v0.3.1

Added a task graph.

• Pipeline developers: See this doc how to update an output definition JSON file for your pipeline.

v0.2.0

IMPORTANT: Upgrade Caper to >=0.5.0 first.

$ pip install caper==0.5.0

Deprecated parameters

• A flag --use-rel-path-in-link: we no longer need this since we can get a presigned/converted URL for both a file table and a UCSC browser track.

Presigning for cloud URIs (gs://, s3://)

• Use a flag --use-presigned-url-gcs and --gcp-private-key [KEY_FILE] to presign GCS URIs.
• Use a flag --public-gcs if your GCS bucket is public. Don't need presigning since it's already public.
• Use a flag --use-presigned-url-s3 to presign S3 URIs.
• Use --tsv-mapping-path-to-url [TSV_FILE] to define mapping of local path prefix to a URL prefix. You may need to have a web server and organized outputs should be (symlinked) on there.

Added USCS genome browser tracks based on presigned/converted URLs

• Use --ucsc-genome-db [GENOME] to define genome DB for UCSC track (e.g. hg38)
• Use --ucsc-genome-pos chr?:?????-?????? to define initial genome coord for tracks.

v0.1.8

Update for sub-workflows without key submittedFiles in its metadata JSON file.

v0.1.7

• allow Array[File] and Map[?, File] type for task's output

v0.1.6

• --method link (default method) works for cloud storages
  • don't make an organized directory on --out-dir if --out-dir's strorage is same as the source storage.
  • e.g. outputs files in metadata.json is on gs:// and --out-dir is also gs://
  • this will make an empty --out-dir and there will be only one HTML report file
  • users can explicitly define --method copy to make actual copies on --out-dir

v0.1.5

• bug fix for outputs key error
• for the contents in metadata.json file, Croo can take in a list of Python dicts [ { } ] but only the first [0] dict object { } will be taken. Others will be ignored.

v0.1.3

• html report
  • filetable
• fix bug in importing croo

v0.1.1

• PIP support

$ pip install croo

v0.1.0

First BETA release