+# Copyright (C) The Arvados Authors. All rights reserved.
+#
+# SPDX-License-Identifier: Apache-2.0
+
import fnmatch
import os
import errno
import urlparse
import re
import logging
+import threading
import ruamel.yaml as yaml
logger = logging.getLogger('arvados.cwl-runner')
+class CollectionCache(object):
+ def __init__(self, api_client, keep_client, num_retries):
+ self.api_client = api_client
+ self.keep_client = keep_client
+ self.collections = {}
+ self.lock = threading.Lock()
+
+ def get(self, pdh):
+ with self.lock:
+ if pdh not in self.collections:
+ logger.debug("Creating collection reader for %s", pdh)
+ self.collections[pdh] = arvados.collection.CollectionReader(pdh, api_client=self.api_client,
+ keep_client=self.keep_client)
+ return self.collections[pdh]
+
+
class CollectionFsAccess(cwltool.stdfsaccess.StdFsAccess):
"""Implement the cwltool FsAccess interface for Arvados Collections."""
- def __init__(self, basedir, api_client=None, keep_client=None):
+ def __init__(self, basedir, collection_cache=None):
super(CollectionFsAccess, self).__init__(basedir)
- self.api_client = api_client
- self.keep_client = keep_client
- self.collections = {}
+ self.collection_cache = collection_cache
def get_collection(self, path):
sp = path.split("/", 1)
p = sp[0]
if p.startswith("keep:") and arvados.util.keep_locator_pattern.match(p[5:]):
pdh = p[5:]
- if pdh not in self.collections:
- self.collections[pdh] = arvados.collection.CollectionReader(pdh, api_client=self.api_client,
- keep_client=self.keep_client)
- return (self.collections[pdh], sp[1] if len(sp) == 2 else None)
+ return (self.collection_cache.get(pdh), sp[1] if len(sp) == 2 else None)
else:
return (None, path)
def exists(self, fn):
collection, rest = self.get_collection(fn)
- if collection:
+ if collection is not None:
if rest:
return collection.exists(rest)
else:
return os.path.realpath(path)
class CollectionFetcher(DefaultFetcher):
- def __init__(self, cache, session, api_client=None, keep_client=None, num_retries=4):
+ def __init__(self, cache, session, api_client=None, fs_access=None, num_retries=4, overrides=None):
super(CollectionFetcher, self).__init__(cache, session)
self.api_client = api_client
- self.fsaccess = CollectionFsAccess("", api_client=api_client, keep_client=keep_client)
+ self.fsaccess = fs_access
self.num_retries = num_retries
+ self.overrides = overrides if overrides else {}
def fetch_text(self, url):
+ if url in self.overrides:
+ return self.overrides[url]
if url.startswith("keep:"):
with self.fsaccess.open(url, "r") as f:
return f.read()
return super(CollectionFetcher, self).fetch_text(url)
def check_exists(self, url):
+ if url in self.overrides:
+ return True
try:
if url.startswith("http://arvados.org/cwl"):
return True
pipeline_template_uuid_pattern = re.compile(r'[a-z0-9]{5}-p5p6p-[a-z0-9]{15}')
def collectionResolver(api_client, document_loader, uri, num_retries=4):
+ if uri.startswith("keep:") or uri.startswith("arvwf:"):
+ return uri
+
if workflow_uuid_pattern.match(uri):
return "arvwf:%s#main" % (uri)