Release Notes for version 3.0

This is the list of changes of taipy version 3.0.

Unsupported version

Version 3.0 of Taipy is no longer supported. We strongly recommend that you upgrade to the latest version of Taipy.

Error when downgrading from Taipy 3.1 to Taipy 3.0

If you are experiencing an error downgrading to Taipy 3.0, please reinstall Taipy with options --no-cache-dir -I like so: pip install --no-cache-dir -I taipy==3.0.0

See issue #1180 for more details.

Published on 2023-10.

taipy 3.0 contains the latest taipy-config 3.0, taipy-gui 3.0, taipy-core 3.0, taipy-templates 3.0, and taipy-rest 3.0 packages.

taipy

3.0.0

New features

• Taipy application can now be run with the Taipy command-line interface (CLI) using the taipy run command. For more information, refer to Run application in Taipy CLI.

taipy-gui

3.0.0

New features

• A new package holds the Page Builder API: a set of classes that let you define the pages for your application entirely with Python.
• You can now update variables on all clients using the shared variables concept. See the Gui.add_shared_variable() and State.dispatch() methods for details.
• You can now invoke a callback for all clients using the broadcast_callback() function.
• The slider control can now handle several knobs, allowing for range selection.
  Please check the example for more information.
• The file_download control now lets developers generate the file content dynamically, at download time.
  Please check the example for more information.
• A new CSS class called toggle-navbar was added to the Stylekit to give a toggle control the aspect of a navbar.
• The chart control now supports the treemap and waterfall chart types.

Improvements and changes

• The navigate() function has an additional parameter params that is used to add query parameters to the requested URL. The query parameters can be retrieved in the on_navigate callback.
• The on_action parameter of the download() function can be a function and not just a function name.
• Setting the debug parameter of Gui.run() to True provides stack traces to be shown in the console when exceptions occur in user code.

Significant bug fixes

• The callback function set to the on_action parameter of the function download() may be called too early. It is now ensured to be invoked after the download operation is performed.
  See issue #916.
• Setting the properties property of a visual element as the returned value from a function may not succeed.
  See issue #897.
• Variables imported by an import * directive are not handled properly in the state of a callback defined in the importing module.
  See issue #908.
• The date control does not use the format property if with_time is not set.
  See issue #909.
• The date control uses the datetime.date type and does not apply time zones if time is not involved.
  See issue #895 and issue #923.
• Updating a chart control data may cause data congestion or display flickering.
  See issue #864 and issue #932.
• Selection in a chart with type pie type is not properly handled.
  See issue #919.
• Hover text doesn't show properly in a selector that is crowded.
  See issue #927.
• Options with a long text in a selector cannot cbe deselected.
  See issue #917.
• The table control does not support undefined date values from Pandas data frames.
  See issue #886.

Deprecations

• The action parameter of the on_action callback was removed for every control.
  The signature of all on_action() callback functions are now unified to the following:

  • state (State): the state of the client invoking that callback;
  • id (str): the identifier of the visual element that triggers that callback;
  • payload (dict): a dictionary that provides additional information to the callback.
    This dictionary now has the additional action key that is set to the action name.

  This change not only impacts the on_action callback of all controls that support it, but in an exactly similar manner the following callback signatures:

  • on_range_change in the chart control;
  • on_edit, on_add, and on_delete in the table control;
  • on_close in the pane block.

taipy-core

3.0.0

New features

• A production version of a Taipy application can now be provided with migration functions to automatically migrate entities and keep them compatible with previous versions.
  For more information, refer to Production mode.
• A GLOBAL scope data node can be created from a data node configuration calling the new taipy.create_global_data_node() method.
  For more information, refer to Create a data node.
• A data node configuration can be built from an existing data node configuration. For more information, refer to the documentation page on data node configuration.

• A new class Submittable models entities that can be submitted for execution. It is an Abstract class instantiated by Scenario and Sequence. It can be handy to use the new following Submittable methods:

  • Submittable.get_inputs() retrieves input data nodes of a Submittable entity;
  • Submittable.get_outputs() retrieves output data nodes of a Submittable entity;
  • Submittable.get_intermediate() retrieves intermediate data nodes of a Submittable entity;
  • Submittable.is_ready_to_run() checks if an entity is ready to be run;
  • Submittable.data_nodes_being_edited() retrieves data nodes that are being edited of a Submittable entity.

• New functions exposed by the taipy module:

  • taipy.is_deletable() checks if an entity can be deleted;
  • taipy.exists() checks if an entity exists.

• The encoding type of CSVDataNode and JSONDataNode can now be configured using the encoding parameter. For more information, please refer to Configure a CSVDataNode and Configure a JSONDataNode sections.

Improvements and changes

• Function DataNode.track_edit has been made public.

Significant bug fixes

• When running the Orchestrator service in development mode, changing the name of the function used by a task then running the application again would raise an error.
  See issue #743.

Deprecations

• A ScenarioConfig graph is now created directly from TaskConfig and DataNodeConfig. Consequently, PipelineConfig has been removed. For more information, refer to Configure a scenario.
• The Pipeline object has been removed and replaced by Sequence. A sequence is held by a Scenario and represents a subset of its tasks than can be submitted together independently of the other tasks of the scenario. For more information, refer to Scenario.add_sequence() and Scenario.remove_sequence().
• Scope.PIPELINE has been removed from possible Scope values.
• The root_folder, storage_folder, read_entity_retry, repository_type, and repository_properties attributes of the GlobalAppConfig have been moved to the CoreSection.
  Please refer to the Core configuration page for details.
• The clean_entities attribute has been removed from the CoreSection. Correspondingly, the --clean-entities option has been removed from the version management CLI.
  To clean entities of a version, please run your application in development mode, or delete your version with the --delete CLI option. For more information, refer to Taipy command-line interface
• The deprecated nb_of_workers attribute of the JobConfig has been removed.
• The deprecated parent_id attribute of a DataNode, Task, Pipeline, or Scenario entity, has been removed.
• The deprecated last_edition_date and edition_in_progress attributes of a DataNode entity have been removed.
• The deprecated DataNode.lock_edition() and DataNode.unlock_edition() methods have been removed.
• The deprecated taipy.create_pipeline() method has been removed.

taipy-template

3.0.0

New features

• A new template named "scenario-management" is available. For more information on creating a new Taipy application with the new "scenario-management" template, refer to the documentation page on templates.

Improvements and changes

• The default template also supports creating a multi-pages application with Orchestrator and Rest services. These options are available when creating a new application from the template.
• The "multi-page-gui" template has been removed. Please use the default instead to create a Taipy multi-pages application.

taipy-enterprise

3.0.0

New features

• Python functions including scenario management methods can be scheduled to run at a specific time using the new taipy.Scheduler API. For more information, refer to Schedule a method.

Improvements and changes

• The job recovery mechanism is now only available when the Orchestrator service is run.