Code cookbook

Cancel a container request

import arvados
arvados.api().container_requests().update(uuid=container_request_uuid, body={"priority": 0}).execute()

Cancel all container requests

import arvados
api = arvados.api()
result = api.container_requests().list(filters=[["state", "=", "Committed"], ["priority", ">", "0"]]).execute()["items"]
for container_request in result:
    api.container_requests().update(uuid=container_request["uuid"], body={"priority": 0}).execute()

List completed container requests

import arvados
api = arvados.api()
result = api.container_requests().list(filters=[["name", "like", name], ["state", "=", "Final"]]).execute()["items"]
container_uuids = [cr["container_uuid"] for cr in result]
containers = api.containers().list(filters=[["uuid", "in", container_uuids]]).execute()["items"]
container_dict = {c["uuid"]: c for c in containers}

for container_request in result:
    container = container_dict[container_request["container_uuid"]]
    print("%s, %s, %s" % (container_request["uuid"], container_request["name"], "Success" if container["exit_code"] == 0 else "Failed"))

Get input of a CWL workflow

import arvados
api = arvados.api()
container_request_uuid="zzzzz-xvhdp-zzzzzzzzzzzzzzz"
container_request = api.container_requests().get(uuid=container_request_uuid).execute()
print(container_request["mounts"]["/var/lib/cwl/cwl.input.json"])

Get output of a CWL workflow

import arvados
import arvados.collection
api = arvados.api()
container_request_uuid="zzzzz-xvhdp-zzzzzzzzzzzzzzz"
container_request = api.container_requests().get(uuid=container_request_uuid).execute()
collection = arvados.collection.CollectionReader(container_request["output_uuid"])
print(collection.open("cwl.output.json").read())

Get state of a CWL workflow

import arvados
def get_cr_state(cr_uuid):
    api = arvados.api()
    cr = api.container_requests().get(uuid=cr_uuid).execute()
    if cr['container_uuid'] is None:
        return cr['state']
    c = api.containers().get(uuid=cr['container_uuid']).execute()
    if cr['state'] == 'Final' and c['state'] != 'Complete':
        return 'Cancelled'
    elif c['state'] in ['Locked', 'Queued']:
        if c['priority'] == 0:
            return 'On hold'
        else:
            return 'Queued'
    elif c['state'] == 'Complete' and c['exit_code'] != 0:
        return 'Failed'
    elif c['state'] == 'Running':
        if c['runtime_status'].get('error', None):
            return 'Failing'
        elif c['runtime_status'].get('warning', None):
            return 'Warning'
    return c['state']
container_request_uuid = 'zzzzz-xvhdp-zzzzzzzzzzzzzzz'
print(get_cr_state(container_request_uuid))

List input of child requests

import arvados
api = arvados.api()
parent_request_uuid = "zzzzz-xvhdp-zzzzzzzzzzzzzzz"
namefilter = "bwa%"  # the "like" filter uses SQL pattern match syntax
container_request = api.container_requests().get(uuid=parent_request_uuid).execute()
parent_container_uuid = container_request["container_uuid"]
child_requests = api.container_requests().list(filters=[
    ["requesting_container_uuid", "=", parent_container_uuid],
    ["name", "like", namefilter]]).execute()
for c in child_requests["items"]:
    print("%s" % c["name"])
    for m in c["mounts"].values():
        if "portable_data_hash" in m:
            print("  %s" % m["portable_data_hash"])

List output of child requests

import arvados
api = arvados.api()
parent_request_uuid = "zzzzz-xvhdp-zzzzzzzzzzzzzzz"
namefilter = "bwa%"  # the "like" filter uses SQL pattern match syntax
container_request = api.container_requests().get(uuid=parent_request_uuid).execute()
parent_container_uuid = container_request["container_uuid"]
child_requests = api.container_requests().list(filters=[
    ["requesting_container_uuid", "=", parent_container_uuid],
    ["name", "like", namefilter]]).execute()
output_uuids = [c["output_uuid"] for c in child_requests["items"]]
collections = api.collections().list(filters=[["uuid", "in", output_uuids]]).execute()
uuid_to_pdh = {c["uuid"]: c["portable_data_hash"] for c in collections["items"]}
for c in child_requests["items"]:
    print("%s -> %s" % (c["name"], uuid_to_pdh[c["output_uuid"]]))

List failed child requests

import arvados
api = arvados.api()
parent_request_uuid = "zzzzz-xvhdp-zzzzzzzzzzzzzzz"
container_request = api.container_requests().get(uuid=parent_request_uuid).execute()
parent_container_uuid = container_request["container_uuid"]
child_requests = api.container_requests().list(filters=[
    ["requesting_container_uuid", "=", parent_container_uuid]], limit=1000).execute()
child_containers = {c["container_uuid"]: c for c in child_requests["items"]}
cancelled_child_containers = api.containers().list(filters=[
    ["exit_code", "!=", "0"],
    ["uuid", "in", list(child_containers.keys())]], limit=1000).execute()
for c in cancelled_child_containers["items"]:
    print("%s (%s)" % (child_containers[c["uuid"]]["name"], child_containers[c["uuid"]]["uuid"]))

