v0.20 Changelog


Released: May 13, 2021


  • [general] [feature]

    Changed the project configuration for samplers and workflow generators. See migration guide for details.

    References: #147, pull request 187

  • [general]

    Removing master/slave language from AWS EMR configuration in project schema.

    References: #164, pull request 205

  • [general] [eagle]

    Fix for create_eagle_env.sh not creating environment.

    References: pull request 226


  • [workflow]

    Adding measure arguments for reporting measures in the workflow generator.

    References: #189, pull request 219


  • [postprocessing]

    Use a map of dask delayed function to combine parquets instead of a giant dask df to avoid memory issues. Default to 85GB memory nodes in eagle with single process and single thread in each node to avoid memory issues.

    References: pull request 212

  • [postprocessing]

    The glue crawler was failing when there was a trailing / character. This fixes that as well as checks to make sure files were uploaded before running the crawler.

    References: #159, pull request 202

  • [postprocessing]

    Moves the eagle.postprocessing.keep_intermediate_files to postprocessing.keep_individual_timeseries and changes behavior to keep only the timeseries parquet files. Also, removes the deprecated aggregate_timeseries key as that aggregation always happens.

    References: #182, pull request 228

  • [postprocessing] [bugfix]

    Previously the postprocessing would fail if an upgrade scenario didn’t have any timeseries simulation output. Now it will skip it and post a warning message. This was fixed previously, but now we have tests for it.

    References: #199, pull request 230



  • [bugfix]

    Fixing issue where the postprocessing fails when a building simulation crashes in buildstockbatch.

    References: #196, pull request 197

  • [bugfix]

    Defaults to the newer datetime encoding in the parquet files now that Athena can understand it.

    References: #221, pull request 224