Ctrl+K

What’s new#

v2406.0.0#

Bug fixes#

  • Fixed plugin batchmode submit bug. Batchmode plugins did not pick up the correct evaluation system config file. This has been fixed now.

  • Revised minor typographical errors in the documentation and code comments.

  • Created a standard_main function in utils.py for cli

Documentations#

  • Added documentation of the freva databrowser rest api.

Internal Changes#

  • Updated unit tests.

  • Added .pre-commit-config.yaml

v2309.0.1#

New Features#

Breaking changes#

Deprecations#

Bug fixes#

Documentations#

  • added documentation to esgf method both for cli and python module.

Internal Changes#

v2309.0.0#

New Features#

  • Plugin and history output can be parsed using the --json flag from the

    command line

  • The plugin method and cli plugin command can be instructed to wait for any batch job. wait=True in the python module or --wait in the cli

  • Users can interact with the plugin output, using the return value of the freva.run_plugin method.

Breaking changes#

Deprecations#

Bug fixes#

Documentations#

Internal Changes#

v2307.0.2#

New Features#

Breaking changes#

Deprecations#

  • cli freva databrowser --all-facets is made deprecated, use --facet '*' or --facet all instead

Bug fixes#

  • cli freva plugin --unique-output False feature reinstated, working as within python module.

Documentations#

Internal Changes#

v2303.0.0#

New Features#

  • User data can be ingested also if not all data files match directory reference standard.

  • A new databrowser key (uri) was added that representing the object storage where the data is stored.

Breaking changes#

Deprecations#

Bug fixes#

  • Only save the hostname when submitting jobs, not the host + domain name

Documentations#

Internal Changes#

v2208.1.1#

New Features#

Breaking changes#

Deprecations#

Bug fixes#

  • Fix typo in deployment routine

Documentations#

  • Fix binder instance

Internal Changes#

  • Make all scheduler options non-mandatory

  • Add metadata-inspector as dependency

v2208.0.1#

New Features#

  • count keyword/flag for the databrowser counts also the number of files found by the databrowser, not only the facet counts.

  • databrowser can subset results by time.

  • add functionality to add new data to databrowser

  • add functionality to delete existing user data from databrowser

  • add new class to handle user data requests

Breaking changes#

  • count_facet_values keyword (count-facet-values flag in cli) has been renamed to count

Deprecations#

  • move freva.crawl_my_data functionality to freva.UsersData.index

  • renamed freva crawl_my_data to freva user-data`

Bug fixes#

Documentations#

  • Documentation of the python modules has been improved

Internal Changes#

  • Files found by the databrowser are alphabetically sorted.

  • follow/unfollow_history_tag receives a user object as argument instead of string

  • lazy loading has been improved

v2206.0.10#

New Features#

Breaking changes#

Deprecations#

Bug fixes#

  • Set plugin status to broken in OS related termination signals that can be handled by python (SIGTERM, SIGINT …) and any other internal python errors.

  • Update databrowser query algorithm.

Documentations#

Internal Changes#

  • Introduce lazy loading to make freva cli a little more responsive

  • Explicitly set EVALUATION_SYSTEM_CONFIG_FILE env variable in workload manager job script.

  • Increase batch size to 5000 as default for querying data.

v2206.0.1#

New Features#

  • Add python module for importing freva() directly in python environments

  • Rename command line interface sub modules:

    • freva --databrowser → freva databrowser,

    • freva --plugin → freva plugin

    • freva --history → freva history

    • freva --esgf → freva esgf

    • freva --crawl_my_data → freva crawl-my-data

  • Add new commands as alternative for freva sub commands

    • freva-databrowser = freva databrowser etc

  • Argument completion works for all sub commands, including plugin setup.

  • Interactive jobs can be controlled from the web user interface

  • Output of interactive jobs can be displayed in the web user interface

Breaking changes#

  • Command line arguments taking boolean values do not require True or False:

    • Instead of freva plugin animtaor --batchmode=True → freva plugin animator --batchmode

  • add_output_to_databrowser() formerly linkmydata() does not infer meta data from file and directory structure anymore (i.e., does not need to follow a particular folder path and naming convention)

Deprecations#

  • The tool-pull-request sub command has been made deprecated.

  • The following methods in the Plugin API Reference have been renamed:

    • runTool() → run_tool()

    • linkmydata() → add_output_to_databrowser()

    • prepareOutput() → prepare_output()

    • getHelp() → get_help()

    • getClassBaseDir() → class_basedir

    • setupConfiguration() → setup_configuration()

    • readConfiguration() → read_configuration()

  • The following methods in the Parameter API Reference have been renamed:

    • parseArguments() → parse_arguments()

Bug fixes#

Documentations#

  • Add user sphinx documentation

Internal Changes#

  • Add support for different workload managers

  • Install freva in dedicated anaconda environment

  • Install each Freva plugin in dedicated anaconda environment

On this page