Get log of a child request

import arvados
import arvados.collection
api = arvados.api()
container_request_uuid = "zzzzz-xvhdp-zzzzzzzzzzzzzzz"
container_request = api.container_requests().get(uuid=container_request_uuid).execute()
collection = arvados.collection.CollectionReader(container_request["log_uuid"])
for c in collection:
    if isinstance(collection.find(c), arvados.arvfile.ArvadosFile):
        print(collection.open(c).read())

Create a collection sharing link

import arvados
api = arvados.api()
download="https://your.download.server"
collection_uuid="zzzzz-4zz18-zzzzzzzzzzzzzzz"
token = api.api_client_authorizations().create(body={"api_client_authorization":{"scopes": [
    "GET /arvados/v1/collections/%s" % collection_uuid,
    "GET /arvados/v1/collections/%s/" % collection_uuid,
    "GET /arvados/v1/keep_services/accessible"]}}).execute()
print("%s/c=%s/t=%s/_/" % (download, collection_uuid, token["api_token"]))

Combine two or more collections

Note, if two collections have files of the same name, the contents will be concatenated in the resulting manifest.

import arvados
import arvados.collection
api = arvados.api()
project_uuid = "zzzzz-tpzed-zzzzzzzzzzzzzzz"
collection_uuids = ["zzzzz-4zz18-aaaaaaaaaaaaaaa", "zzzzz-4zz18-bbbbbbbbbbbbbbb"]
combined_manifest = ""
for u in collection_uuids:
    c = api.collections().get(uuid=u).execute()
    combined_manifest += c["manifest_text"]
newcol = arvados.collection.Collection(combined_manifest)
newcol.save_new(name="My combined collection", owner_uuid=project_uuid)

Upload a file into a new collection

import arvados
import arvados.collection

project_uuid = "zzzzz-j7d0g-zzzzzzzzzzzzzzz"
collection_name = "My collection"
filename = "file1.txt"

api = arvados.api()
c = arvados.collection.Collection()
with open(filename, "rb") as reader:
    with c.open(filename, "wb") as writer:
        content = reader.read(128*1024)
        while content:
            writer.write(content)
            content = reader.read(128*1024)
c.save_new(name=collection_name, owner_uuid=project_uuid)
print("Saved %s to %s" % (collection_name, c.manifest_locator()))

Download a file from a collection

import arvados
import arvados.collection

collection_uuid = "zzzzz-4zz18-zzzzzzzzzzzzzzz"
filename = "file1.txt"

api = arvados.api()
c = arvados.collection.CollectionReader(collection_uuid)
with c.open(filename, "rb") as reader:
    with open(filename, "wb") as writer:
        content = reader.read(128*1024)
        while content:
            writer.write(content)
            content = reader.read(128*1024)
print("Finished downloading %s" % filename)

Copy files from a collection to a new collection

import arvados.collection

source_collection = "zzzzz-4zz18-zzzzzzzzzzzzzzz"
target_project = "zzzzz-j7d0g-zzzzzzzzzzzzzzz"
target_name = "Files copied from source_collection"
files_to_copy = ["folder1/sample1/sample1_R1.fastq",
                 "folder1/sample2/sample2_R1.fastq"]

source = arvados.collection.CollectionReader(source_collection)
target = arvados.collection.Collection()

for f in files_to_copy:
    target.copy(f, "", source_collection=source)

target.save_new(name=target_name, owner_uuid=target_project)
print("Created collection %s" % target.manifest_locator())

Copy files from a collection to another collection

import arvados.collection

source_collection = "zzzzz-4zz18-zzzzzzzzzzzzzzz"
target_collection = "zzzzz-4zz18-aaaaaaaaaaaaaaa"
files_to_copy = ["folder1/sample1/sample1_R1.fastq",
                 "folder1/sample2/sample2_R1.fastq"]

source = arvados.collection.CollectionReader(source_collection)
target = arvados.collection.Collection(target_collection)

for f in files_to_copy:
    target.copy(f, "", source_collection=source)

target.save()

Delete a file from an existing collection

import arvados

c = arvados.collection.Collection("zzzzz-4zz18-zzzzzzzzzzzzzzz")
c.remove("file2.txt")
c.save()

Listing records with paging

Use the arvados.util.keyset_list_all helper method to iterate over all the records matching an optional filter. This method handles paging internally and returns results incrementally using a Python iterator. The first parameter of the method takes a list method of an Arvados resource (collections, container_requests, etc).

import arvados.util

api = arvados.api()
for c in arvados.util.keyset_list_all(api.collections().list, filters=[["name", "like", "%sample123%"]]):
    print("got collection " + c["uuid"])

Previous: Arvados CWL Runner Next: Subscribing to database events