From 0b90b550d157b5e8ef6f3115372798994940eff5 Mon Sep 17 00:00:00 2001
From: Mathieu Leplatre <mathieu@mozilla.com>
Date: Fri, 9 Jun 2017 17:51:45 +0200
Subject: [PATCH 1/2] Limit the number of results returned by default (fixes
 #45)

---
 CHANGELOG.rst                  |  3 ++-
 kinto_elasticsearch/indexer.py |  4 ++--
 kinto_elasticsearch/views.py   |  7 +++++++
 tests/test_elasticsearch.py    | 38 ++++++++++++++++++++++++++++++++++
 4 files changed, 49 insertions(+), 3 deletions(-)

diff --git a/CHANGELOG.rst b/CHANGELOG.rst
index d258b2c..4ee0f89 100644
--- a/CHANGELOG.rst
+++ b/CHANGELOG.rst
@@ -5,8 +5,9 @@ Changelog
 0.2.0 (unreleased)
 ------------------
 
-- Nothing changed yet.
+**Bug fixes**
 
+- Limit the number of results returned by default (fixes #45)
 
 0.1.0 (2017-05-26)
 ------------------
diff --git a/kinto_elasticsearch/indexer.py b/kinto_elasticsearch/indexer.py
index 1291579..d17e0b9 100644
--- a/kinto_elasticsearch/indexer.py
+++ b/kinto_elasticsearch/indexer.py
@@ -26,9 +26,9 @@ def create_index(self, bucket_id, collection_id, schema=None):
                 body = {"mappings": {indexname: schema}}
             else:
                 body = None
-            self.client.indices.create(index=indexname, body=body)
+            return self.client.indices.create(index=indexname, body=body)
         else:
-            self.update_index(bucket_id, collection_id, schema)
+            return self.update_index(bucket_id, collection_id, schema)
 
     def update_index(self, bucket_id, collection_id, schema=None):
         indexname = self.indexname(bucket_id, collection_id)
diff --git a/kinto_elasticsearch/views.py b/kinto_elasticsearch/views.py
index f16f298..ae8381b 100644
--- a/kinto_elasticsearch/views.py
+++ b/kinto_elasticsearch/views.py
@@ -29,6 +29,13 @@ def search_view(request, **kwargs):
     bucket_id = request.matchdict['bucket_id']
     collection_id = request.matchdict['collection_id']
 
+    # # Limit the number of results to return, based on existing Kinto settings.
+    paginate_by = request.registry.settings.get("paginate_by")
+    max_fetch_size = request.registry.settings["storage_max_fetch_size"]
+    if paginate_by is None or paginate_by <= 0:
+        paginate_by = max_fetch_size
+    kwargs.setdefault("size", min(paginate_by, max_fetch_size))
+
     # Access indexer from views using registry.
     indexer = request.registry.indexer
     try:
diff --git a/tests/test_elasticsearch.py b/tests/test_elasticsearch.py
index 245faa4..66df390 100644
--- a/tests/test_elasticsearch.py
+++ b/tests/test_elasticsearch.py
@@ -196,6 +196,44 @@ def test_empty_querystring_returns_all_results(self):
         assert len(result["hits"]["hits"]) == 2
 
 
+class LimitedResults(BaseWebTest, unittest.TestCase):
+    def get_app(self, settings):
+        app = self.make_app(settings=settings)
+        app.put("/buckets/bid", headers=self.headers)
+        app.put("/buckets/bid/collections/cid", headers=self.headers)
+        requests = [{
+            "method": "POST",
+            "path": "/buckets/bid/collections/cid/records",
+            "body": {"data": {"age": i}}
+        } for i in range(5)]
+        app.post_json("/batch", {"requests": requests}, headers=self.headers)
+        return app
+
+    def test_the_number_of_responses_is_limited_by_paginate_by_setting(self):
+        app = self.get_app({"paginate_by": 2})
+        resp = app.get("/buckets/bid/collections/cid/search", headers=self.headers)
+        result = resp.json
+        assert len(result["hits"]["hits"]) == 2
+
+    def test_the_number_of_responses_is_limited_by_max_fetch_size_setting(self):
+        app = self.get_app({"storage_max_fetch_size": 2})
+        resp = app.get("/buckets/bid/collections/cid/search", headers=self.headers)
+        result = resp.json
+        assert len(result["hits"]["hits"]) == 2
+
+    def test_the_number_of_responses_is_limited_by_smaller_limit(self):
+        app = self.get_app({"paginate_by": 4, "storage_max_fetch_size": 2})
+        resp = app.get("/buckets/bid/collections/cid/search", headers=self.headers)
+        result = resp.json
+        assert len(result["hits"]["hits"]) == 2
+
+    def test_the_number_of_responses_is_limited_by_only_defined_limit(self):
+        app = self.get_app({"paginate_by": 0, "storage_max_fetch_size": 2})
+        resp = app.get("/buckets/bid/collections/cid/search", headers=self.headers)
+        result = resp.json
+        assert len(result["hits"]["hits"]) == 2
+
+
 class PermissionsCheck(BaseWebTest, unittest.TestCase):
     def test_search_is_allowed_if_write_on_bucket(self):
         body = {"permissions": {"write": ["system.Everyone"]}}

From 22db03c20e086eaf484bbc37b97fc20eb5eafec1 Mon Sep 17 00:00:00 2001
From: Mathieu Leplatre <mathieu@mozilla.com>
Date: Tue, 13 Jun 2017 15:08:34 +0200
Subject: [PATCH 2/2] @Natim review

---
 kinto_elasticsearch/views.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kinto_elasticsearch/views.py b/kinto_elasticsearch/views.py
index ae8381b..46b1a45 100644
--- a/kinto_elasticsearch/views.py
+++ b/kinto_elasticsearch/views.py
@@ -29,7 +29,7 @@ def search_view(request, **kwargs):
     bucket_id = request.matchdict['bucket_id']
     collection_id = request.matchdict['collection_id']
 
-    # # Limit the number of results to return, based on existing Kinto settings.
+    # Limit the number of results to return, based on existing Kinto settings.
     paginate_by = request.registry.settings.get("paginate_by")
     max_fetch_size = request.registry.settings["storage_max_fetch_size"]
     if paginate_by is None or paginate_by <= 0: