths will be wrong
+ *
+ *
+ * Copyright notice
+ *
+ * (c) 2016
+ * Anna Neovesky  Anna.Neovesky@adwmainz.de
+ * Gabriel Reimers g.a.reimers@gmail.com
+ *
+ * Digital Academy www.digitale-akademie.de
+ * Academy of Sciences and Literatur | Mainz www.adwmainz.de
+ *
+ * Licensed under The MIT License (MIT)
+ *
+ * @package ADWLM\IncipitSearch
+ */
+
+
+/*
+ *
+ * ATTENTION:
+ * This only crawls a small sample of RISM and is only for demo purposes
+ *
+ */
+
+echo "====================\nATTENTION\n=======================\nThe RISM crawler is only for demo purposes and only crawls a small subset of RISM.\nThis is not fit for production\n\n";
+
+$crawler = new RISMIncipitCrawler();
+
+$crawler->createIndex();
+$crawler->crawlCatalog();
\ No newline at end of file
diff --git a/src/RISMIncipitCrawler.php b/src/RISMIncipitCrawler.php
index 5266d23..a9fee24 100644
--- a/src/RISMIncipitCrawler.php
+++ b/src/RISMIncipitCrawler.php
@@ -112,7 +112,7 @@ public function crawlCatalog()
     {
 
         $startID = 400110660;
-        $endID =   400110862;
+        $endID =   400110999;
 
         for ($i = $startID; $i < $endID; $i++) {
             $url = "https://opac.rism.info/id/rismid/" . $i . "?format=marc";