Release Notes for version 2.1
This is the list of changes of taipy version 2.1.
Unsupported version
Version 2.1 of Taipy is no longer supported. We strongly recommend that you upgrade to the latest version of Taipy.
Published on 2023-01.
taipy 2.1 contains the latest
taipy-config 2.1,
taipy-gui 2.1,
taipy-core 2.1 and
taipy-rest 2.1 packages.
Please refer to the Migration page for details on how to migrate from version older than 2.1.
taipy¶
2.1.0¶
New features
- Taipy and all its dependencies
now support Python 3.11.
See Python documentation for details.
taipy-gui¶
2.1.0¶
New features
- A security feature has been added: the file
.taipyignore, located next to the Python main file, can list the paths that you want to prevent access to.
See issue #501 or this section for details. - Charts can use the new
Decimatorclass to cleverly filter data points out to significantly improve performance.
See the paragraph on large datasets for specific information. - Charts now support polar, funnel, candlesticks and
many other types of charts.
See the chart control section for details. - Charts now support the dark theme automatically.
- Tooltips can be set on individual table cells.
See the example for more information. -
Long running callbacks
have been improved to allow for easily returning a value.
See the documentation of theinvoke_long_callback()function or the issue #547 for more details. - Developers can specify the location of the Taipy
webapp, for debugging purposes.
The--webapp-pathcommand line option allows to specify that location.
See issue #564.
Improvements and changes
- The Pie charts now use the values property
to set values instead of x.
See Pie charts for details. - Unselected data points or traces in charts now preserve
their original opacity.
See issue #496. -
class_nameis now a dynamic property.
See issue #480. - The allow_unsafe_werkzeug option of Werkzeug (that Flask depends on for the WSGI part) is forced to True when the Gui instance is run in Debug mode, because of a change in policy in recent updates.
taipy-core¶
2.1.2¶
Significant bug fixes
- The version required for openpyxl has been downgraded from "openpyxl>=3.0.7,<4.0" to "openpyxl>=3.0.7,<3.1" to match the version used by Modin.
2.1.1¶
Improvements and changes
- Add overload type descriptions for the
taipy.get()method that supports multiple different combinations of argument types.
2.1.0¶
New features
- New version management system for Taipy
applications. Users can now run an application in development mode, save a version of
the application as an experiment version, re-run older experiment versions, and push a
version to production.
See the Version management system documentation page for more details. - New data node named MongoCollectionDataNode. It represents the data from a MongoDB collection.
- New data node named ParquetDataNode. It represents tabular data stored in the Apache Parquet format.
- Added support for Modin as a new exposed type.
- Running the Orchestrator service is required to execute
jobs. See
Orchestrator().run()method. - The parent entities of a data node, a task, or a
pipeline can be accessed via
DataNode.get_parents(),Task.get_parents(), orPipeline.get_parents(), or by passing the data node entity, task entity or pipeline entity to the functiontaipy.get_parents(). - New data node property expiration_date computed adding the validity_period duration to the last_edit_date of the data node.
- New data node property is_up_to_date equals to
Trueif the data node has not expired (refer to expiration_date attribute).Falseotherwise. - The sql repository_type is now available on community edition to store Taipy entities in an SQL database.
Improvements and changes
- The last_edit_date attribute of a data node is now updated when the corresponding data is modified by either a Taipy task execution or an external factor. This behavior is limited to file-based data nodes: CSV, Excel, JSON, and pickle data nodes only.
Deprecations
- Deprecation of the data node cacheable property. It is
replaced by skippable property on tasks.
The mechanism remains unchanged but instead of setting cacheable property to
Truefor all the outputs of a task that can be skipped, just set the task skippable property toTrue.