tabpy

TabPy (the Tableau Python Server) is an external service implementation which expands Tableau's capabilities by allowing users to execute Python scripts and saved functions via Tableau's table calculations.

Changelog

v2.13.0

Improvements

• Add support for deploying functions to a remote TabPy server by setting remote_server=True when creating the Client instance.

v2.12.0

Improvements

• Add support for public deployed functions that will be visible to users in Tableau when using the Custom Functions Explorer
• Add functionality to allow users to update existing deployed functions without needing to redeploy the function itself

v2.11.0

Improvements

• Add support for Python 3.10, 3.11, and 3.12. End support Python for 3.7 and 3.8.

v2.10.0

Improvements

• Add TabPy parameter (TABPY_MINIMUM_TLS_VERSION) to specify the minimum TLS version that the server will accept for secure connections. Default is set to TLSv1_2.

v2.9.0

Improvements

• Require confirmation to continue when starting TabPy without authentication, with a warning that this is an insecure state and not recommended.

v2.8.0

Improvements

• Returns 413 error code when request payload exceeds TABPY_MAX_REQUEST_SIZE_MB config setting.

v2.7.0

Improvements

• Adds support for data streaming in Arrow columnar format via Apache Arrow Flight.

v2.6.0

Improvements

• Fixes deprecation of sklearn in favor of current package name scikit-learn

v2.5.1

Improvements

• Gzip encoded requests are now supported by default. This can be disabled in the config file.
• The INFO method will return the enabled status of features.

v2.5.0

Improvements

• A server with Adhoc Disabled Flag on with the wrong credentials will now return wrong credentials error instead of telling the user that Adhoc Scripts are not allowed on this server.
• Added documentation for how to run TabPy projects with local changes

Breaking changes

• Discontinued support for Python 3.6
• Added support for Python 3.9

v2.4.0

Improvements

• Add toggle to turn off evaluate API.

Breaking changes

• Changing error code to 406 when server not configured for authentication but credentials are provided by client.

v2.3.2

Improvements

• Test files added to tar.gz and zip releases.

v2.3.1

Bug fixes

• Overriding deployed models.

v2.3.0

Improvements

• Fixed scrutinizer test run failure.

v2.2.0

Breaking changes

• TabPy fails with 400 when it is not configure for authentication but credentials are provided by client.

Bug fixes

• When TabPy is running with no console attached it is not failing with 500 when trying to respond with 401 status.
• tabpy.query() failing when auth is configured.

Improvements

• Minor code cleanup.

v1.1.0

Improvements

• Authorization is now required for the /info API method. This method did not check authentication previously. This change is backwards compatible with Tableau clients.
• Improved config parsing flexibility. Previously the TABPY_EVALUATE_TIMEOUT setting would be set to a default if tabpy couldn't parse the value. Now it will throw an exception at startup.

v1.0.0

Improvements

• Minor: feature name changed to analytics extensions.
• Startup script files deleted.
• Index page updated.

Other

• TabPy is now Tableau Supported (used to be Community Supported).

v0.9.0

Improvements

• Models deployment doesn't depend on pip._internal anymore.
• Package size reduced.

v0.8.13

Improvements

• TabPy works with Python 3.8 now.
• Documentation updates with referencing Tableau Help pages.
• Added Client.remove() method for deleting deployed models.

Bug Fixes

• Fixed failing Ctrl+C handler.
• Fixed query_timeout bug.
• Fixed None in result collection bug.
• Fixed script evaluation with missing result/return bug.
• Fixed startup failure on Windows for Python 3.8.

v0.8.9

Improvements

• Added Ctrl+C handler
• Added configurable buffer size for HTTP requests
• Added anvoa to supported pre-deployed models in tabpy

v0.8.7

Improvements

• Enabled the use of environment variables in the config file.

v0.8.6

Fixes

• Fixed file names for package building.
• Fixed reading version info for /info call.

v0.8

Improvements

• TabPy is pip package now
• Models are deployed with updated script

v0.7

Improvements

• Added t-test model
• Fixed models call with /evaluate for HTTPS
• Migrated to Tornado 6
• Timeout is configurable with TABPY_EVALUATE_TIMEOUT config file option

v0.6.1

Improvements

• Scripts, documentation, and integration tests for models
• Small bug fixes
• Added request context logging as a feature controlled with TABPY_LOG_DETAILS configuration setting.
• Updated documentation for /info method and v1 API.
• Added integration tests.

v0.4

Improvements

• Added basic access authentication (all methods except /info)
• tabpy-tools can deploy models to TabPy with authentication on
• Increased unit tests coverage
• Travis CI for merge requests: unit tests executed, code style checking

v0.3.2

Breaking changes

• Logger configuration now is in TabPy config file.

Improvements

• Remove versioneer and just replace it with VERSION file
• Require Python 3.6.5
• Require jsonschema to be compatible with 2.3.0
• Added setup instructions (known issues) for CentOS

v0.3.1

• Fixed dependency on tabpy-tools in startup scripts
• Fixed Python version dependency in tabpy-server setup script

v0.3

Breaking changes

• The config file is now not just Python code but an actual config
• Tornado config file has a different setting for CORS
• Setup scripts are deleted - setup (if needed) happens with the startup script
• tabpy-client is tabpy-tools now

Improvements

• Secure connection (HTTPS) is supported with Tableau 2019.2 and newer versions
• Documentation is improved with more examples added
• Versioning is done with Versioneer and github release tags
• Improved logging
• Unit tests are passing now
• Configurations for Postman and Swagger are available to use those against running TabPy

v0.2

• Initial version