client is the in-browser engine for management of sampling and transformation in the Transformer page. is automatically downloaded and updated in each user's web browser when accessing the Transformer page. Within the browser, Photon is used to manage sampling and to process transformations on those samples on the local web client, which limits required interactions with the .
Larger sample sizes (up to 10MB by default).
NOTE: Since supports larger sample sizes, some interactions may be slightly impacted. Loading states have been introduced to enable faster responsiveness from the application.
Tip: For datasets that are smaller than the sample size limit, the Transformer Page displays the entire dataset in its transformed state. So, you can download the dataset from the Recipe Panel in Transformer Page without having to execute it on a remote server. This expanded capability allows for faster and more immediate local use of the product. See Recipe Panel.
This section contains the user-facing configuration for the client. Except as noted below, these configuration changes are applied to the , which then applies the configuration to each instance of the client and its interactions with the node.
NOTE: Some configuration is shared with the running environment. For more information, see Configure Photon Running Environment.
On the , you can make adjustments to the resources claimed by the Photon running environment based on the number of cores on the machine. The following table identifies the recommended settings for a node with 8, 16, or 32 cores. The default settings assume 16 cores.
|Parameter||8 cores||16 cores (default)||32 cores|
The number of simultaneous users is a competing factor.
The following table illustrates some adjustments for a 16-core system:
|Parameter||16 cores (default)||Low number of simultaneous users||High number of simultaneous users|
NOTE: Increasing these values can have a significant impact on load times and performance. Change these values only if you are experiencing difficulties. Make incremental changes.
Increasing the sample size may degrade the user experience in the Transformer page in the following ways:
Time required to complete brushing and linking in histograms
NOTE: If you increase the sample size above the default setting and encounter unacceptable performance in the above areas, you should reduce the sample size settings.
When samples are created using the running environment, their maximum size is determined by the following parameter. Default value is 10 MB.
NOTE: For datasets created from relational sources, if the dataset is larger than this value, a random sample is automatically generated. For smaller datasets, no sample is automatically generated, and the initial sample is used.
The following parameter sets the upper limit to which a sample can grow in the browser. After a sample has been gathered, adding some kinds of transformations, such as joins or adding new columns, can grow the volume of data in the browser. This parameter represents the upper volume of data that can be displayed for a sample + transformations.
Tip: The value for this parameter should be approximately four times the value for
NOTE: If the data volume of the same grows beyond this size, the number of rows available in the sample is reduced until the sample volume is below this threshold. This truncation happens automatically and without warning.
Tip: You can re-sample the dataset at any time, which creates a new sample that is below the value for
The client is an embedded component of the web client. Access to it is transparent to the user.
Photon interacts with backend datastores through the VFS service.
NOTE: The VFS service is enabled by default and must be enabled when running environment is in use.
For more information, see Configure VFS Service.