From 2a0472e6b0e94265d78fb22b5652619a7700e86b Mon Sep 17 00:00:00 2001
From: Peter Amstutz <peter.amstutz@curoverse.com>
Date: Thu, 27 Oct 2016 21:54:14 -0400
Subject: [PATCH] 9849: Cache docker lookup.

---
 sdk/cwl/arvados_cwl/arvdocker.py | 21 ++++++++++++++++++++-
 sdk/cwl/tests/test_container.py  |  4 ++++
 2 files changed, 24 insertions(+), 1 deletion(-)

diff --git a/sdk/cwl/arvados_cwl/arvdocker.py b/sdk/cwl/arvados_cwl/arvdocker.py
index 244dd3d357..f9dba6eff4 100644
--- a/sdk/cwl/arvados_cwl/arvdocker.py
+++ b/sdk/cwl/arvados_cwl/arvdocker.py
@@ -1,19 +1,28 @@
 import logging
 import sys
+import threading
 
 import cwltool.docker
 from cwltool.errors import WorkflowException
 import arvados.commands.keepdocker
 
-
 logger = logging.getLogger('arvados.cwl-runner')
 
+cached_lookups = {}
+cached_lookups_lock = threading.Lock()
+
 def arv_docker_get_image(api_client, dockerRequirement, pull_image, project_uuid):
     """Check if a Docker image is available in Keep, if not, upload it using arv-keepdocker."""
 
     if "dockerImageId" not in dockerRequirement and "dockerPull" in dockerRequirement:
         dockerRequirement["dockerImageId"] = dockerRequirement["dockerPull"]
 
+    global cached_lookups
+    global cached_lookups_lock
+    with cached_lookups_lock:
+        if dockerRequirement["dockerImageId"] in cached_lookups:
+            return cached_lookups[dockerRequirement["dockerImageId"]]
+
     sp = dockerRequirement["dockerImageId"].split(":")
     image_name = sp[0]
     image_tag = sp[1] if len(sp) > 1 else None
@@ -44,4 +53,14 @@ def arv_docker_get_image(api_client, dockerRequirement, pull_image, project_uuid
         raise WorkflowException("Could not find Docker image %s:%s" % (image_name, image_tag))
 
     pdh = api_client.collections().get(uuid=images[0][0]).execute()["portable_data_hash"]
+
+    with cached_lookups_lock:
+        cached_lookups[dockerRequirement["dockerImageId"]] = pdh
+
     return pdh
+
+def arv_docker_clear_cache():
+    global cached_lookups
+    global cached_lookups_lock
+    with cached_lookups_lock:
+        cached_lookups = {}
diff --git a/sdk/cwl/tests/test_container.py b/sdk/cwl/tests/test_container.py
index 8ca5398e82..93100ae9f7 100644
--- a/sdk/cwl/tests/test_container.py
+++ b/sdk/cwl/tests/test_container.py
@@ -1,4 +1,5 @@
 import arvados_cwl
+from arvados_cwl.arvdocker import arv_docker_clear_cache
 import logging
 import mock
 import unittest
@@ -21,6 +22,8 @@ class TestContainer(unittest.TestCase):
     @mock.patch("arvados.commands.keepdocker.list_images_in_arv")
     def test_run(self, keepdocker):
         for enable_reuse in (True, False):
+            arv_docker_clear_cache()
+
             runner = mock.MagicMock()
             runner.project_uuid = "zzzzz-8i9sb-zzzzzzzzzzzzzzz"
             runner.ignore_docker_for_reuse = False
@@ -72,6 +75,7 @@ class TestContainer(unittest.TestCase):
     # For the remaining fields, the defaults will apply: {'cores': 1, 'ram': 1024, 'outdirSize': 1024, 'tmpdirSize': 1024}
     @mock.patch("arvados.commands.keepdocker.list_images_in_arv")
     def test_resource_requirements(self, keepdocker):
+        arv_docker_clear_cache()
         runner = mock.MagicMock()
         runner.project_uuid = "zzzzz-8i9sb-zzzzzzzzzzzzzzz"
         runner.ignore_docker_for_reuse = False
-- 
2.39.5