From 95e2f20a8895d443d594813a41174c751c128e3c Mon Sep 17 00:00:00 2001
From: Adam Kariv <adam.kariv@gmail.com>
Date: Mon, 28 Nov 2016 17:04:40 +0200
Subject: [PATCH] Maintenance mode

---
 requirements.txt   |  2 +-
 tools/reindexer.py | 89 ++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 90 insertions(+), 1 deletion(-)
 create mode 100644 tools/reindexer.py

diff --git a/requirements.txt b/requirements.txt
index 35673a4..2e4c776 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -11,4 +11,4 @@ sqlalchemy
 psycopg2
 cryptography
 elasticsearch>=1.0.0,<2.0.0
-os-package-registry>=0.0.4
+os-package-registry>=0.0.7
diff --git a/tools/reindexer.py b/tools/reindexer.py
new file mode 100644
index 0000000..d4574b6
--- /dev/null
+++ b/tools/reindexer.py
@@ -0,0 +1,89 @@
+import os
+import logging
+
+from elasticsearch import Elasticsearch, NotFoundError
+from os_package_registry import PackageRegistry
+from sqlalchemy import MetaData, create_engine
+
+logging.root.setLevel(logging.INFO)
+
+
+if __name__ == "__main__":
+
+    # Reindex ES
+    es_host = os.environ['OS_ELASTICSEARCH_ADDRESS']
+    es = Elasticsearch(hosts=[es_host], use_ssl='https' in es_host)
+
+    source_index = None
+    target_index = None
+    backup_index = 'packages-backup'
+    existing_alias = None
+    for i in range(2):
+        idx = 'packages-%d' % i
+        if es.indices.exists(idx):
+            if source_index is None:
+                source_index = idx
+                existing_alias = idx
+        else:
+            if target_index is None:
+                target_index = idx
+
+    if source_index is None:
+        source_index = 'packages'
+    if target_index is None:
+        target_index = 'packages-1'
+
+    assert source_index != target_index
+
+    logging.info('SOURCE INDEX %s', source_index)
+    logging.info('TARGET INDEX %s', target_index)
+    logging.info('BACKUO INDEX %s', backup_index)
+
+    try:
+        logging.info('DELETING BACKUP INDEX')
+        es.indices.delete(backup_index)
+    except NotFoundError:
+        logging.info('BACKUP INDEX NOT FOUND')
+
+    try:
+        logging.info('DELETING TARGET INDEX')
+        es.indices.delete(target_index)
+    except NotFoundError:
+        logging.info('TARGET INDEX NOT FOUND')
+
+    source_pr = PackageRegistry(es_instance=es, index_name=source_index)
+    backup_pr = PackageRegistry(es_instance=es, index_name=backup_index)
+    target_pr = PackageRegistry(es_instance=es, index_name=target_index)
+
+    for package in source_pr.list_models():
+        try:
+            params = source_pr.get_raw(package)
+            backup_pr.save_model(*params)
+            target_pr.save_model(*params)
+            logging.info('REINDEXING %s', package)
+        except KeyError:
+            logging.exception('FAILED TO READ DATA FOR %s', package)
+
+    if es.indices.exists_alias(source_index, 'packages'):
+        es.indices.delete_alias(source_index, 'packages')
+
+    # es.indices.delete(source_index)
+    # es.indices.create_alias(target_index, 'packages')
+
+    # Find orphan DB tables
+    used_tables = set()
+    for package in target_pr.list_models():
+        params = target_pr.get_raw(package)
+        fact_table = params[3].get('fact_table') # model
+        if fact_table is not None:
+            used_tables.add(fact_table)
+
+    engine = create_engine(os.environ['OS_CONDUCTOR_ENGINE'])
+    meta = MetaData()
+    meta.reflect(bind=engine,
+                 only=lambda t, _: t.startswith('fdp'))
+    for table in reversed(meta.sorted_tables):
+        if table in used_tables:
+            logging.info('SKIPPING %s', table)
+        else:
+            logging.info('NOT DELETING %s', table)