Release version 0.3.0-alpha.2

Updates in this version:

• xarray version 0.16.0 (#134)
• intake-esm upgraded (#134)
• worker-template.yml always overwritten on boot (#130)

Release version v0.3.0-alpha.1

Major image changes

Package updates

• Interface & hub: jupyterhub (1.0.0), jupyterlab (1.2.1), notebook (6.0.0)
• distributed computing: dask (2.8.1), dask_kubernetes (master)
• data science: xarray (0.14.1), pandas (0.25.3), scipy (1.3.2), zarr (2.3.2), netcdf4 (1.5.3), xesmf (0.2.1), xgcm (0.2.0), NCAR/intake-esm (master)
• data vis: geoviews (1.6.6), holoviews (1.12.7), datashader (0.8.0), plotly (4.3.0), seaborn (0.9.0)

New environments

• octave and oct2py installed in notebook base image, new octave-worker image deployed

Repo layout changes

• Conda specs moved to environment file
• Notebook and workers pull from base_environment.yml, then from any image-specific modifiers

Management utilities

• pinning/unpinning command line tool pin.py introduced

Known issues

• datashader+dask.dataframe results in a numba data type error on staging (#112)
• google fuse buckets not mounting on staging (#113)

Release version 0.2.6

In this release:

• Updates rhg_compute_tools to v0.1.5

Release version 0.2.5

In this release:

• Pulls dask-kubernetes from master (pinned to commit 295f064). This fixes the autoscaling & scale down issue.

Release version 0.2.4

In this release:

• Update jupyterlab and jupyterhub, along with extensions
• Update dask.distributed, dask_kubernetes
• Update xarray, pandas, scikit-learn, scikit-ml, dask, dask-ml, dask-searchcv, zarr, datashader, geoviews, holoviews
• Add dask-jupyterlab extension
• Update rhg-compute-tools

Release version 0.2.3

In this release:

• Notebook and worker Dockerfiles now pull from custom versions of jupyter/base-notebook and continuumio/miniconda3 with netcdf-fortran compiled using gfortran and gcc. intent here is to speed up downstream geoclaw build process by caching this lengthy step.
• Added nose and pyshp packages
• pip updated to version 18.0

Release version 0.2.2

Updates in this release:

• install & pin impactlab-tools, climate-toolbox, rhg-compute-tools
• install google cloud SDK & allow automatic worker authentication using GCLOUD_DEFAULT_TOKEN_FILE environment variable
• pin pip to 0.9.3

Release version 0.2.1

New features

• config is set up correctly, fixing #48 and #49
• dedupe sql proxy, fixing #34

Version 0.2.0 release

Updates in this release:

• Service account credentials are now copied to /opt/gcsfuse_tokens to enable easy zarr/parquet use
• Workers run the worker conda environment rather than root
• Various package updates