Release Notes 2.10 (March 2023)

Lineage Graph

• With the lineage graph, you can find out how the results were generated and whether they are reproducible.

• See exactly which capsules, pipelines, and data were used to produce a particular Results data asset while also showing you if/how the final Results can be reproduced.

• If a data asset can’t be reproduced, you’ll see exactly which step of the process or requisite data are no longer available.

Metadata Search UI

In order to allow data to be findable with respect to FAIR principles, there is now a UI that allows users to find Data Assets based on certain Custom Metadata fields.

This can be done via the “Metadata Filters” button in the filters row. Through there, users can toggle certain metadata fields and enter certain keywords. Datasets appearing under the filter have that field filled out and contain the keyword

Support for GPUs in Pipelines

• When using Capsule configured to utilize GPU instances, GPU resource selection is now available in pipelines allowing parallelization or acceleration benefits present in the original capsule.

• Now, in the resources panel, both the number of GPUs, the number of cores and the number of GB RAM is available to increase/decrease

Ability to Release a Pipeline

• A Released Pipeline allows you to provide a fully functioning, end to end reproducible, stable version of your pipeline to users while work continues on the original.

• For pipelines with multiple released versions, users are able to switch seamlessly between any available version of the released pipeline by selecting the corresponding version from the drop down menu.

• Each released pipeline will be automatically shared with the same users who have access to the original capsule.

Concurrent Runs for Dedicated Machines

Concurrent Runs has been supported in Code Ocean for the flex machine and we now extended the feature to accommodate the dedicated machine as well. This feature allows users to run a capsule multiple times at the same time.

R Shiny Capsules Open on /code folder

When opening a R Shiny in the cloud workstation, the starting page will shows the content on the /code folder instead of the /capsule folder. This way, users won’t see the rest of the folder in the capsule and have a more native experience of R Shiny app.

Run Release Capsule by API Improvements

The run capsule API no longer requires a version number to run a release capsule. When a version number isn’t specified the latest release version will be used

Environment Variables in Computations

Code Ocean specific environment variables are now available in computations to reduce coding complexity.

• CO_CPUS: the number of available CPU cores

• CO_MEMORY: the available RAM in bytes

• CO_COMPUTATION_ID: the unique identifier of the current computation

• CO_CAPSULE_ID/CO_PIPELINE_ID: the capsule or pipeline’s unique identifier, also available in the metadata page

  • In a pipeline, each capsule will have its own CO_CPUS, CO_MEMORY, and CO_CAPSULE_ID variables but they will all share the same CO_PIPELINE_ID and CO_COMPUTATION_ID

  • CO_CPUS and CO_MEMORY are particularly useful in pipelines where typical approaches to get available CPU and memory counts are inaccurate due to the use of AWS batch resources

Data Asset Creation Improvement

• To improve the traceability of data asset sources when created from an S3 bucket, there is a new “Source” block in Data Asset details

• When viewing the contents of an imported/linked S3 bucket data asset, the original s3 bucket of the data source as well as the relative path to a subdirectory (if contents are not at the root of the bucket) are viewable.