18870: Need to declare NODES as array

[arvados.git] / sdk / cwl / arvados_cwl / http.py

diff --git a/sdk/cwl/arvados_cwl/http.py b/sdk/cwl/arvados_cwl/http.py
index 32fc1cf90de09dbfa7e357906fda3a1641102d83..dcc2a51192dfc4d4b573da302b3373fd08d67fff 100644 (file)
--- a/sdk/cwl/arvados_cwl/http.py
+++ b/sdk/cwl/arvados_cwl/http.py
@@ -1,3 +1,11 @@
+# Copyright (C) The Arvados Authors. All rights reserved.
+#
+# SPDX-License-Identifier: Apache-2.0
+
+from __future__ import division
+from future import standard_library
+standard_library.install_aliases()
+
 import requests
 import email.utils
 import time
@@ -5,19 +13,26 @@ import datetime
 import re
 import arvados
 import arvados.collection
-import urlparse
+import urllib.parse
 import logging
+import calendar
+import urllib.parse
 
 logger = logging.getLogger('arvados.cwl-runner')
 
 def my_formatdate(dt):
-    return email.utils.formatdate(timeval=time.mktime(dt.timetuple()),
+    return email.utils.formatdate(timeval=calendar.timegm(dt.timetuple()),
                                   localtime=False, usegmt=True)
 
 def my_parsedate(text):
-    parsed = email.utils.parsedate(text)
+    parsed = email.utils.parsedate_tz(text)
     if parsed:
-        return datetime.datetime(*parsed[:6])
+        if parsed[9]:
+            # Adjust to UTC
+            return datetime.datetime(*parsed[:6]) + datetime.timedelta(seconds=parsed[9])
+        else:
+            # TZ is zero or missing, assume UTC.
+            return datetime.datetime(*parsed[:6])
     else:
         return datetime.datetime(1970, 1, 1)
 
@@ -81,13 +96,13 @@ def http_to_keep(api, project_uuid, url, utcnow=datetime.datetime.utcnow):
         if fresh_cache(url, properties, now):
             # Do nothing
             cr = arvados.collection.CollectionReader(item["portable_data_hash"], api_client=api)
-            return "keep:%s/%s" % (item["portable_data_hash"], cr.keys()[0])
+            return "keep:%s/%s" % (item["portable_data_hash"], list(cr.keys())[0])
 
         if not changed(url, properties, now):
             # ETag didn't change, same content, just update headers
             api.collections().update(uuid=item["uuid"], body={"collection":{"properties": properties}}).execute()
             cr = arvados.collection.CollectionReader(item["portable_data_hash"], api_client=api)
-            return "keep:%s/%s" % (item["portable_data_hash"], cr.keys()[0])
+            return "keep:%s/%s" % (item["portable_data_hash"], list(cr.keys())[0])
 
     properties = {}
     req = requests.get(url, stream=True, allow_redirects=True)
@@ -113,28 +128,30 @@ def http_to_keep(api, project_uuid, url, utcnow=datetime.datetime.utcnow):
         else:
             name = grp.group(4)
     else:
-        name = urlparse.urlparse(url).path.split("/")[-1]
+        name = urllib.parse.urlparse(url).path.split("/")[-1]
 
     count = 0
     start = time.time()
     checkpoint = start
-    with c.open(name, "w") as f:
+    with c.open(name, "wb") as f:
         for chunk in req.iter_content(chunk_size=1024):
             count += len(chunk)
             f.write(chunk)
             loopnow = time.time()
             if (loopnow - checkpoint) > 20:
-                bps = (float(count)/float(loopnow - start))
+                bps = count / (loopnow - start)
                 if cl is not None:
                     logger.info("%2.1f%% complete, %3.2f MiB/s, %1.0f seconds left",
-                                float(count * 100) / float(cl),
-                                bps/(1024*1024),
-                                (cl-count)/bps)
+                                ((count * 100) / cl),
+                                (bps // (1024*1024)),
+                                ((cl-count) // bps))
                 else:
-                    logger.info("%d downloaded, %3.2f MiB/s", count, bps/(1024*1024))
+                    logger.info("%d downloaded, %3.2f MiB/s", count, (bps / (1024*1024)))
                 checkpoint = loopnow
 
-    c.save_new(name="Downloaded from %s" % url, owner_uuid=project_uuid, ensure_unique_name=True)
+
+    collectionname = "Downloaded from %s" % urllib.parse.quote(url, safe='')
+    c.save_new(name=collectionname, owner_uuid=project_uuid, ensure_unique_name=True)
 
     api.collections().update(uuid=c.manifest_locator(), body={"collection":{"properties": properties}}).execute()