This repository contains GitHub Actions workflows for automated BED file uploads to BEDbase.
For complete documentation on the data loading architecture, see: BEDbase Data Loading
Accbase workflows queue ATAC-seq samples for downstream processing on Rivanna HPC.
| Workflow | Schedule | Purpose |
|---|---|---|
accbase_queue_cron.yml |
Daily 19:00 UTC | Queue new samples from accbase namespace |
accbase_queue_gse.yml |
Manual | Queue samples from a specific GSE |
Unlike BEDbase (which processes files in GitHub Actions + Fargate), Accbase only
queues samples in the sample_queue table. Heavy processing (PEPATAC pipeline)
runs on Rivanna HPC.
Configuration: accbase_config.yaml
For Accbase workflows:
ACCBASE_POSTGRES_HOST- PostgreSQL host for accbaseACCBASE_POSTGRES_USER- Database userACCBASE_POSTGRES_PASSWORD- Database passwordPEPHUB_API_KEY- Optional, for higher rate limits
The sample_queue table is consumed on Rivanna:
- Consumer queries for samples with status='pending'
- Generates looper submission from PEPhub PEP
- Submits SLURM jobs via looper
- Updates status to 'processing'
- On completion, updates status to 'completed'
This decouples data discovery (GitHub Actions) from heavy processing (Rivanna HPC).