Execute Python code on the fly and display results in Tableau visualizations:

View the Project on GitHub tableau/TabPy

TabPy Server Configuration Instructions

Custom Settings

TabPy starts with set of default settings unless settings are provided via environment variables or with a config file.

Configuration parameters can be updated with:

  1. Adding environment variables - set the environment variable as required by your Operating System. When creating environment variables, use the same name for your environment variable as specified in the config file.
  2. Specifying a parameter in a config file (environment variable value overwrites configuration setting).

Configuration file with custom settings is specified as a command line parameter:

tabpy --config=path/to/my/config/file.conf

The default config file is provided to show you the default values but does not need to be present to run TabPy.

Configuration File Content

Configuration file consists of settings for TabPy itself and Python logger settings. You should only set parameters if you need different values than the defaults.

Environment variables can be used in the config file. Any instances of %(ENV_VAR)s will be replaced by the value of the environment variable ENV_VAR.

TabPy parameters explained below, the logger documentation can be found at logging.config documentation page.

[TabPy] parameters:

Configuration File Example

Note: Always use absolute paths for the configuration paths settings.

# TABPY_QUERY_OBJECT_PATH = /tmp/query_objects
# TABPY_PORT = 9004
# TABPY_STATE_PATH = <package-path>/tabpy/tabpy_server

# Where static pages live
# TABPY_STATIC_PATH = <package-path>/tabpy/tabpy_server/static

# For how to configure TabPy authentication read
# docs/
# TABPY_PWD_FILE = /path/to/password/file.txt

# To set up secure TabPy uncomment and modify the following lines.
# Note only PEM-encoded x509 certificates are supported.
# TABPY_CERTIFICATE_FILE = /path/to/certificate/file.crt
# TABPY_KEY_FILE = /path/to/key/file.key

# Log additional request details including caller IP, full URL, client
# end user info if provided.

# Limit request size (in Mb) - any request which size exceeds
# specified amount will be rejected by TabPy.
# Default value is 100 Mb.

# Enable evaluate api to execute ad-hoc Python scripts
# Enabled by default. Disabling it will result in 404 error.

# Configure how long a custom script provided to the /evaluate method
# will run before throwing a TimeoutError.
# The value should be a float representing the timeout time in seconds.

# Configure TabPy to support streaming data via Arrow Flight.
# This will cause an Arrow Flight server start up. The Arrow
# Flight port defaults to 13622 if not set here.






args=('tabpy_log.log', 'a', 1000000, 5)

format=%(asctime)s [%(levelname)s] (%(filename)s:%(module)s:%(lineno)d): %(message)s

Configuring HTTP vs HTTPS

By default, TabPy serves only HTTP requests. TabPy can be configured to serve only HTTPS requests by setting the following parameter in the config file:


If HTTPS is selected, the absolute paths to the cert and key file need to be specified in your config file using the following parameters:

TABPY_CERTIFICATE_FILE = C:/path/to/cert/file.crt
TABPY_KEY_FILE = C:/path/to/key/file.key

Note that only PEM-encoded x509 certificates are supported for the secure connection scenario.


TabPy supports basic access authentication (see for more details).

Enabling Authentication

To enable the feature specify the TABPY_PWD_FILE parameter in the TabPy configuration file with a fully qualified name:

TABPY_PWD_FILE = c:\path\to\password\file.txt

Password File

Password file is a text file containing usernames and hashed passwords per line separated by single space. For username only ASCII characters are supported. Usernames are case-insensitive.

Passwords in the password file are hashed with PBKDF2.

It is highly recommended to restrict access to the password file with hosting OS mechanisms. Ideally the file should only be accessible for reading with the account under which TabPy runs and TabPy admin account.

There is a tabpy-user command provided with tabpy package to operate with accounts in the password file. Run tabpy-user -h to see how to use it.

After making any changes to the password file, TabPy needs to be restarted.

Adding an Account

To add an account run tabpy-user add command providing user name, password (optional) and password file:

tabpy-user add -u <username> -p <password> -f <pwdfile>

If the (recommended) -p argument is not provided a password for the user name will be generated and displayed in the command line.

Updating an Account

To update the password for an account run tabpy-user update command:

tabpy-user update -u <username> -p <password> -f <pwdfile>

If the (recommended) -p agrument is not provided a password for the user name will be generated and displayed in the command line.

Deleting an Account

To delete an account open password file in any text editor and delete the line with the user name.

Endpoint Security

All endpoints require authentication if it is enabled for the server.

Arrow Flight

TabPy can be configured to enable Arrow Flight. This will cause a Flight server to start up alongside the HTTP server and will allow for handling incoming streamed data in the Arrow columnar format.

As of May 2023, the Arrow Flight feature can only be used by compatible versions of Tableau Prep. The Arrow Flight feature is not used by Tableau Desktop, Tableau Server, or Tableau Cloud, regardless of the TABPY_ARROW_ENABLE setting. In other words, those products will continue to send the data in a single payload when Arrow Flight is both enabled and disabled.

To leverage the Flight server, use an existing Flight Client API. There are implementations available in C++, Java, and Python. To begin streaming data to the server, a Flight Descriptor (data path) must be generated. One can be obtained via the TabPy Flight server by using the client to submit a getUniquePath Action to the server or it can be randomly generated locally. The client’s do_put interface can then be used to begin sending data to the server.

Structure the data payload in Arrow format according to the client’s API requirements. Continue using the client to append the data path with the data stream.

The mechanism for sending the Python script to the server does not change.


Logging for TabPy is implemented with Python’s standard logger and can be configured as explained in Python documentation at Logging Configuration page.

A default config provided with TabPy is at tabpy-server/tabpy_server/common/default.conf and has a configuration for console and file loggers. Changing the config file allows the user to modify the log level, format of the logged messages and add or remove loggers.

Request Context Logging

For extended logging (e.g. for auditing purposes) additional logging can be turned on with setting TABPY_LOG_DETAILS configuration file parameter to true.

With the feature on additional information is logged for HTTP requests: caller ip, URL, client infomation (Tableau Desktop\Server) and TabPy user name as shown in the example below:

2019-05-02,13:50:08 [INFO] ( Call ID: 934073bd-0d29-46d3-b693-b1e4b1efa9e4, Caller: ::1, Method: POST, Resource: http://localhost:9004/evaluate, Client: Postman for manual testing
2019-05-02,13:50:08 [DEBUG] ( Checking if need to handle authentication, <<
call ID: 934073bd-0d29-46d3-b693-b1e4b1efa9e4>>
2019-05-02,13:50:08 [DEBUG] ( Handling authentication, <<call ID: 934073bd-
2019-05-02,13:50:08 [DEBUG] ( Checking request headers for authentication d
ata, <<call ID: 934073bd-0d29-46d3-b693-b1e4b1efa9e4>>
2019-05-02,13:50:08 [DEBUG] ( Validating credentials for user name "user1",
 <<call ID: 934073bd-0d29-46d3-b693-b1e4b1efa9e4>>
2019-05-02,13:50:08 [DEBUG] ( Collecting Access-Control-Allow-Origin from state file...  
2019-05-02,13:50:08 [INFO] ( function to evaluate=def _user_script(tabpy, _
arg1, _arg2):
 res = []
 for i in range(len(_arg1)):
   res.append(_arg1[i] * _arg2[i])
 return res
, <<call ID: 934073bd-0d29-46d3-b693-b1e4b1efa9e4>>

No passwords are logged.

NOTE the request context details are logged with INFO level.