errorStrategy "skip"

New feature

I wanted to propose a new errorStrategy (skip?) that, if the process fails or keeps failing, continues the pipeline without dropping the sample and forwards the input directly as output.

Use case

Let's say I have ONT data and don't know whether it still contains barcodes or adapters. If I run a tool like porechop_abi, it will detect them if they are present, but fail/crash if they are not. Imagine I have data from multiple sources, some with and some without an adapter. It would be nice to have the ability to run porechop_abi on all, and, in case of failures, still move forward with the input as output.

Currently, errorStrategy ignore allows the pipeline to continue, but will drop the sample that fails. I'd like something that, in case of failure, skips the steps and moves forward directly without dropping the specific sample.

Suggested implementation

If the process fails/error, then forward the channel input as output?

No files or tests are named. Start by tracing the existing errorStrategy "ignore" behavior, then determine how a failed process could forward its input without dropping the sample; done means a defined strategy that preserves failed inputs while allowing the pipeline to continue.