0.13.1

• Lambda service error can now be captured and retried by the step function.
• Changes require redeployment of tibanna as below:

# run the following commands in the order
tibanna deploy_unicorn -b <buckets> -g <usergroup>
tibanna deploy_core -n run_task_awsem -g <usergroup>
tibanna deploy_core -n check_task_awsem -g <usergroup>

• A minor fix for an irrelevant error message as below when buckets were not set properly during deployment.

local variable 'res' referenced before assignment

0.12.1

• Benchmark-4dn dependency version loosened (==0.5.8 -> >=0.5.8)

0.12.0

• availability zone can now be specified in the input json file.
  • "config" : { ... "availability_zone": "us-east-1a"}

0.11.3

• Failure to obtain input file size from lambda no longer throws an error, unless it needs to be used for ebs_size calculation (if user specifies it as a multiplier (e.g. 3x)) or if benchmarking is used (i.e. the workflow is a 4DN workflow that has a benchmarking function associated with it)
  • This is useful for cases with complex external S3 permission (e.g. lambda function does not have permission to the external bucket)

0.11.2

• More informative error message when AWS credentials are not found.

0.11.1

• cloudwatch_dashboard field in the execution input json is disabled (needs redeployment for it to take effect). Use tibanna plot_metrics command instead.

0.11.0

• A recently introduced critical check_task permission error for runs longer than 5min is now fixed. In order to for this change to take effect, a tibanna unicorn must be redeployed - to overwrite an existing unicorn, it may need to be redeployed twice (due to an unknown AWS-related problem).

• Input bucket mounting support

  • The option can be set by specifying "mount": True for individual input file item. If "mount" is not specified or specified False, the file will be downloaded instead (as before).
  • The mount vs download option is set for individual file item and even if two or more files share the same bucket, one can choose one or the other for each file. Mounting reduces EBS volume size, whereas it is usually much faster to read a downloaded file.
  • For this to take effect, a tibanna unicorn must be redeployed.

• Environment variables can be passed for languages shell and snakemake.

  • They can be specified in the format of "input_env" : {"ENVNAME1": "1234", "ENVNAME2": "5678"} in the args field of the input execution json.
  • In the command, the environment variable should be used with double backslash (e.g. $NAME should be written as \\$NAME)

0.10.2

• Benchmark-4dn version 0.5.7->0.5.8 (This change affects only the list of 4DN pipelines with a benchmark function available)

0.10.1

• The problem of EC2UnintendedTerminationException and EC2IdleException not working is now fixed
  • Basically, when an AWSEM EC2 instance for some reason goes into sleep for more than an hour (Idle) or terminated for an unknown reason (unintendedtermination), Tibanna detects and throws these errors. In the former case, it also kills the sleeping instance.
• AMI support for regions other than us-east-1
• Dependency versions (python-lambda-4dn and benchmark-4dn) adjusted.

0.10.0

This repo and pypi package tibanna now contains only the unicorn tibanna components and no longer contains 4DN/CGAP components (pony/zebra) (no change for the unicorn and therefore no change for users other than the 4DN/CGAP developer team).

In other words, the following import / entry points are no longer available from this repo / tibanna package, but now has moved to a different repo and pypi package named tibanna_ff.

tibanna_4dn -h
tibanna_cgap -h

import tibanna_4dn
import tibanna_cgap
import tibanna_ffcommon  # 4dn/cgap shared components that are not part of unicorn