Skip to content

CHANGELOG.md

Latest commit

 

History

History
114 lines (89 loc) · 4.72 KB

CHANGELOG.md

File metadata and controls

114 lines (89 loc) · 4.72 KB

v3.0.1

  • Added short options for projark subcommands (#45, @kopardev):
    • common: -f (--folder), -p (--projectnumber), -d (--datatype)
    • deposit: -t (--tarname), -s (--split-size-gb), -k (--no-cleanup)
    • retrieve: -n (--filenames), -u (--unsplit/--unspilt)
  • Updated projectnumber normalization (#44, @kopardev):
    • remove repeated leading ccbr/CCBR prefixes (with optional _/-)
    • accept any non-empty remainder (for example CCBR-abcd -> abcd)
  • Added absolute-path normalization for --folder handling in projark (#46 @kopardev):
    • relative paths are resolved to absolute paths before use
    • trailing slash/non-trailing slash inputs are both supported
  • Hardened tar command construction with shell-safe quoting for paths containing spaces/special characters.
  • Added ISO 8601 timestamps to projark log lines. (#47, @kopardev)
  • Added completion/failure email notifications for projark (#48, @kopardev):
    • recipient: $USER@nih.gov
    • sender: NCICCBR@mail.nih.gov
  • Added Open OnDemand graphical session detection in runtime checks (future-facing; projark still runs only on Helix today).
  • Updated README/MkDocs docs with:
    • new short-option usage examples
    • current projectnumber behavior
    • path normalization behavior
    • timestamped logging and notification behavior
    • Open OnDemand availability disclaimer (Biowulf-only today; not directly available on Helix)

v3.0.0

  • Replaced legacy bash projark workflow with a Python-native projark CLI (deposit, retrieve) integrated into the parkit package.
  • Archived old bash implementation at legacy/projark_legacy.sh; src/parkit/scripts/projark now points to the Python CLI.
  • Added projark deposit end-to-end archival flow with:
    • checkapisync preflight gate
    • Helix host enforcement
    • tmux/screen session enforcement
    • project/datatype normalization
    • scratch staging, tar/filelist generation, md5 generation
    • transfer via dm_register_directory
    • cleanup enabled by default (--no-cleanup to retain artifacts)
  • Added configurable tar split threshold/chunk size for deposit: --split-size-gb (default 500 GB).
  • Added human-readable tar size reporting (MB/GB/TB + bytes) in projark output.
  • Added projark retrieve enhancements:
    • selected-file retrieval with --filenames
    • full-collection retrieval when --filenames is omitted (dm_download_collection)
    • --unsplit/--unspilt merge support for multiple split tar groups
  • Improved projark CLI output with stepwise status messages and consistent step numbering.
  • Added subcommand version support:
    • projark --version
    • projark deposit --version
    • projark retrieve --version
    • all return the same package-aware message
  • Suppressed bootstrap Java/environment warnings for version-only invocations.
  • Updated checkapisync logic to treat merge-history-only divergence as in-sync when local/upstream trees match.
  • Reworked docs to versioned MkDocs structure with projark-first workflows and updated operational guidance.
  • Updated docs/README guidance:
    • use mamba activate ... directly
    • initialize mamba only if not already in PATH
    • reference HPC_DME setup guide
    • document minimum Java requirement: HPC_DM_JAVA_VERSION >= 23
    • standardize guidance to run all operations in tmux/screen
  • Documentation: Improved code example readability in README. (#34, @kelly-sovacool)

v2.2.0

  • Feature: Automatic object name resolution (fix #30)
    • Added logic to detect and resolve naming conflicts by appending _002, _003, etc. to object names during deposit.
    • Metadata .json and .filelist files are updated accordingly to reflect new names.
  • Enhancement: Java version management (fix #31)
    • Introduced support for setting HPC_DM_JAVA_VERSION dynamically with a default fallback to 23.0.2 on Biowulf/Helix
  • Fixed ccr partition issue (fix #32)

v2.1.3

  • fix scontrol issue on HELIX

v2.1.1

Features/BugFixes

  • fix loading parkit_dev conda env (#26)

v2.1.0

Features/BugFixes

  • adding collectiontype to createmetadata command
  • excluding *.tar.filelist from --cleanup as it may need be put alongside README
  • create readme with --makereadme argument
  • projark subcommand for easy CCBR project backup

v2.0.1

Features/BugFixes

  • package renamed from parkit_pkg to parkit
  • reorganized package by recreating src folder

v2.0.0

Features/BugFixes

  • converted to python package (fix #4)

v1.0.1

Features/BugFixes

  • parkit_tarball2hpcdme added to start with tarballs
  • update_collection_metadata.py added to add more information from AMP dump

v1.0

Features/BugFixes

  • parkit added... built using python
  • parkit_e2e added ... uses parkit and "slurm"