automation_skill
- Shell
0
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill delphine-l/claude_global --skill automation- SKILL.md22.2 KB
Overview
This skill provides practical expertise for automating Galaxy workflows using BioBlend and Planemo. It covers connecting to Galaxy, managing histories and datasets, invoking workflows, polling status, handling errors, rerunning failed invocations, and driving Planemo runs for external Galaxy servers. It is designed for reproducible, scalable Galaxy automation in bioinformatics pipelines.
How this skill works
The skill shows how to create and reuse GalaxyInstance connections via BioBlend, programmatically upload datasets and build collections, invoke workflows with prepared inputs, and poll invocation and dataset states until completion. It also describes patterns for error classification, automated reruns using Galaxy's rerun API, and generating/executing Planemo commands to submit jobs from the command line or scripts. Concurrency controls and Planemo output parsing are included to integrate these steps into batch pipelines.
When to use it
- Automating repeated Galaxy analyses or batch processing many samples
- Integrating Galaxy workflow runs into CI/CD or larger pipelines
- Programmatic management of histories, datasets, and collections
- Robustly invoking workflows and polling for completion/status
- Testing or launching workflows on an external Galaxy using Planemo
Best practices
- Store API keys in environment variables and never hard-code them
- Use HTTPS and mask API keys in logs and displays
- Use thread locks or concurrency limits to avoid API contention
- Poll invocations/datasets with timeouts and clear error reporting
- Categorize failures (memory, timeout, network) to decide retries
- Prefer Galaxy rerun API to reuse successful jobs when possible
Example use cases
- Batch-run a paired-end RNA-seq workflow for hundreds of samples with list:paired collections
- Automate dataset uploads, create histories, invoke a workflow, and wait for completion before downstream steps
- Use Planemo to launch workflows on an external Galaxy from a CI job and capture invocation JSON
- Detect transient job failures (e.g., OOM) and automatically rerun with adjusted resources
- Build a monitoring tool that polls invocation steps and extracts failed job stderr for diagnostics
FAQ
os.system() return codes are shifted; recover the actual exit code with return_code >> 8.
When should I rerun a failed invocation versus inspect manually?
If failures match retriable patterns (out of memory, timeout, network) automate a rerun; for unknown or repeated permanent failures inspect logs and job stderr first.
