From f35490d10721c0fbf78b0c11f909af4f3ea232ea Mon Sep 17 00:00:00 2001
From: Edward Ribeiro <eribeiro@interlegis.leg.br>
Date: Mon, 11 Feb 2019 18:36:46 -0200
Subject: [PATCH] Atualiza schema e adiciona script de stats

---
 solr/docs_stats.py                  | 37 +++++++++++++++++++++++++++++
 solr/sapl_configset/conf/schema.xml |  1 +
 2 files changed, 38 insertions(+)
 create mode 100644 solr/docs_stats.py

diff --git a/solr/docs_stats.py b/solr/docs_stats.py
new file mode 100644
index 000000000..5343cfb15
--- /dev/null
+++ b/solr/docs_stats.py
@@ -0,0 +1,37 @@
+import requests
+
+"""
+    Imprime quantidade de colletions, qtd de documentos por collection e
+    total de documentos indexados.
+"""
+
+BASE_URL='http://localhost:8983/solr'
+
+
+if __name__=='__main__':
+
+    resp = requests.get(BASE_URL+'/admin/collections?action=LIST')
+
+    collections = sorted(resp.json()['collections'])
+
+    largest_col = (None,-1)
+    total_docs = 0
+    
+    print("Collection\t\t\tNumber of documents")
+    print("--------------------------------------------------")
+    
+    for c in collections:
+        r = requests.get(BASE_URL+'/{}/select?q=*:*&rows=0'.format(c))
+        num_docs = r.json()['response']['numFound']
+        total_docs += num_docs
+
+        if num_docs >= largest_col[1]:
+            largest_col = (c, num_docs)
+        
+        print("%30s\t%6s" % (c, num_docs))
+
+    print("------------------------------------------")
+    print("- Number of collections: %s\n" % len(collections))
+    print("- Largest collection: '%s' (%s docs)\n" % largest_col)
+    print("- Total documents accross all collections: %s\n" % total_docs)
+
diff --git a/solr/sapl_configset/conf/schema.xml b/solr/sapl_configset/conf/schema.xml
index 597033929..e230533e5 100644
--- a/solr/sapl_configset/conf/schema.xml
+++ b/solr/sapl_configset/conf/schema.xml
@@ -151,6 +151,7 @@
     <field name="django_id" type="string" indexed="true" stored="true" multiValued="false"/>
     <field name="_version_" type="plong" indexed="true" stored ="true"/>
     <field name="text" type="text_pt" indexed="true" stored="true" multiValued="false" />
+    <field name="indextime" type="pdate" default="NOW" />
 
   </fields>